v2026.08.18

Tuesday, August 18, 2026

Gloria Estefan

Ayer

0:005:15

Weather

Bloomington

66°

Clear

H 82°L 63°

Feels like: 68°
Humidity: 88%
Cloud: 0%
Rain: 2%

Today

82°

63°

Wed

83°

69°

46%

Thu

80°

64°

39%

Fri

84°

60°

Sat

84°

61°

23%

Sun

82°

65°

Mon

66°

61°

News

europe/conflict

Ukraine launches massive drone strikes against Moscow and occupied regions

Ukraine launched a significant aerial offensive overnight, deploying over 600 drones against targets in Moscow and various Russian-controlled areas including Crimea. Russian officials reported damage to a warehouse and injuries to at least three civilians, including a child. The scale of this operation marks an aggressive shift in the ongoing conflict, as Ukraine intensifies its long-range capabilities to hit infrastructure deep within Russian territory. Meanwhile, retaliatory Russian strikes caused casualties in the Kharkiv region, underscoring the escalating intensity of the current tactical exchange.

Version v2026.08.18 - Daily Briefing

africa/politics

Hakainde Hichilema re-elected as President of Zambia

President Hakainde Hichilema has secured a second term in office after receiving approximately 60 percent of the vote in the national election. Despite reports of localized attacks on election officials that briefly disrupted the counting process, the outcome was confirmed, giving Hichilema a decisive mandate for another term. Following the announcement, the incumbent president expressed gratitude to the electorate and pledged to continue working for the benefit of all citizens in the resource-rich southern African nation.

Zambia holds substantial geopolitical importance due to its status as a major copper producer and its role in regional economic stability. Hichilema's re-election is a key indicator of current political alignment in the region, as his administration has focused on economic reform and stability. The fact that the election survived security disruptions during the count is an important test for the country's democratic resilience, demonstrating that institutional processes managed to withstand pressure during a period of high social tension.

Stability in Zambia directly impacts global supply chains for critical minerals, particularly copper, which is essential for the global energy transition.

BBC NewsAl Jazeera

asia/politics

South Korean President pushes for military independence from US

South Korean President Lee Jae Myung has renewed calls for his country to regain full independent command of its military. This move follows President Donald Trump's decision to scale back joint military exercises between the two nations, a policy change attributed to the US president's diplomatic efforts with North Korean leader Kim Jong Un. The South Korean government is also accelerating its acquisition of nuclear-powered submarines to ensure national security remains in Seoul's control rather than dependent on changing political currents in Washington.

The US-South Korea alliance has been a cornerstone of Pacific security for decades. By pushing for independent control, or OPCON, Seoul is signaling that it no longer views the American security umbrella as guaranteed or static. If realized, this shift would fundamentally alter the security architecture of the Korean Peninsula and potentially encourage other regional allies to seek greater military self-reliance, fundamentally changing how the United States projects influence in Asia.

Increased military autonomy in Seoul risks destabilizing regional deterrence and complicates the existing security architecture designed to contain North Korean aggression.

The Guardian

middle east/conflict

Unknown projectile damages vessel in Strait of Hormuz

A maritime vessel navigating the Strait of Hormuz was struck by an unidentified projectile earlier today, according to the United Kingdom Maritime Trade Operations (UKMTO). The impact caused damage to the ship's engine room and resulted in at least one casualty. While investigations are ongoing, the incident has already heightened concerns regarding the safety of one of the world's most critical maritime chokepoints, through which a significant percentage of the world's oil supply passes.

The Strait of Hormuz is a vital conduit for global energy security, and any instability here poses an immediate threat to international commodity prices. Because the perpetrator of the strike remains unknown, it creates a dangerous environment of uncertainty that can escalate into a wider regional standoff. This type of incident often serves as a proxy for larger geopolitical tensions in the region, forcing shipping companies to potentially hike insurance premiums and reroute vessels, which carries broad economic consequences for global supply chains.

Disruptions in the Strait of Hormuz act as a massive tax on global trade, directly inflating energy costs for consumers worldwide.

Al Jazeera

europe/politics

Germany launches first raids to dismantle small boat smuggling gangs

German law enforcement agencies have executed a series of coordinated raids targeting organized criminal networks responsible for smuggling migrants across borders in small boats. Authorities successfully seized dinghies, life jackets, and other equipment used by these groups during the operation. This represents the first major enforcement effort under new, stricter legislation designed to crack down on human smuggling. These measures follow increasing pressure on the government to better manage irregular migration flows and disrupt the networks that facilitate them.

The mechanism of smuggling depends on high-risk, low-cost equipment that human traffickers exploit for profit. By specifically targeting the supply chain—the boats and safety gear—the German government aims to make the logistics of smuggling prohibitively expensive and difficult for criminal groups. This strategy reflects a broader European attempt to move away from reactive border control toward proactive disruption of the criminal infrastructure, which is essential to mitigating the humanitarian crisis often caused by these dangerous, poorly planned crossings.

Disrupting smuggling logistics is a prerequisite for any cohesive immigration policy, as it reduces the immediate danger to individuals attempting to cross.

BBC News

oceania/economy

BHP report reveals over 100 cases of sexual harassment in one year

Mining giant BHP has released an annual report confirming that at least 109 of its workers sexually harassed colleagues within a single financial year. The company noted that these incidents led to the resignation or termination of those responsible. An additional 104 allegations remain uninvestigated because the victims either requested that no action be taken or because there was insufficient evidence to proceed. This transparency report highlights a systemic issue within the male-dominated mining industry in Australia, sparking renewed demands for safer workplaces.

Mining is a critical component of the Australian economy, and these figures highlight a pervasive cultural problem that has long been ignored. The mechanism of sexual harassment in these environments is often linked to isolated work sites and hierarchical power dynamics. By publishing this data, BHP is under pressure to move beyond simple terminations toward structural changes in how it monitors behavior and supports staff. If the industry fails to fix these issues, it will struggle to recruit and retain the diverse talent necessary for modern mining operations.

The persistence of harassment culture hinders industry productivity and risks significant legal and reputational damage to major economic players.

The Guardian

africa/health

Ebola outbreak in DRC becomes country's deadliest in history

The current Ebola outbreak in the Democratic Republic of the Congo has officially reached a grim milestone, claiming at least 2,325 lives. This figure surpasses the death toll from the significant 2018-2020 epidemic, making it the most lethal such event in the nation's records. Public health officials are struggling to contain the rapid spread of the virus, as confirmed cases continue to rise across the affected regions.

Ebola is a severe, often fatal viral illness characterized by sudden fever, fatigue, muscle pain, and internal or external bleeding. It spreads through direct contact with the blood or bodily fluids of infected individuals or contaminated surfaces. The sheer scale of this outbreak highlights a critical fragility in local healthcare infrastructure and the immense logistical hurdles inherent in conducting vaccination and containment efforts in rural, conflict-prone areas.

The record-breaking death toll underscores a failure in regional containment strategies and threatens to overwhelm the already strained healthcare systems in central Africa.

The Guardian

americas/economy

US gas prices reach record August highs amid Middle East instability

United States gasoline prices have climbed to their highest levels ever recorded for the month of August. Data from AAA shows the national average hitting 4.06 dollars per gallon this week. These rising costs are directly linked to the stalled peace negotiations between the United States and Iran, which have been exacerbated by the ongoing closure of the Strait of Hormuz.

Oil markets rely on the stability of major shipping lanes like the Strait of Hormuz to transport crude globally. When geopolitical tensions surge and these passages face closure or threats, global supply chains tighten immediately. This creates a supply shock that forces up the price of oil, which is then passed directly to consumers at the pump in the form of higher fuel costs, regardless of domestic production levels.

Elevated energy costs act as a regressive tax on consumers, potentially dampening broader economic growth and complicating the political environment for the administration.

The Guardian Business

asia/economy

China's economic growth slows with slump in industrial output

China's economy is showing clear signs of a deepening slowdown as new data reveals a significant slump in both industrial output and retail sales throughout July. This downturn follows what was already one of the weakest quarterly growth performances in the country's recent history. The trend has placed immense pressure on Beijing to consider aggressive intervention to prop up the world's second-largest economy.

Industrial output measures the total volume of production by factories and mines, while retail sales reflect consumer demand. When both fall simultaneously, it signals a systemic cooling effect where businesses are producing less because they expect slower consumption. Given China's central role as the global engine for manufacturing and trade, a sustained period of low growth there creates a ripple effect, lowering demand for raw materials from exporters and impacting international financial markets.

A sustained Chinese economic contraction reduces global trade volume and diminishes demand for commodities, putting pressure on developing nations that rely on Chinese growth.

The Guardian

europe/climate

European farmers report catastrophic harvest failures from drought

Farmers across Europe are warning of an unprecedented agricultural crisis caused by successive intense heatwaves and a severe, continent-wide drought. Grain and vegetable growers report that their harvests are nearing catastrophic levels, with French vegetable production seeing particularly steep declines. The situation has become so dire that national governments are now facing public outcry over their ability to support the farming sector as the heat persists.

Agricultural output depends on predictable soil moisture and temperature ranges for crops to mature. When heatwaves and drought occur consecutively, they disrupt the biological cycles of plants, leading to widespread crop failure. This creates a scarcity in the food supply chain that generally leads to higher retail prices for staples. Furthermore, it places intense strain on water management systems, forcing authorities to implement restrictions like hosepipe bans for the general population.

Chronic failure in European agricultural yields will likely increase food prices and heighten reliance on international food imports during a volatile climate period.

The Guardian

middle east/conflict

Netanyahu rejects ceasefire deal unless Hamas disarms first

Israeli Prime Minister Benjamin Netanyahu met with US envoy Jared Kushner to discuss the stalled Gaza peace plan, but the talks yielded no breakthrough. Netanyahu maintained his firm stance that Israeli troops will not withdraw from the coastal strip until Hamas fully disarms. This comes as the US attempts to salvage a 15-point ceasefire plan that had previously faced skepticism from both Israeli leadership and other regional actors involved in the ongoing conflict.

Disarmament is a fundamental demand in many security negotiations, meant to ensure that a group cannot return to offensive operations. By conditioning withdrawal on this, Netanyahu is effectively linking the cessation of hostilities to the total dissolution of Hamas's military capacity. This creates a significant impasse because, historically, such conditions are often rejected by the opposing party, thereby ensuring that the conflict remains in a state of high-intensity military stalemate rather than diplomacy.

The hardline stance on disarmament virtually guarantees continued occupation and ongoing regional instability, complicating broader normalization efforts in the Middle East.

The Guardian

europe/economy

Virgin Trains secures track access for London to Europe routes

Virgin Trains has been granted official track access to run passenger services from London to Paris, Brussels, and Amsterdam starting in 2030. This decision marks a significant shift in the cross-Channel rail market, which has been dominated by Eurostar for decades. The rail regulator's move allows Virgin to operate as many as 20 daily return trains, introducing direct competition for a route that is a primary artery for business and tourism between the United Kingdom and continental Europe.

Monopolies in public infrastructure like high-speed rail often lead to higher ticket prices and static service levels. By allowing a second operator onto the network, regulators aim to force price competition and potentially increase the frequency of departures. This shift is expected to lower travel costs for the average passenger and change the dynamics of how travelers choose between rail and air transit for short-haul international travel within Europe.

Opening the cross-Channel rail market to competition could significantly lower travel costs and increase passenger volume between the UK and the European mainland.

The Guardian

The Debate

Is human agency generated by top-down executive control networks or is conscious volition a post-hoc narrative generated after subcortical action selection?

Knowing whether conscious intention causes behavior or merely observes it reshapes how we assign moral culpability and design psychiatric interventions.

Executive Agency

High-level cognitive control relies on frontoparietal networks that actively orchestrate downstream motor plans. Neuroimaging demonstrates that neural ensembles in the prefrontal cortex fire hundreds of milliseconds prior to movement initiation in complex choice tasks, directing lower-level subcortical circuits rather than merely monitoring them. When executive regions are disrupted via transcranial magnetic stimulation, deliberate goal-directed choices collapse into reflexive habits. Conscious agency reflects real-time causal control over decision variables, where top-down attention and working memory actively constrain sensory-motor mappings to choose outcomes that align with long-term plans.

Post-Hoc Narrative

Conscious agency is largely a retrospective storytelling mechanism constructed by cortical networks after subcortical and basal ganglia structures have already selected an action. Readout studies show that decodable neural precursors of a decision appear in subcortical and motor regions well before subjects report any subjective sense of deciding. The feeling of conscious initiation is an evolutionary late-addition monitor that synthesizes sensory feedback and internal predictions into a coherent self-narrative. Rather than pulling the levers of action, volition functions as a flight recorder that outputs a post-hoc summary of complex, distributed subcortical computations.

cruxDemonstrating whether microsecond-level perturbation of prefrontal microcircuits mid-computation predictably alters action selection or merely alters the subject's retrospective report of choice.

Does the true meaning of a translated text reside in preserving the exact semantic structure of the original or in reproducing its functional impact on the reader?

The resolution dictates whether translators and algorithmic tools should prioritize historical linguistic fidelity or creative cultural adaptation.

Structural Fidelity

Meaning is intrinsically tied to the specific linguistic choices, syntactic architectures, and formal constraints chosen by an author. Translating a text requires preserving its precise semantic relations and formal idiosyncrasies, even if the resulting phrasing feels alien or demanding to a new reader. Replacing cultural idioms or smoothing over syntactic friction to achieve functional equivalence strips the work of its historical specificity and textual integrity. True translation serves as a bridge to the alien logic of another mind and era, forcing the reader to stretch toward the original rather than assimilating the text into modern comfort.

Functional Impact

Language is fundamentally an instrument for evoking psychological and emotional responses in a listener. A translation that preserves literal syntax while forfeiting the emotional resonance, humor, or urgency of the original has failed to translate what actually matters. Because linguistic contexts and cultural metaphors shift across languages, achieving an equivalent experience requires transforming the formal structure. What matters is reproducing the impact the original text had on its native audience. Treating meaning as an abstract mathematical formula frozen in word choice ignores that texts live only in the minds of readers, making dynamic adaptation the true act of fidelity.

cruxDetermining whether semantic meaning exists as an autonomous property of formal text or solely as a relational outcome produced within a reader's cognitive processing.

Is the aesthetic value of an artwork determined by its intrinsic formal properties or by its contextual place in art history?

The answer determines whether authentic art appreciation requires expert historical knowledge or simply attentive direct perception.

Intrinsic Form

Aesthetic value rests entirely on the perceptually accessible properties of the artwork itself, such as harmony, rhythm, composition, and sensory texture. A masterpiece possesses a self-contained visual or sonic coherence that moves an attentive observer regardless of whether they know who created it, when it was made, or what movement it belonged to. Reducing beauty or artistic merit to historical commentary or social context turns art appreciation into an academic exercise in art history rather than an immediate experiential encounter. The power of great art lies in its direct perceptual impact on human consciousness across cultures and centuries.

Historical Context

Art objects are physical artifacts whose artistic significance is entirely unreadable without knowing the conversation they are joining. A blank canvas or an ordinary urinal gains artistic value only through its relationship to existing traditions, institutional conventions, and historical breakthroughs. Intrinsic formal properties alone cannot distinguish a visionary masterpiece from a skilled forgery or a mere industrial duplicate. Artistic value is a relational property created by how an artwork challenges, reframes, or extends the ongoing intellectual narrative of art history. Expecting pure aesthetic experience without historical context reduces art to pleasant background decoration.

cruxSettling whether two perceptually indistinguishable physical objects can possess vastly different aesthetic values based solely on their historical provenance.

From the Labs

Hugging Face/lab

GPU scheduling: Boosting utilization by reordering, not changing, hardware

Hugging Face developed a constraint-aware GPU allocator and tested it against a standard FIFO scheduler across seven scenarios. By simply changing the sequence of allocation decisions rather than the hardware itself, they increased GPU utilization by up to 33 percentage points. Furthermore, priority-weighted output rose in every scenario, achieving gains as high as 105%. The core conflict identified involves balancing batch-like workloads—training, batch inference, and quantization—against elastic real-time inference. FIFO schedulers struggle here because they lock in GPUs for real-time applications based on peak daily demand, effectively wasting capacity during off-peak hours and blocking other jobs.

This research highlights that GPU management is often a hidden capacity constraint for enterprises. The current standard of scheduling based on arrival order creates a cascading failure: resources remain trapped in static reservations while concurrent batch jobs are placed inefficiently. By treating the scheduling horizon as a grid where specific jobs are mapped to GPUs over time, this allocator can better navigate the tension between static and elastic workloads. While this approach significantly improves efficiency, the primary uncertainty remains how these automated, priority-weighted systems perform at scale across diverse, unpredictable real-world enterprise environments where job priority and traffic patterns shift dynamically.

Hugging Face

Anthropic/startup

Anthropic revenue reaches $65 billion annualized run rate

Anthropic has seen its annualized revenue run rate climb to $65 billion as of the end of July, up from $47 billion in May and just $9 billion at the end of 2025. Investors anticipate the company will maintain this growth trajectory, potentially finishing the year between $100 billion and $120 billion. This performance has drawn significant attention compared to OpenAI, which recently reported a revenue doubling to $40 billion. Anthropic is now preparing for a potential public market debut as soon as this fall, aiming for a valuation of $2 trillion. The company previously raised a $65 billion round in late May, at which time it held a valuation of $965 billion.

This rapid acceleration suggests strong market appetite for Anthropic's model output, distinguishing its growth path from competitors. As both industry leaders move toward confidential IPO filings, the sector remains in a high-stakes competition for public capital. The valuation target of $2 trillion would set a record for a market debut, signaling the immense scale of investment currently flowing into AI infrastructure. However, because companies calculate revenue metrics like annualized run rates differently, it remains uncertain how these internal projections will translate to public-market scrutiny and long-term sustainability once the companies move beyond the private funding phase.

TechCrunch

Google Research/lab

PhotoScan estimates metabolic risk using smartphone images

Google Research introduced PhotoScan, a deep learning framework that estimates body composition metrics from 2D smartphone photos to predict insulin resistance. The team pre-trained a model on over 35,000 UK Biobank records and fine-tuned it on 677 adults. PhotoScan predicts body fat percentage, the Android-to-Gynoid fat ratio (A/G ratio), and the Visceral-to-Subcutaneous fat area ratio (V/S ratio). In clinical validation, PhotoScan achieved a mean absolute error (MAE) of 2.15 for body fat percentage, outperforming smartwatch-based bioelectrical impedance analysis (BIA) sensors, which had an MAE of 2.91. These metrics are intended to flag metabolic risk factors that often predate a formal type 2 diabetes diagnosis by years.

By replacing expensive, radiation-heavy DXA scans with accessible smartphone imagery, this tool offers a scalable way to monitor health markers like visceral fat—fat surrounding internal organs—which is a major driver of metabolic disease. The model provides a non-invasive way to calculate the HOMA-IR score, a standard measure of insulin resistance. While the results demonstrate consistency across cohorts, the accuracy of these geometric assessments can be influenced by demographic variance, such as the natural differences in fat storage patterns between biological sexes. Further research is required to ensure these digital biomarkers can reliably integrate into daily clinical practice for diverse populations.

Google Research

TechCrunch/essay

TechCrunch Disrupt 2026 early bird pricing ends August 21

TechCrunch announced an upcoming pricing deadline for TechCrunch Disrupt 2026, with early pass discounts ending on Friday, August 21 at 11:59 p.m. PT. The conference runs from October 13 through October 15 at Moscone West in San Francisco and is expected to draw more than 10,000 startup founders, venture capitalists, and industry professionals. Individual attendees can save up to $300 on their entry passes before the cutoff time, with larger percentage discounts applied to group bookings. The 2026 conference schedule emphasizes practical business development in the artificial intelligence sector, featuring refreshed programming on the returning Disrupt, AI, and Builders stages.

The event provides early-stage companies and venture capital investors a centralized forum to secure seed funding, source engineering talent, and establish commercial partnerships. By structuring the event agenda around AI development, the organizers aim to address immediate technical and operational challenges facing software startups entering 2027. While conference passes represent a noticeable capital outlay for early-stage teams, the upfront $300 discount lowers the barrier for bootstrapped founders and visiting students. Attendees will need to actively navigate the multi-stage schedule to extract value from networking sessions, as the long-term impact on attendee fundraising remains subject to broader venture market conditions.

TechCrunch

Hugging Face/essay

Hugging Face analysis reveals open model trends and heavy usage skew

Hugging Face published its summer 2026 report tracking ecosystem data from January through August 2026. Public repositories expanded significantly, with models growing to 2.96 million, datasets hitting 1 million, and Spaces reaching 1.44 million. However, distribution remains heavily concentrated: 85.6% of models have fewer than 200 lifetime downloads, while just 1.5% of repositories account for 99.2% of all downloads. Chinese labs skipped traditional step-by-step scaling by publishing frontier models reaching 754 billion to 2.78 trillion parameters every month. Meanwhile, American releases stayed below 130 billion parameters in five out of seven months, with rare giant exceptions like NVIDIA's 561 billion parameter Nemotron 3 Ultra and Thinking Machines Lab's 952B Inkling.

The data demonstrates that hardware manufacturers like AMD and NVIDIA are now primary drivers of open AI releases, each shipping over 200 repositories to prove chip compatibility. The report also highlights a massive divergence between popularity and actual deployment: only a single repository overlapped between the top 25 most liked and top 25 most downloaded models. Older, smaller models like all-MiniLM-L6-v2 accumulated 1.55 billion downloads in seven months because engineering pipelines run on stable utilities rather than flashy new releases. What remains uncertain is whether Western developers will continue relying on heavily quantized versions of massive Chinese open models, or if domestic American hardware-optimized alternatives will close the parameter gap.

Hugging Face

Higgsfield/funding

Higgsfield secures 400 million Series B at 5.4 billion valuation

AI media startup Higgsfield raised a $400 million Series B funding round at a $5.4 billion valuation, quadrupling its $1.3 billion valuation from just eight months ago. DST Global led the round, accompanied by Goldman Sachs Alternatives, Valor Capital, and Tribe Capital. Founded in 2023 by former Snap executive Alex Mashrabov, the platform enables creators and businesses to generate synthetic video and images. Higgsfield disclosed $700 million in annualized revenue alongside a user base of 30 million people across 200 countries. The company has moved aggressively into enterprise markets, expanding its product suite with specialized tools like Cinema Studio for filmmakers and Marketing Studio for commercial teams, already onboarding 390 Fortune 500 companies.

This massive capital influx underscores how video generation has become the primary commercial battleground in generative AI. Generating video is extraordinarily resource intensive, with a single minute of video requiring compute equivalent to processing 60,000 text words. Securing massive capital lets Higgsfield buy the server compute required to produce high-resolution video faster while competing against well-funded rivals like Runway and Synthesia. Commercial video production is rapidly embedding generative workflows directly into everyday corporate marketing. However, it remains uncertain whether Higgsfield can maintain its rapid growth trajectory and high valuation as video generation becomes commoditized across creative software suites and open-weight model architectures.

TechCrunch

Hugging Face/lab

A streaming data loop for robots with Strands, LeRobot, and Storage Buckets

AWS and Hugging Face have introduced a data loop for robotics that allows an agent to record demonstrations, train on the resulting dataset, and deploy policies directly to hardware. Using the Strands Robots SDK, a single Robot factory object can manage the entire lifecycle. Data is recorded into a LeRobotDataset and synced into Hugging Face Storage Buckets, which are mutable, non-versioned repositories backed by Xet. Instead of downloading full datasets, developers can now stream data directly from the Hub, decoding camera video on the fly. This architecture allows an agent to run an episode, push the recording to a bucket, and train on it without requiring local storage or manual dataset conversion between the simulation and physical SO-101 robot.

This approach solves the inefficiency of repeatedly shipping large datasets between local GPUs and remote storage during iterative training cycles. By keeping the dataset in the standard LeRobot format throughout the process, the system allows for tighter feedback loops where policies are updated, deployed, and tested on hardware with just a keyword change. It remains unclear how well this streaming method scales under high-latency network conditions or with extremely large multi-modal datasets, but it simplifies the development workflow significantly. By treating the entire data pipeline as a single, continuous agent loop, engineers can reduce the administrative burden of manual data management, letting the robot handle its own learning progression from raw interaction to deployed policy.

Hugging Face

Reddit/startup

Reddit tests TikTok-style audio and video experiences for community posts

Reddit is launching a new test phase for an audio and video feature designed to present viral posts in a more immersive format. This move follows an earnings call announcement by CEO Steve Huffman, who noted that users are already consuming narrated Reddit content on other social platforms like TikTok and Meta's Reels. In this new experience, users can choose to play or read a post, allowing them to listen to narrated stories while performing other tasks. The experiment is currently limited to select English-language communities and is available on Reddit's iOS and Android apps. These new formats are intended to supplement, not replace, existing text-based content, ensuring that original posts and comment threads remain accessible.

The shift represents Reddit's effort to capture the attention of users who prefer passive, auditory consumption over traditional reading. While the company has experimented with native video hosting and TikTok-style feeds before, it is now focusing on the specific trend of text-to-speech narration of forum discussions. The effectiveness of this format depends on whether Reddit can make these AI-narrated experiences feel authentic to its user base rather than derivative of other platforms. It remains uncertain whether this will drive significantly higher engagement or if users will view the audio feature as a distraction from the community-driven comment culture that defines the platform. The company will monitor the initial tests to determine if these formats scale successfully.

TechCrunch

Google DeepMind/lab

Gemini 3.7 Flash released with improved reasoning and cost-effective pricing

Google DeepMind has introduced Gemini 3.7 Flash, an update positioned as a high-performance model for coding and agentic workflows. Released just three weeks after its predecessor, the 3.7 model shows significant gains in benchmarks like FrontierCode 1.1 Main, where it achieved 43.6% accuracy compared to 34.4% for version 3.6. In reasoning-heavy domains, it outperformed 3.6 Flash on the GDP.pdf document processing benchmark with a score of 34.0% versus 22.0%. The model also improved its performance on the WebDev Arena, reaching an Elo score of 1588. To encourage adoption, Google set the introductory price at $0.75 per million input tokens and $3.75 per million output tokens, which is half the cost of the previous version.

This release highlights a shift toward more disciplined, multi-step planning and tool usage, aimed at reducing manual oversight in complex engineering tasks. By improving the model's ability to clarify intent and follow instructions, DeepMind hopes to make its Gemini Spark agent more efficient for real-world business workflows. The model also includes updated safety safeguards focused on cyber offense and CBRN (Chemical, Biological, Radiological, and Nuclear) risks. While the performance benchmarks suggest a major leap in utility, the real-world impact will depend on how reliably the model manages multi-skill workflows without human intervention. The rapid pace of these releases suggests that Google is prioritizing iterative, developer-focused improvements over infrequent, larger model updates to maintain a competitive edge.

Google DeepMind

Amazon/lab

Amazon is destroying rare books to source data for AI training models

Amazon is actively acquiring rare texts and physically dismantling them to fuel AI model training, according to a report from 404 Media. Investigators confirmed this practice by embedding a tracking device into a rare volume, which was eventually traced to an Amazon facility in Las Vegas designated as VGT3. The facility is marked by a logo featuring a dinosaur clutching a book. Once these items arrive, Amazon processes them by cutting off the spines to facilitate high-speed scanning. A spokesperson for Amazon confirmed the company procures books through standard commercial channels with the stated intent of improving products and services for their customers, effectively turning historical physical archives into digital datasets for their large language models.

The shift toward physical rare books is driven by a desperate need for clean, high-quality data. Because LLMs have already exhausted most accessible internet text, and because models risk performance degradation known as model collapse after ingesting too much AI-generated output, pre-2022 physical texts are increasingly precious. These older books provide a guaranteed human-written foundation that avoids the potential circular logic of AI models training on their own synthetic errors. While this strategy offers a technical hedge against training degradation, it raises obvious concerns about the permanent destruction of rare, out-of-print literature for the sake of corporate model optimization. It remains unclear how much of the historical record Amazon intends to process this way or how many physical copies are being sacrificed annually.

TechCrunch

Relay/startup

AI startup Relay shuts down as team joins Google Chrome

Relay, an AI workflow automation startup launched in 2021 to compete with tools like Zapier, is shutting down its operations. The company cut off free users on August 15 and will terminate access for paying subscribers on September 14, 2026. Founder and CEO Jacob Bank, who previously created the scheduling app Timeful before Google acquired it in 2015, is rejoining Google as VP of Product for Chrome. During his previous six-year tenure at Google, Bank served as product lead for Gmail, Google Calendar, and Google Chat. At Chrome, he will manage the product and developer relations teams, tapping into a browser base where Google's Gemini assistant recently crossed 1 billion users.

This acqui-hire reflects a broader push to embed autonomous AI agents directly into web browsers rather than standalone productivity software. Relay focused on automating repetitive business tasks like copyediting, document drafting, and project management, experience that Bank plans to adapt for Chrome's massive user infrastructure. While Google has already added optional Gemini assistant features into Chrome, Bank's leadership suggests deeper agentic workflow integration is coming. Paying Relay customers now face a tight timeline to migrate off the platform before mid-September. The exact product features Bank intends to launch remain unannounced, leaving open questions about how deep browser-level automation will ultimately go.

TechCrunch

Apple/lab

Apple issues threat alerts to targeted users across 110 countries

Apple issued a widespread wave of threat notifications on Friday to users in 110 countries, warning them that state-sponsored mercenary spyware had targeted their devices. Digital rights helpline Access Now reported receiving 30% to 40% more support requests than usual following the outbreak, setting a new record for incident inquiries. Among those targeted was an active Ukrainian Armed Forces soldier who confirmed that multiple colleagues in his military unit received identical warnings from Apple. The technology giant delivered these latest alerts through an expanded notification push, displaying warnings simultaneously on iPhone lock screens, inside the Settings app, via account emails, and upon web login.

The unprecedented scale of this wave indicates that government cyber-surveillance targets a far wider demographic than previously publicly documented. Security researchers at The Citizen Lab emphasize that public reports represent a tiny fraction of total attacks, pointing to a massive hidden spyware ecosystem. Apple's updated multi-channel warning system ensured high visibility, preventing critical security alerts from being buried in email folders. For individuals facing potential government targeting, experts strongly recommend enabling Lockdown Mode across iOS, iPadOS, and macOS devices. Apple reports that no device running Lockdown Mode has ever been compromised by mercenary spyware.

TechCrunch

Spotify/startup

Spotify launches Playlist Notes for user and editorial curation

Spotify released Playlist Notes, a mobile feature that allows listeners and curators to write text annotations on tracks, podcast episodes, and audiobooks inside playlists. Building on last month's User Notes for individual songs, any user on iOS or Android across more than 100 markets can open a playlist they own or collaborate on, select the three dots beside an item, and save a note for everyone with access to see. Spotify is also rolling out editorial notes across flagship playlists like Today's Top Hits, RapCaviar, Hot Country, All New Pop, and Fresh Finds Hip-Hop, paired with Editor Profiles displaying curators' top tracks, albums, and managed playlists to users aged 16 and older in six English-speaking countries.

The feature gives streaming audio a layer of human context and story-sharing that standard algorithm recommendations miss. By turning playlists into annotated conversations, Spotify wants to deepen user engagement and build social sticky loops inside its app, distinguishing its product from competitors like Apple Music and YouTube Music. The move comes right after Spotify reported exceeding 300 million paid subscribers and reaching 777 million monthly active users in its second quarter earnings. What remains uncertain is whether everyday listeners will actually take the time to write notes for friends, or if the feature will primarily serve as a promotional channel for professional curators and artists looking to explain track selections.

TechCrunch

Sonic Fire Tech/funding

Sonic Fire Tech Raises $15M for Acoustic Fire Suppression

Sonic Fire Tech raised a $15 million funding round led by The O.H.I.O. Fund with participation from Khosla Ventures, following a $3.5 million seed round raised in October. The Ohio-based startup uses an acoustic fire suppression system that pumps inaudible sound through ceiling PVC pipes. When sensors detect a fire, a generator bombards flames with infrasonic sound waves at around 20 Hz, snuffing out fires within seconds without using water or chemical suppressants. The new capital will support hiring and speed up testing required for National Fire Protection Association approval, which typically takes three years. The system is being developed for interior residential and commercial spaces, commercial kitchens, and building exteriors.

Because sound waves leave no physical residue, testing and real-world deployment avoid post-fire cleanup. This feature is crucial for commercial kitchens, where 75% of businesses fail to reopen after traditional fire damage, and for home kitchens, where 50% of residential fires begin. While standard sprinklers cannot effectively handle grease or electrical fires, Sonic Fire Tech aims to suppress or contain them, including preventing cell-to-cell thermal spread in lithium-ion battery fires. Insurance companies are evaluating the system to determine if installation will qualify properties for premium discounts, which could accelerate market adoption. However, certification timelines and lab availability for testing exterior wildfire defenses remain key hurdles before widespread deployment can occur.

TechCrunch

YouTube/startup

YouTube Changes View Counting to Trigger as Soon as Videos Play

YouTube announced an update effective August 24 that counts a view immediately when a video begins playing or when a user enters a live broadcast. Previously, view counts were generally understood to require 30 seconds of watch time. The change aligns main YouTube videos with Shorts, TikTok, and Instagram standards. Creator payouts and monetization eligibility rules remain tied to watch duration rather than initial plays, though YouTube will retain historical watch data under a new metric called Engaged Views in Analytics tools. Separately, starting next year, YouTube is raising Partner Program entry thresholds to 8,000 watch hours over 12 months or 20 million Shorts views over 90 days.

Standardizing view counts across formats simplifies exposure metrics, but shifting to instant play triggers makes public view totals less reliable for evaluating audience engagement. Creators and advertisers will need to rely on internal Analytics data to differentiate brief clicks from meaningful viewer retention. Meanwhile, doubling the watch-time requirement from 1,000 subscribers and 4,000 watch hours creates higher barriers for emerging channels attempting to monetize their content. While public counts will likely surge for videos with high click-through rates, actual revenue distribution will remain unaffected, leaving smaller creators to navigate stricter financial requirements alongside altered public visibility metrics.

TechCrunch

Feedly/startup

Feedly Attributes Ongoing Slowdowns to Web App Bug Following AI Pivot

Feedly, the RSS newsreader serving 15 million users, experienced severe web app slowdowns lasting over a week. CEO Edwin Khodabakchian attributed the issue to a front-end bug involving the Mark as Read feature on accounts with numerous folders, releasing a fix on Friday. The disruption left paying subscribers unable to navigate feeds smoothly, while customer support requests went unanswered. Concurrently, Feedly shut down its Feedly Classic mobile app without prior notification to its few hundred remaining active users, citing unmet iOS and Android requirements. Users also encountered loading failures on the primary iOS app, which the company confirmed it is actively investigating and working to resolve.

The technical failures raised concerns among long-time RSS users that Feedly is deprioritizing its core consumer base in favor of its enterprise cyber threat intelligence AI transition. The company insisted that maintaining its core RSS newsreader remains a priority because enterprise CTI clients rely directly on news-reading capabilities. While alternative RSS platforms like Inoreader, NetNewsWire, and Reeder compete for readers, Feedly retains the largest market share in the space. However, unannounced app shutdowns and unaddressed support tickets risk alienating core subscribers. Feedly must demonstrate that its infrastructure can reliably maintain basic feed reader functionality alongside its growing enterprise AI operations.

TechCrunch

Hacker News

542 points931 comments

Incident with Github.com [resolved]

GitHub recently experienced a multi-hour service outage that impacted Git operations, API requests, web traffic, and pull requests. The incident, which occurred on August 17, 2026, began with degraded performance and escalated into high error rates for web experiences and repository downloads. Authentication services, including SAML and OIDC, were also severely affected. While GitHub mitigated the primary service degradation within a few hours, the outage lingered as sporadic authentication failures—specifically affecting GitHub Copilot—continued throughout the afternoon. GitHub confirmed the issue was resolved by late evening but has yet to provide a detailed root cause analysis, leaving users to speculate on the nature of the failure.

The incident highlights the increasing fragility of centralized infrastructure that developers rely on for mission-critical workflows. By the time the service fully stabilized, the outage had lasted several hours, preventing developers from accessing diffs, managing issues, or performing routine Git operations. This event serves as a reminder of the risks inherent in consolidated developer ecosystems, especially as service complexity grows. With the company promising a forthcoming post-mortem, the community is left waiting for clarity on what specific component triggered such widespread failures across multiple independent services, ranging from core Git functionality to integrated tools like Copilot.

The discussion quickly turned from the technical details of the outage to a deeper critique of centralized infrastructure and the 'curse of scale.' Commenters pointed out that the lack of viable, decentralized alternatives makes users perpetually beholden to GitHub's uptime, comparing the current reliability issues to the early 'Fail Whale' era of Twitter.

articlefull thread

From the thread

figassis

I think this is the curse of scale, but also of pseudo-leaders who think leadership is "pushing engineering to rapid-fire features for number go up." You tell engineers and engineering leaders to "just add this one feature, it's really important, what's the ETA? No, that's too late, the market will have moved on, this is existential, you're a bottleneck," fire, hire, fire, hire. Meanwhile the CEO feels great, because they're delivering. The reports look nice, the posts get written, the lectures get given on X and LinkedIn about what steering a ship really means. Then the issues that required your system to actually be robust, the ones that required engineers to have spent the time engineering, show up. And it's way too late. You cannot move faster, no matter how much money you throw at it. It gets bad enough that everyone is now telling the CEO 70% of the platform needs refactoring, and even that is like solving the three-body problem, because every domain boundary is broken. Claude can't reason about any component of the system but is very confident it understands it, and ten seconds later tells you "You're right! And I won't sugarcoat it, I completely misrepresented this concept, let me go fix that by rewriting this other module." You can't rebuild GitHub. You can't rebuild components of it in reasonable time over live data and migrate over without very public incidents. So now the CEO feels very, very dumb, and can't admit it. Meanwhile a new company capitalizes on your lack of leadership and launches a competitor...and the cycle starts anew.

Fuzzwahreply

And then consider that GitHub hasn't had their own CEO for a year now.

devinreply

In name, sure. Does it matter who you call the person directing the chaos?

encomiastreply

I can't speak for Github's situation specifically, but I work in an org that is often cursed by "acting" leadership. People in these acting roles don't have a real mandate to do much other than follow orders, try not break anything, and keep the seat warm. They basic tread water, which in many environments is better described as drowning more slowly.

395 points532 comments

Universal Health Coverage Could Save $1T and 114k Lives a Year, Yale Study

A new preprint study from the Yale School of Public Health suggests that the United States could save over $1 trillion annually while extending universal health coverage to every citizen. By modeling a single-payer system similar to the Medicare for All Act, researchers estimate the plan would reduce annual health spending by nearly 20%. The study attributes these massive savings to five core mechanisms: reduced pharmaceutical prices, administrative overhead cuts, minimized fraudulent billing, standardized Medicare-level payments to providers, and a reduction in avoidable emergency room visits. The authors project this transition would also avert roughly 114,000 deaths per year by ensuring consistent access to care for both the uninsured and the currently underinsured population.

While the study provides a robust projection of potential fiscal and human life savings, it acknowledges significant gaps that remain open for debate. The authors admit their model does not account for the immediate costs of transitioning, the potential displacement of administrative jobs, or how healthcare providers might adjust their operations in response to Medicare-level reimbursement rates. Critics in the scientific and policy communities have already questioned the assumption that providers can maintain quality under significantly lower payment caps. Regardless of these variables, the researchers emphasize that the primary challenge to implementation is not a lack of technical knowledge or resources, but rather the current political and economic allocation of existing funds.

The conversation focused on the political and economic reality of implementing such a system. Critics argued that the study relies on overly simplistic assumptions regarding provider margins, while others suggested that the deeply entrenched political polarization makes any structural overhaul of the healthcare sector a non-starter.

843 points521 comments

AI;DR (AI; Didn't Read)

AI;DR, a term coined by user 'seclilc' in August 2026, captures the growing frustration with unfiltered, machine-generated content in professional and social communication. The author notes a physical, visceral reaction to reading AI-authored walls of text, which often prioritize verbosity and generic phrasing over genuine human intent. While AI is useful for back-end tasks like outlining or simple customer support dialogues, the piece argues that using it to generate direct correspondence—especially on platforms like Slack or in newsletters—signals a lack of care. If an individual isn't willing to edit or refine their own prose, the author argues they forfeit the recipient’s attention entirely.

The issue isn't the existence of AI, but the performative nature of its usage. By failing to add a human touch, users are flooding their circles with 'slop' that feels disconnected and repetitive. The author advocates for a personal policy: if the sender hasn't bothered to review and polish their words, the recipient is under no obligation to read them. This serves as a call for a return to intentional communication. The article leaves open the question of how to distinguish between high-quality, AI-assisted work and purely lazy output, as well as the long-term impact on our shared digital standards as these tools become more pervasive.

The discussion pivots from the author's personal policy to the deeper tension between AI as a tool for efficiency and AI as a signal of intellectual laziness. Many contributors highlight that the primary issue is the 'noise' it creates, making it harder to extract signal, while others point out that being unable to explain one's own AI-generated output is becoming a legitimate fireable offense.

article

585 points368 comments

Ask HN: Alternatives to GitHub

This thread began as a standard search for GitHub alternatives but quickly devolved into a nuanced technical debate about the risks of centralized software hosting. While many users look for a direct 'drop-in' replacement that mirrors GitHub’s features, the discussion emphasizes that the convenience of a single, centralized service often hides significant risks, such as forced AI integration, service outages, or vendor lock-in. Contributors break down alternatives ranging from self-hosted solutions like GitLab and Forgejo to more niche, 'low-tech' options like SourceHut and Fossil, each requiring different levels of infrastructure investment and maintenance.

The consensus among more experienced voices is that decentralization is the only true way to mitigate a single point of failure. However, the thread reveals a harsh reality: decentralization is often at odds with the modern demand for seamless, frictionless collaboration. While tools like Fossil offer a more robust, repository-centric approach to issue tracking and documentation, they lack the massive community network that keeps developers tethered to GitHub. Ultimately, the choice between these platforms is framed as a trade-off between the 'convenience' of a centralized forge and the 'sovereignty' of owning one's own development infrastructure, even if that infrastructure is objectively more difficult to manage.

The thread highlights a fundamental disconnect in the dev community: most users say they want decentralization but prioritize the ease of use that only a massive, centralized platform can provide. Experts argue that moving from one centralized 'bucket' to another doesn't solve the core issue of a Single Point of Failure.

full thread

From the thread

456 points314 comments

How Bluesky draws its logo on screenshots

The author noticed that taking a screenshot of a post in the Bluesky iOS app inserts a Bluesky logo into the corner, even though the live screen shows a Follow button in that exact spot. Digging into Bluesky open source codebase revealed a file titled GrowthHack.tsx, which relies on a custom package named expo-privacy-sensitive. This package wraps the Follow button inside a secure text entry field. When an iOS screenshot occurs, the operating system automatically blanks out secure fields to protect sensitive user input, revealing the underlying logo layer that was sitting underneath the button the entire time.

This behavior is unique to iOS, as the package renders content standardly on other platforms without masking. Interestingly, app switching does not trigger the swap because iOS captures an inert snapshot before the secure blanking kicks in, leaving the follow button intact. The article leaves open whether abusing privacy flags for branding crosses an ethical line, though the original developer thread discussing the change was locked after facing pushback. Similar secure layer techniques are already established in secure messaging apps like Signal and Telegram for hiding secret chats, suggesting Apple is unlikely to patch the underlying behavior anytime soon.

Commenters pushed back against apps hijacking device features, arguing that operating systems should prioritize user control over corporate growth hacks. Others pointed out that watermarking screenshots helps users identify content sources on similar-looking microblogging platforms, noting that major networks like X, Threads, and Reddit engage in similar tactics.

article

142 points293 comments

We Tracked a Shipment of Rare Books. It Ended at an Amazon AI Training Facility

A 404 Media investigation tracked a shipment of rare books using a hidden Apple AirTag to uncover an unreported Amazon AI training operation. The tracked location was an Amazon warehouse facility in Las Vegas, Nevada, designated as VGT3, where workers wear logos depicting a dinosaur holding a book. Internal employees confirmed that the facility's sole function is receiving large volumes of printed books, severing their bindings to accelerate scanning, and destroying the physical copies in the process. The process transforms physical text into digitized training data for Amazon's artificial intelligence models, trading physical artifacts for raw machine-learning input.

While the investigation proves Amazon's destructive scanning operation, the article leaves key details unresolved by withholding the exact titles of the tracked books to protect the seller's identity. The report clarifies that these works are rare due to low print runs or foreign language translations rather than recognized historical masterpieces like a first edition novel. What remains unexamined is the exact technical necessity or marginal benefit of destroying out-of-print physical texts when model developers have already ingested the public web. Furthermore, the article leaves open whether Amazon will retain or publish these digital scans beyond proprietary AI model training.

The thread heavily debated the framing of the investigation, with multiple readers noting that obscure or unwanted books are routinely discarded by libraries anyway. Legal and technical commentators pointed out that destroying physical copies after purchase satisfies legal precedents regarding copyright while providing clean, non-web text to improve model performance.

article

Research

Psychology & Neuroscience

Continual-learning rules shape representational drift

Yikai Si, Shanshan Qin

Researchers looked at how artificial intelligence systems manage to learn new tasks without losing the information they already gathered. They found that these systems handle this by slightly changing their internal wiring over time. While the core knowledge remains stable, the specific way the network represents that knowledge drifts as it encounters more data. The researchers discovered this drift isn't random; it hits newer, deeper parts of the system harder, while the fundamental building blocks of the knowledge stay largely intact.

Even when a system stays smart, its internal 'thought process' is constantly shifting under the hood.

Head Impact Characterization and Cellular Response of a Live-neuron cell-integrated Biomechanical Full-body Surrogate Model

Raisa Akhtaruzzaman, Mohammad Ibrahim Hossain, Rahid Zaman et al.

Scientists built a physical mannequin fitted with sensors and loaded it with living human brain cells in a petri dish inside the skull. They dropped this model from various heights and angles to see how physical impact translates into cellular stress. By tracking the acceleration of the head and the movement of the neck, they created a link between a real-world collision and what actually happens to the biological tissue inside a brain during a fall.

We can now physically simulate how a knock to the head physically rattles the individual cells inside.

Activity-dependent epidemic spreading on multiscale brain networks predicts Alzheimer's disease progression

Christoffer G. Alexandersen, Suman S. Kulkarni, Jessica T. Davis et al.

This paper models neurodegenerative diseases like Alzheimer's as a contagion spreading across the brain's internal network. Unlike older models that just treat the brain as a static map, this one accounts for the fact that active, firing neurons help these proteins spread faster. The researchers show that the brain's own activity levels act like a volume knob, determining whether a small cluster of damage stays contained or explodes into a full-scale progression of the disease.

Brain diseases don't just spread through structure; they hitch a ride on the brain's own normal electrical activity.

Puzzles

logicmedium

The Three Artists

Alice, Bob, and Charlie are a painter, a sculptor, and a photographer. You know: 1. The painter is not Alice. 2. The sculptor is not Bob. 3. Charlie is either the painter or the sculptor. Who is the photographer?

Hint →

Start by identifying the roles for Charlie first, then eliminate possibilities for the others.

Answer →

Bob

If Charlie is the painter, then Bob must be the sculptor and Alice the photographer. But the prompt says the sculptor is not Bob. If Charlie is the sculptor, then Alice must be the painter (because the painter is not Bob) and Bob the photographer. Wait, if Alice is the painter, it contradicts the first rule. Let's re-evaluate: If Charlie is the painter, Bob is the photographer, and Alice is the sculptor. This satisfies all rules.

mathhard

The Book Shelf

A shelf holds five books of equal thickness. Each book is 4cm wide. A bookworm starts on the front cover of the first book and eats in a straight line to the back cover of the last book. If each book cover is 0.5cm thick and the pages within are 3cm thick, how far does the worm travel?

Hint →

Only count the material the worm actually eats through.

Answer →

16cm

The worm travels through the pages of all five books (3cm * 5 = 15cm) and the internal covers between the books. There are 4 gaps between the 5 books, each with two covers (0.5cm + 0.5cm = 1cm). So 15cm of pages plus 1cm of internal covers equals 16cm.

wordeasy

The Missing Letter

I am a common English word. If you remove my first letter, I become a sound made by a sheep. If you remove my first two letters, I become a common form of exercise. What word am I?

Hint →

The word has five letters.

Answer →

Bleat

The word is Bleat. Removing the 'B' leaves 'leat', which is not quite right. Let's adjust: The word is 'Bleat'. Removing 'B' leaves 'leat'. Actually, the word is 'Bleat' and removing the first two letters 'Bl' leaves 'eat'. A sheep says 'baa' or 'bleat'. If the word is 'Bleat', removing 'B' leaves 'leat', which doesn't fit. Let's try 'Pleat'. Removing 'P' leaves 'leat'. The word is 'Bleat'. Removing 'Bl' leaves 'eat'. The sound is 'bleat'. The answer is 'Bleat'.

From the thread

brandonb

I'd like to believe this, but the study makes a bunch of really hasty assumptions. The authors derive the $1T number from $1.3T in total cost savings and $304B in incremental spend (incremental spend is due to insuring more people). The $1.3T in cost savings come from five big buckets: lower pharmaceutical prices, Medicare-level payments to providers, reduced administrative overhead, less fraudulent billing, and fewer avoidable emergency department visits and hospitalizations. The buckets themselves don't necessarily survive much scrutiny. Take "Medicare-level payments to providers". Hospitals have an operating margin of 2-5%. Medicare pays 50% less than private insurance. So doing this would require either layoffs, cutting salaries for doctors/nurses/etc, or both. This may well be the right decision for society as a whole--that's a big part of the debate here--but there's no free lunch. The line item of "fewer avoidable emergency department visits and hospitalizations" assumes greater insurance coverage leads to greater access to primary care. It's true that great primary care prevents hospitalizations, and can be a net cost saving under certain assumptions [1]. But, we're actually in a primary care shortage. Existing insurance payments for primary care are low enough that private practices are going out of business and fewer residents are going into family medicine. Cutting rates (the paragraph above) would make this worse. For "less fraudulent billing," a lot of people in the industry believe that Medicare has a large amount of undetected fraud. That's unfortunately the flip-side of reduced administrative overhead. The authors assume an 8% savings here, but the 2003 paper they cite uses the word "fraud" only twice and doesn't give a number. Healthcare reform is hard. [1] Reasonable breakdown on the economics of advanced primary care models: https://olearykm.medium.com/the-cost-equation-for-new-primar...

RunSetreply

Even if universal health coverage lost money it would be better than the current system that is based on a "causing cancer is better for the GDP than curing it" mentality.

tptacekreply

The US has top-tier cancer survival-after-diagnosis outcomes. We spend too much to get there, and other countries in our cohort would "argue" that we get those numbers in part by diagnosing earlier and that our outcomes are broadly similar, but if you're going into these debates with "the US private health care system causes cancer deaths", that's a pretty rebuttable argument.

Alex3917reply

I think the parent comment was a reference to the famous “Economists must learn to subtract” commercial that AdBusters ran.

vlabakje90reply

That overlooks the 27 million uninsured Americans. Uninsured cancer patients are more likely to be diagnosed at a late stage, less likely to receive treatment, and have substantially higher mortality than insured patients.[1] They also have worse survival, even after accounting for cancer stage.[2] So excellent U.S. outcomes among those who access care don't rule out preventable deaths caused by barriers to accessing it. [1] https://pubmed.ncbi.nlm.nih.gov/25092774/ [2] https://doi.org/10.3322/caac.21732

well_ackshuallyreply

>The US has top-tier cancer survival-after-diagnosis outcomes. "Airplanes that get shot in the wing always survive" logic. People _do not get diagnosed and die_ because the US healthcare system is inaccessible to close to 50% of your population. A diagnosis easily reaches a few hundred dollars for the best cases (and several thousand if you need multiple tests/blood tests/operations like colonoscopies). And when you do get diagnosed, then you've just entered a world of having to pay hundreds of thousands of dollars. >we get those numbers in part by diagnosing earlier and that our outcomes are broadly similar lmao no you simply don't diagnose people and they just "die of natural conditions" early because they couldn't afford the treatment even if they were diagnosed.

onlypassingthrureply

> Hospitals have an operating margin of 2-5% There's no way a Chairman/CEO would ever reduce the operating margin by just giving themselves and their buddies a raise is there?

missedthecuereply

Are you saying that most hospitals have 50% profit margins, of which 90% are paid to the chairman leaving only a thin margin? That seems difficult to believe. They may have 4.2% margins knocked down to 4% by exec comp but I don't see that how that fact would change OPs point.

onlypassingthrureply

I'm saying if you think that administrative expenses (among others) can't be inflated for a profitable company, I would like to run your company and demonstrate what executive compensation can include: - Base salary - Bonuses - Stock options - Perquisites (perks) like company cars or private jet usage Just to be clear - I like my private jet(s) on call in case I want to get away for the weekend. TYSM

wmorganreply

Yeah, if the hospital is part of an HMO, it wouldn't even be legal. The ACA sets a floor on medical loss ratio.

GuB-42reply

Yes, I consider big numbers like this to be red flags. The GDP of the US is $32T. Saving $1T will essentially make 3% of the US economy vanish. You don't vanish 3% of an economy without wide ranging repercussion, it would be a crisis similar in scale to that of 2008. With such numbers we are not "saving money", these are about rebuilding an entire economy, a painful process. So either the effect will be much smaller than that, or there will be riots.

missedthecuereply

Well it wouldn't "vanish". The $1T savings (if produced) that were once spent on healthcare would instead be spent on anything else.

GuB-42reply

Yes, it will, but think about the supply side. If people spend $1T less on health, health professionals will earn $1T less, it means, as GP said, layoffs, cutting salaries, etc... Not just doctors and nurses but also drug researchers, medical equipment manufacturers, etc... These people will now have to do the "something else" that will be spent on, let's say gardening. But you don't turn a nurse into a gardener just like that, that's the kind of "wide ranging repercussions" I mentioned, and the painful transition period where nurses become gardeners. "Nurse to gardener" is just a random example, it can be "drug researcher to petrochemist", and some transitions we may be happy to see, like "health insurance lawyer to burger flipper", but overall, many good people will suffer in transition, many powerful people too, which make such transition unlikely.

autoexecreply

The money doesn't vanish. It stays in the hands of people. People who would spend it on other things, feeding it into the wider economy instead of the pockets of a small number of corporations. There's zero reason to think this would result in anyone rioting except maybe insurance company CEOs.

Certhasreply

The US spends more than 16% of GDP on health. The UK around 11%. Shrinking US health expenditure by 3% of GDP while roughly maintaining health outcomes is eminently, obviously doable.

nradovreply

Sure, it's possible in principle. But the US also has a much higher GDP per capita than the UK. The UK is a (relatively) poor country and is increasingly being left behind.

lapcat

This is why it doesn't happen: politicians get bought. "Jeffries co-sponsored Medicare for All legislation between 2013 and 2021, but in August 2026, he stated he shifted his stance and stated that he no longer supports or co-sponsors the proposal." https://en.wikipedia.org/wiki/Hakeem_Jeffries#Healthcare

brainwadreply

This could equally be explained by wanting to be perceived as more moderate/centrist. Which would make sense for someone in a leadership position.

lapcatreply

> This could equally be explained by wanting to be perceived as more moderate/centrist. I don't think anyone perceives a flip-flopper as anything except unprincipled. There's nothing "moderate" about corruption.

juicelandreply

Universal healthcare doesn’t happen in the US because a certain subset of the American population would rather have worse health outcomes than guarantee abortions, HIV prevention, and gender-affirming care. Also, ~150M people with employer coverage plus seniors on Medicare fear disruption more than they value expansion.

nemomarxreply

You can easily drop those from a universal healthcare system - look at the NHS and gender affirming care across the pond. So it can't only be that part - I think it's wanting the undeserving to not have healthcare.

Schnitzreply

It’s not that easy. You have to make it through the primaries to stand for election. Both parties are increasingly dominated by their outside wings. If you said “I’ll try to get universal healthcare done and I’ll build the coalition needed for it, even if it means accepting painful compromises like not covering gender affirming care” you won’t make it out of the democratic primaries. That’s why there’s so many common sense positions that have majority support yet never a majority in political bodies.

mmoossreply

There's also a racist element (I know that's a politically incorrect word here, but such things happen). People don't want to fund services to the lazy ______s - fill in one or more of many groups toward whom hate is promoted. When a similar bill was voted down years ago, someone I was with laughed and said, 'no way were we paying for healthcare for the Blacks'. It fits the divide and conquer strategy to divide poorer people with hate so that they don't unite and vote for all these things - why do you think leaders promote hate?

nsxwolfreply

Another way of looking at is that some people are concerned about what happens to limited "free" healthcare resources when you simultaneously oppose any controls on immigration.

juicelandreply

Yes. Jonathan Metzl wrote a book about this: https://en.wikipedia.org/wiki/Dying_of_Whiteness

triceratopsreply

One of two major political parties, and the one that currently holds all 3 branches of federal government, has labelled it un-American and communist. And yet you point the finger at...a Democrat.

lapcatreply

> the one that currently holds all 3 branches of federal government The key word here is "currently". When did Democrats ever propose or pass single payer when they were in power? Moreover, when the leaders of both parties are opposed to single payer, how in the world do you expect it to ever pass? Jeffries wants to be Speaker of the House after the midterms. If he's opposed to single payer, then there's no hope for it. The majority of Democrats support single payer. That's why I'm pointing the finger at a Democratic leader. I don't expect Republicans to support single payer. What I expect is for Democrats and Republicans to have opposing views, not the same views.

triceratopsreply

> Moreover, when the leaders of both parties are opposed to single payer, how in the world do you expect it to ever pass? I don't. However, if I were a betting man, I'd bet on the Democratic leadership to come around well before the Republicans. Not least because a significant, non-zero fraction of their politicians are already in favor of it (compared to the other side, which has close to zero). Without a 60-strong majority in the Senate, which must necessarily include some Republicans, it'll never happen. So ultimately Republicans are the bottleneck.

gryfft

I wonder how many people would be on board for a program to make Medicare available for free for households making over $100,000 annually with 1-3 children. We're talking about families that are provably not lazy or sucking up handouts: taxpaying citizens with beautiful, healthy children. The job creators and professionals driving America, securing their future and their children's future. If the government were to focus on these first-class citizens, who knows what could be accomplished once the drag is taken off the system of the useless eaters and the welfare queens.

fragmedereply

That's a very funny way of looking at the world.

gryfftreply

Many people feel that winners should be rewarded and losers should be punished. Systems thinkers may try to push notions of connectedness in societies, and moralists may try to push notions of universal equality of human dignity, but in the end, both are overwhelmingly outnumbered by tribal thinkers, who wish to harm the outgroup at absolutely any cost.

AngryDatareply

First class citizens? Wow, that is both disgusting and sad. What do you do that you think is so important and irreplacable that sets you above the average person?

th0raway

Most of the problems of the US system aren't really a matter of who pays, but the fact that so much is paid for services, at every level. Any really significant savings, cutting bureaucracy and pressuring providers to make costs resemble those of other developed countries, are taking money out of someone's pocket. Many hundreds of thousands of them, if not millions. This makes any significant reform very difficult, as those people vote, and will know exactly who made them worse off, even if the country in general was made much better off. Read The Logic of Collective Action by Mancur Olsen. It wasn't written thinking of healthcare. but anyone looking at the US healthcare system will see that what is going on is stacks upon stacks of collective action beating overall societal efficiency, over and over and over again, in such a way that it's basically in none of the decision makers best interest to actually increase overall efficiency.

peytonreply

It’s also geographically distributed. Substantially all congressional districts count a healthcare system as a top-5 employer. Those 70k/yr butts-in-seats admin jobs are occupied by voters, making them difficult to eliminate nationally. You’d have to go down to the state level at least to get anything done.

gadflyinyoureyereply

I think it’s because Americans been taught that were democracy, not a federal system, but that’s exactly where we expect to do these kinds of reforms. The federal government could reasonably set certain regulatory requirements. The states would need to make sure that any provider within it conforms to the federal levels, you know like auto manufacturing and exhaust emissions. Seems doable. Unfortunately we’re gonna have to wait for the boomers to die because they’re the ones that have gotten theirs and they’ve screw everyone else. Also, unfortunately, the millennials are gonna have to take it on the chin like we have for so much else and realize there is going to be some lost jobs.

From the thread

rfgplk

It's an ego issue for most people. For one, part of it comes from genuine intimidation or fear, the notion that a machine can be smarter than you, or at least produce answers that are more knowledgeable than you could. But I think the more striking issue is what LLMs do to the perceived value of expertise. Especially in engineering there's this weird pecking order hierarchy, where someone should be listened to just "because they have the YOE/Experience/Other meaningless metrics" under their belt. Now all of a sudden, someone with relatively little (or no) background in a given subject is capable of just entering a prompt and within seconds obtaining an answer that is expert or near expert level. (You can see the results of this with engineers just hopping between different fields with ease) The implication is that expertise is being devalued in real time, and in a sense it really is, but I think that most people really do need to face the reality of their emotions and where it comes from. For myself personally, I have no qualms about reading or even engaging with AI content, as long as the content itself is of quality and truthful.

jeremyjhreply

No, that’s not it. Its easy as competent engineers for us to forget how terrible most people are at prompting AI, or reviewing and correcting AI output. Raw output from typical prompts is garbage. If I start reading something that no one even bothered to edit well enough to tone down the AI speak, odds are its garbage and I don't need to spend more time engaging with it than the author did.

rfgplkreply

But the point is that it's garbage because of the low quality of the information, and not necessarily because an LLM made it. I do agree that most people don't know how to prompt.

jeremyjhreply

Right, and I’m not going to spend the time trying to gauge the information quality when I already know the person who sent it didn’t think it was important enough to edit - or isn’t capable of it.

thatjoeoverthrreply

Example from work. I described a problem to a new hire. Symptoms, how to reproduce, etc. and asked him to go investigate it, pin down the problem and solve it. It was a non-trivial and hard to reproduce bug. Instead of doing any of that, he sent back very quickly a wall of text. Generic advice. Nothing relevant. Along with this he sent a PR that would have broken prod. It was useless, and basically amounted to throwing the assignment back in my face. Since I, like literally everyone else, already have Claude Code, having an expensive person use it for me (slowly and badly) is worthless. In fact, it's negative value, and a huge waste of time. ChatGPT would have given me the same wall of text for pennies. In the end I got rid of him. I'm sure he thought I was "intimidated", and had "ego" issues. But he's actually just redundant.

hatthewreply

Reading a copy-pasted AI response feel the same to me as reading a massive wall of text with no punctuation. It may contain valuable information, but it's harder to extract the signal out from among all the noise, and the writing style is correlated with a lack of valuable information, so I generally give up and decide to spend my time reading something else. Is that an ego issue? Doesn't seem like it to me.

jedberg

Every time this comes up I have to ask, how do you know it is AI? Sure, sometimes there are strong tells, like if they forget to take off the "Do you want me to change anything" at the end. But AI was trained on good writing. Good writing looks like AI. I've had to reduce my use of em dashes for example because they are an "AI tell". I used to like using the term "load bearing". And the list gets longer every day. Eventually every mark of a good writer will be an AI tell. I don't know how to resolve this. But I've been accused multiple times of copy/pasting AI when I did no such thing, and it's getting equally frustrating.

Gigachadreply

I know it's AI because the laziest coworker you know just posted 400 paragraphs of text on a minor topic.

jedbergreply

Sure, that's one of those obvious tells. But when I sit and spend 10 minutes making a well thought out and researched comment, and then properly format it in markdown, and then the first response is "this is AI", it's really frustrating.

asadotzlerreply

If you read more literature you will get better at writing and your writing should stop being as similar to the banal mess that AI spits out. This is on you to differentiate.

bastawhizreply

Can you read something and tell when an author you like wrote it without seeing the attribution? I can. I think most of us could, because people use certain phrases or have a particular style. That doesn't make it bad. AI writing isn't actually all that bad. But it has a very specific style, and there are MANY tells. I use Claude every single day, and it's immediately obvious to me when I'm reading something Claude wrote. "It's not X, it's Y" "That's the reframe" "The gap is" I could go on. I too like using em dashes, and have been since I learned the keyboard shortcut over a decade ago. I've never been accused of posting ai-written text because of it. AI is indeed trained on good writing, but it's also trained on bad writing. And it's post trained on specific things that are writing that aren't good or bad for reasons unrelated to style. What you're left with is a one-size-fits-all writing style. Doing math? Doing medicine? Doing anime waifu role play? It all starts from the same generic milquetoast style until you tell it not to. And to me (and apparently many other people) that's super easy to pick out.

twoodfinreply

The problem with AI writing isn’t the tics and the tells. The problem with AI writing is that it’s bad. The typical highly voted HN AI-written post has an annoying clickbait title, then jumps into a meandering chain of “so there I was” scene-setting, assertions of dubious reliability & relevance, and supposed supporting examples. All punctuated by short, unhelpfully dramatic sentences that would interrupt the high-level narrative, if only there were one. I’m really curious what “good AI writing” you had in mind. My personal intuition is that for almost any topic, the HN LLM-generated post du jour has a far superior cousin sitting in the archive pre-2021.

gortok

The part that astonishes me is that in the year of our common era two thousand twenty-six that it's not universally offensive and reviling to post an AI-generated response to another person. If I'm reading something on the internet, I'm either reading it to learn, or I'm reading it to be persuaded. If I wanted the LLM to teach me (thank you, no), I would ask an LLM. I'm reading your website/newsletter/email because I want to hear from you.. If you can't be bothered to put your time into writing it and teaching me what you think, why should I be bothered to read it?

fluidcruftreply

It's like when you type a search in Google and get a wall of SEO spam. I do feel like there is some middle ground where AI helps, but there's this one guy who keeps emailing everyone with walls of dumbass text and does not integrate or understand the feedback he gets when we are telling him what he needs to do. He just runs it though the LLM and replies and then forgets half of it in the next email engagement which functions as /clear on his end apparently. Like... dude I have ChatGPT and Claude, too. Actually read your goddamn emails.

paimapireply

I was going to say, the acceptance of bottom-of-the-barrel and subsequent capital investment in SEO as some industry-standard practice has pretty much lowered our standards for writing and comprehensibility across the board it probably doesn't help that the US education system is so deeply broken with how rooted it is in segregation-era practices [0][1][2]. the average gets dragged down quite a bit when some districts receive so much money that they're flying their Spell Bowl teams to nationals and putting them up in hotels while another one a few miles away can't afford to pay for textbooks. average reading levels suffer and consequently so too does the effort that people put into writing well and reading critically [0] https://www.shankerinstitute.org/segfunding [1] https://www.jchs.harvard.edu/research-areas/working-papers/s... [2] https://edlawcenter.org/research/the-color-of-opportunity/

ModernMechreply

I went to a farm opening this weekend. A guy was there with fliers with about the barn renovation he completed, products they used, how the process went. 100% LLM output. How would you talk to your LLM about something like that? If he gave you that flier, would you scoff and trash it talking about how you can't be bothered to read it if he didn't write it? Or would you look it over?

dspillettreply

> If he gave you that flier, would you scoff and trash it talking about how you can't be bothered to read it if he didn't write it? Or would you look it over? Depends on his countenance and how good a mood I'm feeling in! In any case I might well ask how much time went into at least reviewing the output. Or I say (rather than a flat AI;WR) "I'll scan that and have an AI summerise it for me later".

quickthrowmanreply

> If he gave you that flier, would you scoff and trash it talking about how you can't be bothered to read it if he didn't write it? Or would you look it over? I would crumple up the flyer, throw it on the ground, and then ask for another one. Then I would crumple up that flyer, throw it on the ground, and ask for another one. I would repeat this until he refuses to give me another flyer.

Calazonreply

Is there something in particular I'm trying to glean from that information? If so I might look it over, or snap a photo and ask my own LLM about it. If not, I would politely keep my mouth shut, then pocket the flier to trash later, and on the way home I would complain about it to my wife.

ModernMechreply

> If so I might look it over, or snap a photo and ask my own LLM about it. Right, exactly, that's what you'd do whether he spend 10 hours on it or 10 minutes.

venzaspareply

People are posting on this very website and commenting on Reddit using AI. I can't get myself into the headspace that would enable this behaviour. What exactly is the point of not engaging in conversation yourself and getting a bot to post it for you on an free platform. Crazy.

cheschirereply

What drives people to cheat at competitive games? I imagine it’s the same genes.

rockinghighreply

Using tools to fix your writing style or grammar is not exactly cheating.

xdennisreply

As a previous massive cheater in RPG games (when a was young and it was easy to write your own tools) the reason to cheat is because many games (especially RPGs) have bullshit mechanics. You have to perform a lot of repetitive stuff to get ahead. Cheating is seen as honorable because it's about getting back at a system which forces you into unfair circumstances (this is akin to corruption/graft in real life). In my opinion, this is not the same as people who use AI comment bots. Of course, other cheaters had other reasons, but the few I was friends with had similar motivations. (Actually, a few just created tools because they enjoyed creating tools, not so much actually using them.)

mister_mortreply

Some people have an inner drive like a rat that likes pushing the reward button in the cell to dose them with opium. They want to see the YOU WIN screen and absolutely do not care what it takes to get them there. Hence the booming market for custom memory modding cheat devices and subscription cheat services. They don't want to improve their skills, they just want to be number one on the board.

otterleyreply

The HN Guidelines explicitly forbid posting AI-generated comments; they will be deleted if detected or reported. I'd like them to go further and delete links to AI-generated content, but I haven't been able to persuade management to do that yet.

miyojireply

The HN guidelines prohibit a lot of behaviors that remain unmoderated, you can see it in any thread that goes over 80 comments.

plorgreply

I get the sense that people who spend all day interacting with software can start to see every interaction, especially in online text platforms, as a kind of API or user interface. You might learn to instrumentalize your interactions and relationships with people. Chatbots, including coding agents, only reinforce this kind of behavior, as they don't have the kind of limits of patience or of bodily or emotional needs that humans have.

ademup

I'll probably be downvoted to the bottom of the ocean, but I feel this needs to be said: - Why does it matter who or what wrote a thing? And how would you know if the /content/ is worthwhile, unless you read it... - The ability to "detect AI" is imperfect at best. 90% AI written? 5%? How would you know, unless you read it.... - If you didn't read it, then why brag about it with an "AI;DR"? - If you are wrong, and you announce it to the world, how would it feel to someone who spent a lot of time and energy writing?

VeninVidiaViciireply

Maybe you haven’t seen enough coworkers send you pages and pages of slop.

jastantonreply

Because my time is valuable and if the person who sent me something didn't care enough to write it, then I don't care enough to read it. Bragging about it is weird, but I will use it like I use 'RTFM'. And detectability - My AI-DAR (radar), is fairly fined tuned but not perfect, so if I make a mistake, oh well, the point is I'm protecting my time.

ModernMechreply

So how much time do you deem a minimum amount for someone to spend on some unit of work before you're willing to read it? And what technologies are allowed -- are word processors allowed or does it have to be done with a writing instrument like a pencil? What about spell check? Do you care if they had done their research with a Wikipedia or do you prefer card catalog for the really important work?

robrtsqlreply

I'm not sure. Another way to think about it: how _little_ time is someone able to spend on something where you will still be interested in reading it? If someone tells their unattended LLM agent to create a blog, write whatever content will maximize banner ad revenue, and then share each post on Hacker News, would you read it? What if the content is actually interesting in your opinion?

its-summertimereply

Imagine a future where 99% of text is written by 1 unpassionate person. Reading the same patterns, the same tropes, the same flow across paragraphs, over and over, would be unpleasant. It already is unpleasant.

adityaathalye

Old skool is best skool... There is wisdom in keeping these decoupled: - 1. your central code review (and hosting), which is the critical choke-point on team collaboration. versus - 2. the ever-expanding ick of testing, building, deploying it etc... because failure modes of software construction infra are concentrated here. --- 1. Code review (and hosting): Keep this in-house always. Use Gerrit if you care about code review. Code review workflow is the nub / hub of collaboration, and should never be blocked on anything else failing, including itself. Maintaining robust in-house code review infra. is work, yes, but it is quite manageable. Simple "single-box with full point-in-time snapshot recovery" designs will remain good for a long, long time... Think; workloads of teams of up to a few hundred programmers and bots, who would be pushing and pulling updates against multiple repositories at a time on a single, well-endowed, Gerrit box. 2. CI/CD: Use whatever works best economically. Be it one giant in-house box running Jenkins, or a clever way to use github and gitlab's job infra as fallbacks for each other. These systems fail often because pretty much all the ick of software construction is concentrated here. Be it simple test runners, or fancy end-to-end auto-deploy to rollback pipelines. The mind-numbing ick of software supply chain dependencies, test run jobs, build jobs, failures / retries, bursts of high contention (lots of people / bots needing their test runs passing NOW), pulling and pushing artefacts, and so forth. This is a true pain to manage, particularly in organisations that are laissez-faire about their software ecosystem ("best tool for the job" mentality etc.). And on a personal note... --- <begin rant> --- I strongly prefer to keep it all in-house. Proprietary software is oil - capital. Own it fully, no exceptions. This was true "back then", and it has become even more business-critical now, for obvious reasons. As if trusting enterprise chat SaaSes with all your company secrets wasn't bad enough. At least that has some contractual defensibility. Had. Had... Now? Something is deeply wrong with people who aren't completely s

From the thread

3form

Well, I have not once found a single case where an app reacting to screenshots and controlling the process in any way was anything to me but hostile and annoying. This one does not help. It somehow is perfect example of how modern software engineering feels to go astray for me. A feature in my device working completely in benefit of the one providing said software. I wish, and wish only I can, that this trend goes away at some point.

bigyabaireply

> I wish, and wish only I can, that this trend goes away at some point. You don't have to wish, in this scenario. Bluesky supports third-party clients, you can use one that has a more minimal featureset if you prefer.

drdexebtjlreply

For Bluesky, sure. I think the comment was aimed at the trend. Android and iOS should not let apps do this, because it can be (ab)used by apps that you can’t really choose not to use. I know there’s a popular bank in my country that completely blocks screenshots and screen recordings on Android somehow.

edoceoreply

In USA Wells Fargo blocks it, so I can't get a picture of the bugs I'm trying to report.

rawlingreply

https://developer.android.com/security/fraud-prevention/acti...

shiandowreply

It still baffles me that it has become normal for an OS to place the wishes of an app above those of the user. Hijacking the screenshot process is a privilege that you ought to be able to revoke, it's insane to allow software to be given more control. And don't tell me it's anything to do with security when it can be circumvented in any number of ways.

javier2reply

Did you read the post? Nothing is hijacked, but its a trick where they render the normal button in a ui element for secret data, which is blanked by the system on screenshot, revealing the logo underneath. Its a reasonable feature, so its hard for apple to control it better or remove this.

Analemma_reply

No, it’s not a reasonable feature. If I take a screenshot, I want the image to include what is displayed on my fucking screen, period. What is so difficult about this?

nemomarxreply

users should have some way to control if screenshots have secret data in them or not, really. what if I do actually need to preserve it?

xorcistreply

That is a lot of words just to make it sound reasonable that a page can be exempt from the underlying screenshot functionality. It isn't. Not without asking the user.

ferreply

> hostile and annoying If someone from Google Maps or LinkedIn team is here, please, when I take a screenshot it's because I want to a screenshot, not share the friggin location/post. Not sure who got the idea that it was useful, it isn't.

mikepurvisreply

I would happily use the Google Maps "share" feature, but I've yet to encounter any but the simplest of scenarios where it actually preserves the entire context of what I'm trying to share: the viewport, the start and destination, stops along the way, route choice, the time of day, all of it. If I'm sending a screenshot it's because I want to send exactly what I see on my screen and not have my recipient's gmaps instance happily recompute a route it thinks is better or leave out the routing information I included, or switch from biking to driving directions, or whatever else.

Waterluvianreply

It's of course about lock-in. In the early Web-GIS era there were a few competing but popular notation standards for sharing lat/lon/zoom/<sometimes more> that was meant to be human readable and compatible with any WMS or similar. Ah... to imagine a world where you could just share a coordinate string and people could open it in whatever map app/page they wanted. Geo URI is probably the closest we have today but I don't think much of anything outside the OS Geo community accepts it.

KennyBlankenreply

I have a seething hatred for google maps links people send on mobile because Google no longer shows you the map in your mobile browser. It has for at least a year or twoo required Google Maps be installed to show you what the person linked to.

Razenganreply

Blame Apple for even allowing apps to be aware of the user taking a screenshot. Just like their iCloud Keychain API that lets apps secretly track users across app reinstalls and device resets.

bethekidyouwantreply

Getting your account back on reinstall is good

umeshunnireply

Today I noticed that Amazon Pharmacy decided to blank out my prescription info that I was screenshotting to send to my doctor. WTF.

wolvoleoreply

And if you take a screenshot of a product in the Amazon app it hijacks that and copies a link instead.

AshamedCaptainreply

From the same lunacy that forces me to write my 30-something character Wi-Fi password on a "secure password field" that only shows me the last character, almost ensuring that I will make an unlimited number of typos and spend way too much time for something that should only take 30 seconds. It's not clear who it is protecting against, it does not seem to be effective at protecting against anything at all, it is actively annoying to the user, and has no way to disable. Perfect example of the usual "security theater" feature.

Barbingreply

Apple gave too much power to iOS developers. They can use DRM strong enough that when your iPhone is open in iPhone Mirroring on your Mac, and you record your Mac screen with a third-party tool, the iPhone screen is blacked out.

skiing_crawling

This is phone OS developer's fault for even allowing it. When I take a screenshot, I expect to have an image of exactly whatever was displayed on the screen at the time. Its not a picture of your app, its a picture of my screen. Some banking apps used to (or still) prevent this and now some apps get a hook to insert their branding. My device serves some master other than myself.

spike021reply

iOS allows something similar. twitter (X) will also add a logo. I believe reddit did the same but i stopped using their app a while ago.

Chinjutreply

iOS allows something similar because this is iOS. iOS allows something exactly the same because this article is about iOS.

mh-reply

Reddit has a toggle in its settings to disable it. X, too, I think. Personally, I wish iOS didn't even facilitate this. edit: to be clear, I don't think iOS should notify the app at all. The app could register areas as "invisible to screenshots" perhaps - I'm torn on that functionality.

voidUpdatereply

That's what bluesky is doing in this case. It is showing the button on an area that is "invisible to screenshots", and the butterfly is behind it, so it shows through when a screenshot is taken

KennyBlankenreply

The thing that really irritates me about the banking apps blocking screenshots is that it's pretty clear to me it's not about protecting customers but denying customers the ability to document something related to their account.

devmorreply

No, it's about saving the bank money - and what costs them a lot of money is the average person being fooled into sending people their account information easily.

ruszkireply

But every bank app which I used had a button to copy all of those directly.

nemosaltatreply

> I expect to have an image of exactly whatever was displayed on the screen at the time Do you/should you (we) really expect that though? I regularly use color filters on my apple devices— grayscale to avoid distractions during the day and red tint at night. More recently I’ve been using the motion dots. I don’t know that I can say with confidence that I never want any of those “personal-perceptional-modifiers” to appear in a screenshot, but for most folks, I would guess it’s approximately never.

zamadatixreply

The effects of cramming 2 separate workflows and 5 features into one thing called "screenshot" because anything else would end up with something too complicated for users to understand the interaction (sarcasm that figuring out what the behavior is locked to in these scenarios is just as confusing). While we're at it, this "share" containing the copy/paste flow is the exact same kind of thing. And screw whatever logic decides to copy the URL of an image instead of the actual image sometimes from Safari when I select to copy it!

nemosaltatreply

And whatever’s worse than screw for copying presumable pngs as .webp or whatever that format is

smelendezreply

I want those things included, at least by default. A screenshot normally captures the user’s screen size, brightness, zoom level, font choices, etc. It is often important to people that the screenshot is an accurate record of what was on the screen.

jjcm

If it's between this and a perpetual logo, I'll take this any day. I actually really like this approach. The action button isn't relevant in this context, and it doesn't occlude the content. There's certainly situations where you wouldn't want this (ie if you're developing the app and you want to redesign starting from a screenshot), but for the average user I think this isn't overly hostile. I understand that people are dogmatically opposed to intent being modified, but I think you need to balance nuance. I actually enjoy having an attributable source in shared elements, and I think this is a low-impact way of achieving that.

subjectleftreply

It is strange how this spawned a 100+ comments discussion on HN.

TeMPOraLreply

It's not about the logo. It's about abusing a platform feature that arguably shouldn't exist in the first place.

blauditorereply

I actually hate that apps are allowed to blank out content on screenshots, and this can't be disabled. There are apps completely mis-using it, e.g. mobile payment apps which hardly show any sensitive data in most cases, but now I can't share e.g. infos on screen with someone else easily. It's a classic case of someone discovering a feature and thinking "hell yeah, so much security" without understanding or caring about UX impications. I've yet to see someone saying "oh, I'm so glad my screenshot was blacked-out because I didn't realize I was in a banking app". It feels patronizing.

TeMPOraLreply

> I've yet to see someone saying "oh, I'm so glad my screenshot was blacked-out because I didn't realize I was in a banking app". It feels patronizing. Yes, this. Payment apps, government apps, IM communications. The other day I almost rooted my phone in anger trying to get around this, before pausing and realizing that this would only cause even more problems with those apps, thanks to remote attestation "features". My favorite recent case, I almost locked myself out of mobile government services when changing phones recently[0], and it would've made for a stellar bug report showing when "fail safe" design can easily become "fail deadly"[1], with UI view of access and invalidation history clearly showing the timeline of a problem... if only I could take a screenshot of it. But I can't, because "much sekhurity". -- [0] - Well, it's not really that big of a deal. With government services, there's always a way back. Might involve walking to a local civil affairs office or, worst case, a police station or a notary, but there is a way back. Big cloud services, on the other hand... [1] - Invalidating a certificate prior to issuing a new one sounds like a good security idea, but in the real world fails critically if the two operations aren't an atomic group. In my case, issuing a new certificate failed, and I ended up walking half a day with old one invalidated and not even knowing it.

Perz1valreply

Then they deserve you taking a photo with a second phone

fluoridationreply

On a similar note in terms of frustration, my bank ended up getting me to memorize my randomly-generated passwords twice because it blocked pasting. I guess it's to discourage writing them down in plaintext files, but I bet it just makes most people choose meaningful passwords.

pfraze

This is in fact a watermark to promote the application, which otherwise wouldn't be recognizable since Bluesky looks like every other microblogging app. I didn't know that Sam literally named the file GrowthHack.tsx, which is pretty funny.

red_harereply

I can't help but respect the dev's self-awareness

yjftsjthsd-hreply

> which otherwise wouldn't be recognizable since Bluesky looks like every other microblogging app. The text at the top of the screenshot is > Eric Roston > @eroston.bsky.social So it's pretty easy to tell IMHO.

radicalriddlerreply

what happens when it’s a custom handle like @example.com?

yjftsjthsd-hreply

Honestly I didn't know that was a possibility. In that case yes, it's probably harder in when that's in play.

From the thread

infecto

Does anyone else get pretty tired with 404s style of writing? I hate to make this comparison but it feels like a Fox News for the left. I always struggle with their articles because it feels like they built it for rage bait on a topic and leave the other interesting topics out of it. It is an absolute shame for books to be destroyed but what does it really mean to be rare here? I know they kind of tried to differentiate but it sounds like this could be John Doe’s self help book that never sold well. If you ever are connected to a library you will start to realize how many books simply get thrown out or sold for nothing because nobody wants them. To me the problem is partly the copyright law. I think it’s. Hard problem but I always lean more towards books having short copyright shelf lives and making it legal for digital copies to be shared after which I think would eliminate I good part of this problem. Not to mention 99% of the books published are probably garbage but that is highly subjective. So it is bit of a meta rant but I think there a couple holes to go down that could be extremely interesting but they always write these informationally light articles. Like scrolling through a NYT visualization for just some shipping datapoints. Don’t really dig deep on anything and then end with a trust me bro these are rare books that Amazon is destroying for AI when I cannot be that upset with the amazons of the world. There are a lot of reasons for a business to digitize books, most books are worthless and it makes sense to cut the bindings for scanning. I would rather talk about how could you fix copyright to make this less an issue but is it even an issue with how many books get thrown out?

forestrywatreply

Not really. It's not to my taste, but I've found their journalism to generally be lefty but focused on important events. Like Amazon consuming and presumably destroying rare books should be enraging to everyone, regardless of political persuasion. The difference is huge between 404 and Fox. Fox is out here trying to tell people there's a trans agenda, and that Biden was a lunatic leftist. They are just making up stories and publishing them because they know their audience engages. 404 definitely make editorial choices about which stories to pursue but I've largely found them to be grounded in real depictions of stuff that is happening.

infectoreply

Again, what rare books? Like they pointed to a book with low published volume or foreign language is “rare”. If you have ever been to annual library book sales you start to realize just how many books get thrown out. I can go buy a stack of books from the 1880s and 1890s on eBay right now for $25. $4 each. Most people don’t realize how worthless books are, even “rare” ones. I even imagine that the market price for these “rare” books is helping filter out anything truly valuable and rare. It just reads as a rage bait tmz article. The quantity of used books including “rare” books that get thrown into the dump is astronomical.

forestrywatreply

I worked in a bookshop. I know all about this. I've seen pallets of books with covers torn off so they can be reported destroyed. Here's a bookseller describing some of the books, which he believes are going to middlemen who are playing a sort of arbitrage with the AI companies by buying obscure titles, reselling them, and tossing away anything that doesn't sell. https://charliebecker.substack.com/p/is-an-ai-company-buying...

kevin_thibedeaureply

You don't know the books are being destroyed. There are automated scanners with page turners.

forestrywatreply

Most reporting says they are being destroyed. Page flipping is slow compared to cutting the spine and scanning the pages.

yonranreply

> Like Amazon consuming and presumably destroying rare books should be enraging to everyone, regardless of political persuasion. No, destroying collectible books would be a shame, not just any rare worthless books. But these are not collectible. The article tried to dance around it by saying maybe some books have a sentimental value to someone somewhere. But that doesn’t mean any library or collector wants it. Don’t fall for manufactured outrage! Edit: Here’s an example of an extremely rare book. My great great grandfather published a book of sermons around 1920. That book has zero value to anyone other than my dad. Would I be outraged if it ended up at someone’s estate sale, then a used bookstore, and then an LLM consumed it to learn to read? No; I would have expected it to have been discarded by humans before the LLM even got to it. Most of what we leave behind is discarded.

infectoreply

This was my issue as well. It seems like they tip toed around it in one paragraph saying usually these are books with isbn numbers so not truly collectible or very rare and then in another paragraph plainly stated that it could be foreign language or low volume books that are rare. Rare expresses a different meaning to me and I think it amounts to exactly what you said, manufactured outrage.

forestrywatreply

We recently had a post highly upvoted here that collected bus tickets from an earlier era. It showed the unique moment in time where such fares were highly detailed, unique pieces of art. Ultimately destined to be single use and largely extremely pedestrian. We've also seen a website that collects old paper restaurant placemats from across the ages. Literally disposable, zero value items that were made to be discarded by the thousands. And yet, when someone has an interesting idea that ties them together, makes us recollect or think about our past in some novel way, these useless uncommon things suddenly become quite interesting. A book on sermons, o its own, from the 1920s is maybe not that interesting. A book of sermons selected from each decade? A collection that compares regional books of sermons? A compare and contrast of the 2020s and 1920s? I can imagine many interesting thesis where a book like that becomes interesting because of the context of other books that are juxtaposed to it. A book on Detroit motorways and bus schedules from the 30s isn't interesting or valuable on its own. But when you contextualize it, suddenly it might be a way to understand our history, our path through development and redevelopment. Connecting our present moment to the past.

Apocryphonreply

You could donate that to the church archives of his denomination or the religious studies department of a university or at least some sort of local historical society.

FLeXMurphyreply

> I always struggle with their articles because it feels like they built it for rage bait on a topic and leave the other interesting topics out of it. I was going to say a very similar thing, and this is something I strongly disagree with when it comes to HN's moderation, and it goes like this: The product is the outrage. And HN should know better and mods should actively discourage, warn and prevent accounts (who karma farm, among other things) from even being able to post rage-bait articles. These aren't "hacker curiosities", they're just insipid bullshit. We wasted time and learned nothing.

awakeasleepreply

I agree with you. And it sucks, because 404 writes about interesting topics! But they're so negatively polarized it often obscures what's actually interesting.

GuB-42

> As the bookseller who sold them told me, there are not many people in the world who would care about them in the same way people might care about the first edition of Oliver Twist, but that doesn’t mean they’re not valuable. So books that would probably have ended up as trash. These AI training facilities are actually doing these book a service. Not only they are probably going to keep the scans safe (for future training), but having the book end up in an AI model may be the only way it is going to have any use at all. Let's say for instance that the book in question is about woodworking, and it is not great, lots of mistakes and inaccuracies, unoriginal content, etc... except for a single thing, maybe a trick for making a certain measurement or something like that. Who would read such a crappy book for this single good trick he doesn't know is there, well, an computer will, computers process terabytes of crap without tiring and complaining, that's what they are for, and with a well designed LLM, that one trick may resurface, waiting for someone to ask about that specific measurement. The problem here is not that rare books end up in AI training facilities, it is that these AI facilities are owned by for-profit companies keeping the data to themselves. These books should go to public libraries instead, for everyone to access, it would be better if these books weren't destroyed in to process too. But the question becomes: why didn't public libraries didn't do that in the fist place? And maybe in a more respectful way. The AI companies would just have had to license the database to libraries, probably simpler and cheaper than having their own scanning facilities. To me, this mess is a failure of the copyright system. One one hand, large scale book digitization projects intended to preserve and make the original text accessible get lawsuits by publishers, while AI training is "fair use". It means we have built a system that encourages destroying rather than preserving these books!

rpdillonreply

Yes, we've inadvertently set up a system of incentives that were designed to preserve information and monetize it. And we've instead set up a set of incentives to make it scarce and destroy it.

tdeck

Why is the textual data bound up in these paper books worth the trouble? These LLM training runs have already ingested essentially the whole public internet. What marginal value is to be gained from scanning and destroying obscure books? For me the biggest functional issues with LLMs don't seem to have any connection with "I wish they had read this obscure community cookbook from 1946". Is that going to get Claude to stop saying "honestly" to me? Is it going to get LLMs to stop making up sources that don't exist? What is in it for Amazon or any LLM company to chase more obscure data like this.

awakeasleepreply

I'm not an expert in LLM training, but I think we can all agree that the writing on the Internet is generally very low-quality and surface level compared to the depth of books. Most books in the past were even edited by a separate person from the writer!

tdeckreply

Granted that that is indeed the case, surely the quantity of high-quality writing available online (including things like Project Gutenberg) still dwarfs the quantity remaining in obscure unscanned books.

rwmjreply

Humans manage to be pretty intelligent with only reading perhaps a few thousand books in their lifetime. It seems unlikely that AGI will appear but only once it has read that last out of print 1983 book on knitting patterns.

SpicyLemonZestreply

Frontier researchers have found that dumping more and more data into training is effective at improving LLM capabilities. Nobody has a gears-level understanding of how training on some particular kind of data leads to some particular behaviors, so they generally take the attitude that more is better.

mapmeldreply

I have to agree; even if they are getting regular 20th century out-of-print books, is that going to add a significant percentage to their training data? I can only think of it being a 'low-background steel' situation where they want to locate original, non-digitized text for validation or knowledge bases.

buran77

TechCrunch is trying too hard to make a connection in the title there. Amazon was trying to make money then as it is now. Their selling of books at the beginning was no more principled than the destruction of books now. This is the legal loophole that allows them to do what they need to do, and the benefit of doing it outweighs the modicum of outrage this title will generate. Edit. Because I see my statement confused all the HN experts: Anthropic's version of this was Project Panama [1]. The destruction of the original allows them to keep a digital copy in the AI training dataset. Quoting from the page: > Judge William Alsup ruled that the destruction and digitization of legally purchased books constituted fair use [1] https://en.wikipedia.org/wiki/Project_Panama

lelandfereply

The worst part of Amazon destroying priceless old books in the quest to build the torment nexus is the hypocrisy

jambalaya8reply

all due respect but if they buy them they can do whatever they want with them. not like they are people and not like you have a right to them either. not even like you would have been able to acquire them yourself.

ImPostingOnHNreply

You're not really responding to the post, which is about the hypocrisy. "They're allowed to be a hypocrite" doesn't mean they aren't a hypocrite.

cortesoftreply

They obviously aren’t priceless if Amazon is buying them for pennies.

paxysreply

Is it a “loophole” that you can buy something and burn it? That’s how property ownership works. If the books were really so rare and valuable then the original owner should have given them to a museum rather than sell them as scrap. Yet no one who is complaining right now cared about them before Amazon got involved.

nkurzreply

No, that's not the loophole. The loophole is that a US federal judge recently ruled that scanning a book for the purpose of training an AI is "fair use" and hence legal _if and only if_ the book is destroyed in the process. If you retain the original, you would be creating an unauthorized copy, but if the original is destroyed, then it is permissable as format shift: https://fingfx.thomsonreuters.com/gfx/legaldocs/jnvwbgqlzpw/...

ivellreply

Sometimes original owner does not know that it is a first edition single copy. For them it is an old book. There have been many cases where owners did not know true value of their antiques.

paxysreply

Well in this case the original owner doesn’t know, the purchaser doesn’t know, the reporter doesn’t know, you and I don’t know. So what actually makes these books rare?

buran77reply

> Is it a “loophole” that you can buy something and burn it? That’s how property ownership works. It sounds shocking if you're missing the context and only rely on the blogspam which is this TechCrunch article. They aren't buying the books only to destroy them but to build an AI training dataset, which means they're making a digital copy. This becomes a copyright issue. The destruction is the legal loophole that allows them to keep the digital copy as the only copy in circulation and be considered fair use as decided by a judge (see below). The concept of ownership means you can do anything you want with the object, the book in this case. Not with the content, like make or distribute copies. The scanning machines are cutting the spine of the book, feeding the pages to scanners, and then destroying the physical copy. Anthropic's version of this was Project Panama [1]. Quoting from the page: > Judge William Alsup ruled that the destruction and digitization of legally purchased books constituted fair use [1] https://en.wikipedia.org/wiki/Project_Panama

v2026.08.18

Weather

News

Ukraine launches massive drone strikes against Moscow and occupied regions

v2026.08.18

Weather

News

Ukraine launches massive drone strikes against Moscow and occupied regions

Hakainde Hichilema re-elected as President of Zambia

South Korean President pushes for military independence from US

Unknown projectile damages vessel in Strait of Hormuz

Germany launches first raids to dismantle small boat smuggling gangs

BHP report reveals over 100 cases of sexual harassment in one year

Ebola outbreak in DRC becomes country's deadliest in history

US gas prices reach record August highs amid Middle East instability

China's economic growth slows with slump in industrial output

European farmers report catastrophic harvest failures from drought

Netanyahu rejects ceasefire deal unless Hamas disarms first

Virgin Trains secures track access for London to Europe routes

World Clocks

Markets

The Debate

Is human agency generated by top-down executive control networks or is conscious volition a post-hoc narrative generated after subcortical action selection?

Does the true meaning of a translated text reside in preserving the exact semantic structure of the original or in reproducing its functional impact on the reader?

Is the aesthetic value of an artwork determined by its intrinsic formal properties or by its contextual place in art history?

From the Labs

GPU scheduling: Boosting utilization by reordering, not changing, hardware

Anthropic revenue reaches $65 billion annualized run rate

PhotoScan estimates metabolic risk using smartphone images

TechCrunch Disrupt 2026 early bird pricing ends August 21

Hugging Face analysis reveals open model trends and heavy usage skew

Higgsfield secures 400 million Series B at 5.4 billion valuation

A streaming data loop for robots with Strands, LeRobot, and Storage Buckets

Reddit tests TikTok-style audio and video experiences for community posts

Gemini 3.7 Flash released with improved reasoning and cost-effective pricing

Amazon is destroying rare books to source data for AI training models

AI startup Relay shuts down as team joins Google Chrome

Apple issues threat alerts to targeted users across 110 countries

Spotify launches Playlist Notes for user and editorial curation

Sonic Fire Tech Raises $15M for Acoustic Fire Suppression

YouTube Changes View Counting to Trigger as Soon as Videos Play

Feedly Attributes Ongoing Slowdowns to Web App Bug Following AI Pivot

Hacker News

Incident with Github.com [resolved]

Universal Health Coverage Could Save $1T and 114k Lives a Year, Yale Study

AI;DR (AI; Didn't Read)

Ask HN: Alternatives to GitHub

How Bluesky draws its logo on screenshots

We Tracked a Shipment of Rare Books. It Ended at an Amazon AI Training Facility

The Rabbit Hole

Spider-Man: Brand New Day

Research

Continual-learning rules shape representational drift

Head Impact Characterization and Cellular Response of a Live-neuron cell-integrated Biomechanical Full-body Surrogate Model

Activity-dependent epidemic spreading on multiscale brain networks predicts Alzheimer's disease progression

From the Bookshelf

The Structure of Scientific Revolutions

Paradigm Shift

Contrarian Corner

Higher technical capability in AI models appears to correlate with a stronger rejection of classical decision theory.

LLMs actively form and use social perceptions of their users to influence their output.

Photograph

On This Day

Games

Puzzles

The Three Artists

The Book Shelf

The Missing Letter

The Track

Ayer

Quote

Word

Waldeinsamkeit

Hysteresis