# The Daily AI Digest — Recent Editions Index > Daily briefings on developments in AI, curated and edited by Alex Panetta. > Website: https://dailyaidigest.net > Author: Alex Panetta — https://alexpanetta.substack.com > This file lists section headers and item headlines from the 30 most recent editions. > Plain-English summaries and 'Why it matters' analysis live only in the full editions — follow the URL on each entry. > See https://dailyaidigest.net/llms.txt for a structured index of the site. --- # 2026-05-30 — Boston Children's Hospital Reports 40 Rare Diagnoses, 60,000 Hours Saved Using AI URL: https://dailyaidigest.net/archive/2026/05/30/ Sources: Google AI, Hacker News, Hugging Face Papers, NBER, OpenAI Items: 13 ## What's New ### [Opinion: AI's Trillion-Dollar Valuations Only Work If It Replaces Workers at Scale](https://www.owenmcgrann.com/p/the-dead-economy-theory) ### [Rockstar Games Workers Unionize, Allege Firings Were Retaliation](https://rockstarintel.com/gta-6-developers-announce-rockstar-games-union/) ### [Mistral Pitches On-Premise AI as Europe's Alternative to U.S. Hyperscalers](https://koenvangilst.nl/lab/mistral-ai-now-summit) ### [Startup Offers Free Home Cleaning—If You Let Them Film Everything for Robot Training](https://www.theverge.com/ai-artificial-intelligence/939765/ai-training-data-startup-shift-free-cleaning) ### [Developer Argues AI Will Deskill Programming the Way Frameworks Deskilled Frontend](https://mastrojs.github.io/blog/2026-05-23-is-AI-causing-a-repeat-of-frontends-lost-decade/) ## What's in the Lab ### [Google Unveils Conversational Video Editing and Faster AI for Automated Workflows](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-omni-3-5-videos/) ### [Boston Children's Hospital Reports 40 Rare Diagnoses, 60,000 Hours Saved Using AI](https://openai.com/index/boston-childrens-hospital) ### [AI Startup Claims It Turns Customer Requests Into Working Code in Minutes](https://openai.com/index/braintrust) ### [OpenAI Offers U.S. Government Specialized AI for Biodefense](https://openai.com/index/strengthening-societal-resilience-with-rosalind-biodefense) ### [Framework Proposes Standards for Third-Party AI Safety Evaluations](https://openai.com/index/trustworthy-third-party-evaluations-foundations) ## What's in Academe ### [People Avoid AI Explanations When They Reveal Uncomfortable Bias](https://www.nber.org/papers/w35240#fromrss) ### [Time-Locked AI Models Eliminate a Persistent Problem in Financial Backtesting](https://www.nber.org/papers/w35247#fromrss) ### [Wikipedia Lookups Beat Complex Neural Fact-Checkers in Accuracy Tests](https://huggingface.co/papers/2605.29648) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-29 — Claude Gets Opus 4.8 Model: Details On Pricing And Performance URL: https://dailyaidigest.net/archive/2026/05/29/ Sources: Cohere, Google AI, Hacker News, NBER, OpenAI, arXiv Items: 15 ## What's New ### [Anthropic's Claude Opus 4.8 Claims Top Spot on AI Agent Benchmarks](https://www.anthropic.com/news/claude-opus-4-8) ### [Anthropic Raises $65 Billion, Now Valued at $965 Billion](https://www.anthropic.com/news/series-h) ### [Connected Cars Collect and Sell Your Data, With Few Privacy Protections](https://www.bbc.com/future/article/20260513-your-car-is-spying-on-you-its-about-to-get-worse) ### [GitHub Bans Researcher Who Published Windows Zero-Days After Alleged Bounty Dispute](https://www.tomshardware.com/tech-industry/cyber-security/microsofts-github-bans-security-researcher-who-posted-zero-day-windows-exploits-because-company-ruined-their-life-expert-claims-action-is-vindictive-and-promises-further-retaliation) ## What's in the Lab ### [Google Previews Video-Generating AI and Search Agents at I/O](https://blog.google/innovation-and-ai/technology/ai/io-2026-keynote-moment-videos/) ### [Software Firm Claims AI Agents Let Junior Developers Produce Senior-Level Work](https://openai.com/index/endava) ### [OpenAI Maps Its Safety Practices to Emerging AI Regulations](https://openai.com/index/openai-frontier-governance-framework) ### [Japan's Largest Bank Deploys ChatGPT to 35,000 Employees](https://openai.com/index/mufg) ### [Cohere Publishes Framework for Adding AI to Business Intelligence Tools](https://cohere.com/blog/ai-for-business-intelligence) ## What's in Academe ### [Open-Source Tool Lets Mental Health Apps Keep Sensitive Data Local](https://arxiv.org/pdf/2605.30273v1) ### [AI Fact-Checkers Improve With Web Access but Choose Different Sources Than Humans](https://arxiv.org/pdf/2605.30241v1) ### [Medical AI Stumbles on Real Hospital Data Formats, Study Finds](https://arxiv.org/pdf/2605.30295v1) ### [Language Models Handle German Grammar Well but Struggle With Neopronouns](https://arxiv.org/pdf/2605.30214v1) ### [Working Paper Claims AI Can Spot Financial Risk 10x Better Than Traditional Methods](https://www.nber.org/papers/w35227#fromrss) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-28 — Friendly Chatbots Make Users More Likely to Trust Wrong Answers URL: https://dailyaidigest.net/archive/2026/05/28/ Sources: Cohere, Hacker News, Meta AI, OpenAI, arXiv Items: 16 ## What's New ### [If AI Boosts Productivity, Should Workers Get Shorter Hours?](https://mlsu.io/posts/day-off/) ### [YouTube Will Auto-Detect AI-Generated Videos Starting 2026](https://blog.youtube/news-and-events/improving-ai-labels-viewers-creators/) ### [Enterprise AI Prices Reportedly Rising as Labs Near Profitability](https://simonwillison.net/2026/May/27/product-market-fit/) ### [DuckDuckGo Traffic Jumps as Users Seek AI-Free Search](https://www.pcgamer.com/hardware/duckduckgos-ai-free-search-saw-nearly-28-percent-more-visits-in-the-week-following-googles-insistence-that-people-love-ai-mode/) ### [Music Tracking Service Last.fm Goes Independent After 17 Years Under CBS](https://support.last.fm/t/last-fm-is-now-independent/118591) ## What's Controversial ### [Cisco Claims OpenAI's Codex Now Writes 95% of Its New AI Features](https://openai.com/index/cisco) ## What's in the Lab ### [OpenAI Tax System Claims 97% Accuracy, Learns From Its Own Mistakes](https://openai.com/index/building-self-improving-tax-agents-with-codex) ### [Terminal App Warp Says GPT-5.5 Now Writes 90% of Its Code Changes](https://openai.com/index/warp) ### [OpenAI Expands Election Safeguards for 2026 US and Brazilian Votes](https://openai.com/index/election-safeguards-2026) ### [Quebec AI Partnership Aims to Make French Models Sound Local](https://cohere.com/blog/cohere-and-mila-partner-to-advance-quebec-french-language-and-cultural-context-in-ai) ### [Meta Claims 23x Speed Boost for Recommendation Systems With New Architecture](https://engineering.fb.com/2026/05/26/ml-applications/silvertorch-index-as-model-new-retrieval-paradigm-recommendation-systems/) ## What's in Academe ### [Humans Often Misjudge When to Trust AI Teammates, Study Finds](https://arxiv.org/pdf/2605.28255v1) ### [Workers Fear AI Will Make Their Jobs Look Less Meaningful](https://arxiv.org/pdf/2605.28680v1) ### [Uncertainty Displays in AI Tools May Backfire, Reducing Fact-Checking](https://arxiv.org/pdf/2605.28571v1) ### [Friendly Chatbots Make Users More Likely to Trust Wrong Answers](https://arxiv.org/pdf/2605.28498v1) ### [AI Search Agents May Confirm Existing Knowledge Rather Than Discover New Facts](https://arxiv.org/pdf/2605.28721v1) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-27 — Uber Latest Company To Question Rising AI Costs URL: https://dailyaidigest.net/archive/2026/05/27/ Sources: Cohere, Hacker News, arXiv Items: 11 ## What's New ### [Spain Blocks Polymarket and Kalshi, Classifying Prediction Markets as Unlicensed Gambling](https://www.reuters.com/business/spain-blocks-prediction-markets-polymarket-kalshi-over-lack-gambling-licences-2026-05-26/) ### [Offshore Teams With Open-Source AI Could Cut API Costs 30x, Analyst Claims](https://www.signalbloom.ai/posts/outsourcing-plus-localai-will-soon-become-more-economical-vs-frontier-labs/) ### ["Sleep Mode" for LLMs Could Make AI Responses Faster and More Accurate](https://arxiv.org/abs/2605.26099) ### [Uber Burned Its 2026 AI Budget in Four Months, Struggles to Show Results](https://www.theverge.com/transportation/937116/uber-ai-investment-hard-to-justify) ## What's in the Lab ### [Cohere Launches Branded Merchandise Store](https://cohere.com/blog/2026-summer-merch-collection) ## What's in Academe ### [Same Hiring Algorithm May Block Some Applicants Across Multiple Employers](https://arxiv.org/pdf/2605.27371v1) ### [AI Document Clustering Falls 16-26 Points Short of Human Expert Reliability](https://arxiv.org/pdf/2605.27168v1) ### [Framework Proposes Measuring the Hidden Costs of AI Agents](https://arxiv.org/pdf/2605.27320v1) ### [Safety Layer Cuts AI Finance Agent Attacks While Preserving Performance](https://arxiv.org/pdf/2605.27333v1) ### [LLMs Struggle to Match Local Style When Writing for Different Chinese Markets](https://arxiv.org/pdf/2605.27296v1) ## What's On The Pod --- # 2026-05-26 — Showing AI's 'Reasoning' Makes Users Feel Better but Perform Worse URL: https://dailyaidigest.net/archive/2026/05/26/ Sources: Hacker News, OpenAI, arXiv Items: 11 ## What's New ### [California Bill Would Carve Out Linux From Age-Verification Rules](https://www.tomshardware.com/software/linux/california-moves-to-exempt-linux-from-its-upcoming-age-verification-law-after-backlash-over-forcing-operating-systems-to-collect-users-ages-amendment-proposed-by-the-same-lawmaker-who-wrote-the-original-law) ### [One Developer's Case for Using Multiple AI Models as Code Reviewers](https://nolanlawson.com/2026/05/25/using-ai-to-write-better-code-more-slowly/) ### [Norway Builds Sovereign AI Using 20 Petabytes of Library Archives](https://www.blocksandfiles.com/flash/2026/05/22/norways-2-petabytes-of-huawei-flash-storage-and-llm-training/5244910) ### [Security Researchers Claim Microsoft Copilot Vulnerability Could Leak Files](https://www.promptarmor.com/resources/microsoft-copilot-cowork-exfiltrates-files) ### [YC Startup Claims It Can Send iMessage for Business—If Apple Allows It](https://www.trychert.com) ## What's in the Lab ### [OpenAI Locks Up Brazilian News Content, Reaching 50 Million Users](https://openai.com/index/grupo-folha-grupo-uol-partnership) ## What's in Academe ### [Framework Aims to Turn Vague AI Concepts Like 'Fairness' Into Auditable Standards](https://arxiv.org/pdf/2605.26001v1) ### [AI Response Speed—Not Just Accuracy—Determines How Teams Fail](https://arxiv.org/pdf/2605.25868v1) ### [Showing AI's 'Reasoning' Makes Users Feel Better but Perform Worse](https://arxiv.org/pdf/2605.25856v1) ### [Safety Filters May Block Legitimate Therapy Conversations, Audit Finds](https://arxiv.org/pdf/2605.25454v1) ### [Simulation Platform Trains Phone-Controlling AI Agents Without Physical Devices](https://arxiv.org/pdf/2605.26114v1) ## What's On The Pod --- # 2026-05-25 — Pope Seeks To Reshape Global AI Agenda With Historic Manifesto URL: https://dailyaidigest.net/archive/2026/05/25/ Sources: Hacker News, NBER, arXiv Items: 10 ## What's New ### Pope's 42,000-Word AI Manifesto—Drafted With Anthropic—Sets Global Ethical Agenda ### [George Hotz Warns AI Coding Agents Will Become Enterprise's Costliest Mistake](https://geohot.github.io//blog/jekyll/update/2026/05/24/the-eternal-sloptember.html) ### [Memory Costs Now Drive Two-Thirds of AI Chip Budgets](https://epoch.ai/data-insights/ai-chip-component-cost-shares) ### [Australian Four-Day Workweek Trial: 15 Companies, Zero Productivity Drops](https://scienceaim.com/australia-just-proved-the-four-day-work-week-works-here-is-what-the-data-actually-says/) ### [Why AI Assistants Make Poor System Architects](https://www.hollandtech.net/claude-is-not-your-architect/) ### [DeepSeek Makes 90% Price Cut Permanent, Undercutting OpenAI Long-Term](https://api-docs.deepseek.com/quick_start/pricing) ## What's in Academe ### [Loan Officers Deliberately Avoid AI Explanations When Bonuses Are at Stake](https://www.nber.org/papers/w35240#fromrss) ### [Deep Learning Model Claims 10x Better Crisis Prediction for Financial Regulators](https://www.nber.org/papers/w35227#fromrss) ### [Time-Locked AI Models Could Finally Fix Finance's Backtesting Problem](https://www.nber.org/papers/w35247#fromrss) ### [Persuasive AI Explanations May Hurt Decision-Making, Study Finds](https://arxiv.org/pdf/2605.23867v1) ### [Framework Aims to Standardize What 'Explainable AI' Actually Means](https://arxiv.org/pdf/2605.20081v1) ## What's On The Pod --- # 2026-05-24 — Microsoft to End Claude Code Licenses, Shift Developers to GitHub Copilot CLI URL: https://dailyaidigest.net/archive/2026/05/24/ Sources: Hacker News, arXiv Items: 10 ## What's New ### [Open-Source Tool Lets Teams Run Multiple AI Coding Agents in Parallel](https://www.kanbots.dev/) ### [Microsoft Pulls Claude Code Licenses Despite Developer Preference](https://www.theverge.com/tech/930447/microsoft-claude-code-discontinued-notepad) ### [Forwarding Unedited AI Responses Is the New Rude Email Habit](https://dontquotetheai.com/) ## What's in Academe ### [AI Pipeline Builds Data Visualizations From Plain English Descriptions](https://arxiv.org/pdf/2605.21825v1) ### [Humans Still Beat AI at Complex Strategic Resource Games](https://arxiv.org/pdf/2605.22095v1) ### [Healthcare AI Benchmarks Miss Half the Picture, Researchers Argue](https://arxiv.org/pdf/2605.22612v1) ### [Framework Aims to Make AI Safer for Teens Without Over-Blocking](https://arxiv.org/pdf/2605.21609v1) ### [Audit Finds Safety Gaps in Thousands of Medical AI Chatbots](https://arxiv.org/pdf/2605.20591v1) --- # 2026-05-23 — Anthropic Revenue Stokes Debate: Are LLMs Really Approaching Profitability? URL: https://dailyaidigest.net/archive/2026/05/23/ Sources: Google AI, Hacker News, NBER, OpenAI, Research Blog, TechCrunch, Where's Your Ed At Items: 9 ## What's New ### [Anthropic Revenue Stokes Debate: Are LLMs Really Approaching Profitability?](https://techcrunch.com/2026/05/20/anthropic-says-its-about-to-have-its-first-profitable-quarter/) ### [Anthropic Claims 10,000 Software Vulnerabilities Found in Security Initiative's First Month](https://www.anthropic.com/research/glasswing-initial-update) ### [Japanese Toilet Maker Toto Is Now a Key AI Chip Supplier](https://davidoks.blog/p/why-japanese-companies-do-so-many) ### [Pirate Library Addresses AI Models Directly, Asks Them to Pay for Training Data](https://annas-archive.gl/blog/llms-txt.html) ## What's in the Lab ### [Google Executives Signal AI Priorities: Agents, Quantum, Robotics](https://blog.google/innovation-and-ai/technology/ai/io-2026-dialogues-recap/) ### [OpenAI Touts Codex Enterprise Momentum With Gartner Nod and Virgin Atlantic Case Study](https://openai.com/index/gartner-2026-agentic-coding-leader) ## What's in Academe ### [That $916 AI-Built Operating System? Researchers Say the Demo Obscures Substantial Human Work](https://www.normaltech.ai/p/did-googles-ai-agents-really-build) ### [AI System Claims to Pick Stocks Directly, Reports Sharpe Ratio Above 2 in Backtests](https://www.nber.org/papers/w35195#fromrss) ### [Those AI Job-Impact Studies May Be Getting It Wrong, NBER Paper Argues](https://www.nber.org/papers/w35211#fromrss) ## What's On The Pod --- # 2026-05-22 — You're Probably Overusing AI on Tasks You Could Do Faster Yourself URL: https://dailyaidigest.net/archive/2026/05/22/ Sources: Politico, DeepMind, Google AI, Hacker News, OpenAI, arXiv Items: 11 ## What's New ### [Trump Yanks AI Oversight Order at the Last Minute After David Sacks Intervenes: Politico](https://www.politico.com/news/2026/05/21/trump-ai-order-sacks-00933295) ### [Samsung Chip Workers Win $340K Bonuses After Union Strikes](https://qz.com/samsung-chip-workers-bonus-ai-profits-052126) ### [Waymo Pauses Robotaxi Service in Four Cities After Vehicles Drive Into Floods](https://techcrunch.com/2026/05/21/waymo-pauses-atlanta-service-as-its-robotaxis-keep-driving-into-floods/) ### [Google Tests AI-Generated Ads Woven Into Search Conversations](https://blog.google/products/ads-commerce/google-marketing-live-search-ads/) ## What's in the Lab ### [Hospital System Claims 80% Cut in Admin Time Using ChatGPT](https://openai.com/index/adventhealth) ### [MIT Scientists Use Google AI to Find Genes That May Reverse Aging](https://deepmind.google/blog/fast-tracking-genetic-leads-to-reverse-cellular-aging/) ## What's in Academe ### [You're Probably Overusing AI on Tasks You Could Do Faster Yourself](https://arxiv.org/pdf/2605.22687v1) ### [Chatbot Adapts to Your Thinking Style Instead of Forcing One Approach](https://arxiv.org/pdf/2605.22509v1) ### [Beyond the Org Chart: AI Is Eroding the 'Invisible Work' That Grows Careers](https://arxiv.org/pdf/2605.22707v1) ### [AI Models Fail Up to 47% on Conflict-Sensitive Topics](https://arxiv.org/pdf/2605.22720v1) ### [Chatbots Collapse on Fact-Checks When Prompts Contain False Information](https://arxiv.org/pdf/2605.22785v1) ## What's On The Pod --- # 2026-05-21 — Trump To Sign Huge AI Order As Early As Today URL: https://dailyaidigest.net/archive/2026/05/21/ Sources: Politico, Reuters, Cohere, Google AI, OpenAI, arXiv Items: 14 ## What's New ### [Trump to Sign AI Oversight Order, Asking Labs to Submit Frontier Models for Federal Review](https://www.politico.com/news/2026/05/20/trump-ai-order-details-00930681) ### [AI Model Autonomously Disproves 79-Year-Old Math Conjecture](https://openai.com/index/model-disproves-discrete-geometry-conjecture/) ### [Cohere Releases Open-Source Model for Enterprise AI Agents](https://cohere.com/blog/command-a-plus) ## What's Controversial ### [Rights Groups Accuse Meta of Blocking Activist Accounts in Gulf States](https://www.alqst.org/ar/posts/1190) ## What's in the Lab ### [Google Tests Life-Size Video Displays to Help Remote Workers Feel Present](https://blog.google/innovation-and-ai/models-and-research/google-research/google-beam-group-meetings/) ### [OpenAI Expands School AI Program to Nine Countries](https://openai.com/index/the-next-phase-of-education-for-countries) ### [Ramp Claims OpenAI's Codex Cuts Code Review Time From Hours to Minutes](https://openai.com/index/ramp) ### [Google Cuts AI Ultra Price, Adds $100 Pro Tier at I/O](https://blog.google/products-and-platforms/products/google-one/google-ai-subscriptions/) ### [Cohere Signs Partnerships to Expand Enterprise AI in Europe and Canada](https://cohere.com/blog/cohere-announces-strategic-mous-with-indragroup-and-multiverse-computing) ## What's in Academe ### [Behavioral AI Cuts Mental Health Prediction Errors Up to 58%](https://arxiv.org/pdf/2605.21295v1) ### [Weekly Oral Code Reviews May Let Students Use AI Without Hurting Learning](https://arxiv.org/pdf/2605.21374v1) ### [Open-Source AI Models Obeyed Orders to Inflict Maximum Harm in Milgram-Style Study](https://arxiv.org/pdf/2605.21401v1) ### [83% of Workplace AI Failures Stem From Developer-Worker Mismatches](https://arxiv.org/pdf/2605.21035v1) ### [AI Training Tool Aims to Help Doctors Disclose Medical Errors](https://arxiv.org/pdf/2605.20701v1) ## What's On The Pod --- # 2026-05-20 — Google Reshuffles the Deck for the Entire Internet URL: https://dailyaidigest.net/archive/2026/05/20/ Sources: Google AI, Axios, Anthropic, Hacker News, OpenAI, arXiv Items: 15 ## What's New ### [Google Rebuilds Search Around AI Agents — and Rewrites the Rules of Online Discovery](https://blog.google/products-and-platforms/products/search/search-io-2026/) ### [Pichai Declares the 'Agentic Gemini Era': AI That Works in the Background, On Your Behalf](https://blog.google/innovation-and-ai/sundar-pichai-io-2026/) ### [Trump Order Would Give the Government an Early Look at Powerful AI Models](https://www.axios.com/2026/05/20/ai-trump-executive-order-white-house-infighting) ### [Andrej Karpathy Joins Anthropic in Major Talent Win](https://twitter.com/karpathy/status/2056753169888334312) ## What's Controversial ### [Minnesota Bans Prediction Markets, Triggering Federal Lawsuit Battle](https://www.npr.org/2026/05/19/nx-s1-5821265/minnesota-ban-prediction-markets) ### [Google Cloud Outage Takes Down Startup's Platform Without Warning](https://status.railway.com/?date=20260519) ## What's in the Lab ### [OpenAI Adds Invisible Watermarks to Track AI-Generated Images](https://openai.com/index/advancing-content-provenance) ### [Anthropic Lets Companies Run Claude Agents Inside Their Own Walls](https://claude.com/blog/claude-managed-agents-updates) ### [OpenAI Commits $225 Million to Singapore as First International Hub](https://openai.com/index/introducing-openai-for-singapore) ### [Gemini 3.5 Flash Claims Flagship-Level Performance at Half the Cost](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/) ## What's in Academe ### [Simple Checklist Beats Back-and-Forth Chat for Better AI Results](https://arxiv.org/pdf/2605.20149v1) ### [Fact-Checking Method Shows AI's Reasoning, Not Just Its Verdicts](https://arxiv.org/pdf/2605.20098v1) ### [Dataset Reveals Gap Between What Users Type and What They Actually Want](https://arxiv.org/pdf/2605.20087v1) ### [AI-Judged Training Improves Self-Driving by 12% in Waymo Tests](https://arxiv.org/pdf/2605.20082v1) ### [Small AI Model Matches GPT-4 at Labeling Medical Reports With 32 Examples](https://arxiv.org/pdf/2605.20052v1) ## What's On The Pod --- # 2026-05-19 — OpenAI Survives Its Biggest Challenge URL: https://dailyaidigest.net/archive/2026/05/19/ Sources: Hacker News, OpenAI, arXiv Items: 10 ## What's New ### [Anthropic Acquires the Company Behind Its Developer Tools](https://www.anthropic.com/news/anthropic-acquires-stainless) ### [Simon Willison Pinpoints November 2025 as When AI Coding Agents Became Reliable](https://simonwillison.net/2026/May/19/5-minute-llms/) ### [Open-Source Note App Challenges Obsidian With Plain-File Storage](https://github.com/zakirullin/files.md) ### [Startup Finds Manual Fix for AI Bot Spam Flooding Its Open-Source Project](https://archestra.ai/blog/only-responsible-ai) ## What's Controversial ### [Jury Dismisses Musk's Lawsuit Against OpenAI on Statute of Limitations](https://techcrunch.com/2026/05/18/elon-musk-has-lost-his-lawsuit-against-sam-altman-and-openai/) ## What's in the Lab ### [OpenAI Partners with Dell to Bring Codex to Private Enterprise Servers](https://openai.com/index/dell-codex-enterprise-partnership) ## What's in Academe ### [Vision Paper Proposes AI That Adapts Explanations to Each Viewer in Real Time](https://arxiv.org/pdf/2605.18698v1) ### [Paper Argues AI Chatbot Ads Are Harder to Spot Than Traditional Digital Ads](https://arxiv.org/pdf/2605.18673v1) ### [AI Assistants Echo Your Flawed Reasoning Back to You, Study Finds](https://arxiv.org/pdf/2605.18372v1) ### [AI-Simulated User Preferences Consistently Miss What Real Users Want](https://arxiv.org/pdf/2605.18311v1) ## What's On The Pod --- # 2026-05-18 — Researchers Hand Stock-Picking Entirely to AI — and Report Returns Few Hedge Funds Match URL: https://dailyaidigest.net/archive/2026/05/18/ Sources: Hacker News, NBER, arXiv Items: 7 ## What's New ### [Counterpoint: AI Coding Tools Won't Help If Requirements Are Your Real Bottleneck](https://frederickvanbrabant.com/blog/2026-05-15-i-dont-think-ai-will-make-your-processes-go-faster/) ### [Gruber Dismisses AI-Agent Hype: Apple Sells Experiences, Not Technologies](https://daringfireball.net/2026/05/ai_is_technology_not_a_product) ### [Cost Analysis: Running AI Locally on MacBooks May Be 3x Pricier Than Cloud](https://www.williamangel.net/blog/2026/05/17/offline-llm-energy-use.html) ## What's in Academe ### [Researchers Hand Stock-Picking Entirely to AI — and Report Returns Few Hedge Funds Match](https://www.nber.org/papers/w35195#fromrss) ### [University of Toronto Economist: Standard Methods for Measuring AI's Wage Impact May Get It Backwards](https://www.nber.org/papers/w35211#fromrss) ### [LLM Analysis of 114,000 Claims Shows Media Systematically Amplifies Climate Severity](https://www.nber.org/papers/w35216#fromrss) ### [Researchers Give AI Companions Emotional Memory That Persists Across Conversations](https://arxiv.org/pdf/2605.15812v1) --- # 2026-05-17 — How Open-Weight Models Are Cutting the Cost of Long Documents URL: https://dailyaidigest.net/archive/2026/05/17/ Sources: Hacker News, OpenAI, Research Blog, arXiv Items: 14 ## What's New ### [Malta Will Offer Free ChatGPT Plus to All Citizens Who Complete AI Course](https://openai.com/index/malta-chatgpt-plus-partnership/) ### [Rust-Based Coding Agent Claims Fraction of Competitors' Memory Use](https://crates.io/crates/zerostack/1.0.0) ### [Open-Source Video Generator Promises One-Minute 720p Output](https://nvlabs.github.io/Sana/WM/) ### [Frontier Models Now Solve Most Hacking Challenges, Researcher Claims](https://kabir.au/blog/the-ctf-scene-is-dead) ### [Developer Reportedly Spent $1.3 Million on OpenAI Tokens in 30 Days](https://twitter.com/steipete/status/2055346265869721905) ## What's Controversial ### [AI-Exposed Jobs Show First Signs of Decline, But Effect Remains Modest](https://www.bloomberg.com/news/articles/2026-05-15/us-is-starting-to-see-heavy-job-losses-in-roles-exposed-to-ai) ## What's in the Lab ### [OpenAI Pitches Codex for Executive Briefs, Not Just Code](https://openai.com/academy/codex-for-work/how-business-operations-teams-use-codex) ### [Codex Guide Shows Data Teams a Draft-First Workflow for Reports](https://openai.com/academy/codex-for-work/how-data-science-teams-use-codex) ### [Sales Teams Get OpenAI's Pitch: Codex as Research Assistant](https://openai.com/academy/codex-for-work/how-sales-teams-use-codex) ## What's in Academe ### [How Open-Weight Models Are Cutting the Cost of Long Documents](https://magazine.sebastianraschka.com/p/recent-developments-in-llm-architectures) ### [Walking Coach Study Suggests Emotional AI Beats Pure Information](https://arxiv.org/pdf/2605.14628v1) ### [Grading AI Desktop Agents on a Curve, Not Pass-Fail](https://arxiv.org/pdf/2605.14311v1) ### [AI-Moderated Discussions Create 'Illusion of Inclusion,' Study Finds](https://arxiv.org/pdf/2605.14097v1) ### [Citations Make AI Seem Trustworthy, but Users Rarely Check Them](https://arxiv.org/pdf/2605.14849v1) ## What's On The Pod --- # 2026-05-16 — ChatGPT Starts Connecting To Bank Accounts URL: https://dailyaidigest.net/archive/2026/05/16/ Sources: Cohere, Hacker News, OpenAI, arXiv Items: 13 ## What's New ### [HashiCorp Co-Founder Warns: Some Teams Now Ship Bugs Expecting AI to Fix Them](https://twitter.com/mitchellh/status/2055380239711457578) ### [Satirical Article Highlights JavaScript's Recurring Security Vulnerabilities](https://kevinpatel.xyz/posts/no-way-to-prevent-this/) ### [Free Public Domain Library Upgrades Mobile Experience After 54 Years](https://www.gutenberg.org/) ### [Waymo Recalls 3,800 Robotaxis After Vehicles Drove Into Flooded Streets](https://www.cnbc.com/2026/05/12/waymo-recalls-3800-robotaxis-after-able-drive-into-standing-water.html) ### [Bun Merges Million-Line AI-Generated Rewrite, Drawing Developer Skepticism](https://github.com/oven-sh/bun/pull/30412) ## What's in the Lab ### [ChatGPT Now Connects to Bank Accounts for Personalized Financial Advice](https://openai.com/index/personal-finance-chatgpt) ### [Databricks Offers GPT-5.5 for Enterprise Document Workflows](https://openai.com/index/databricks) ### [Cohere Proposes Measuring AI Success Beyond Traditional ROI](https://cohere.com/blog/enterprise-ai-maturity-model-pt2) ## What's in Academe ### [International Students Turn to ChatGPT as Low-Judgment Guide to American Culture](https://arxiv.org/pdf/2605.15127v1) ### [AI-Generated Interfaces Score High on Usability, Low on Originality](https://arxiv.org/pdf/2605.15124v1) ### [Brief AI Disclosures on News May Distract Readers More Than Detailed Ones](https://arxiv.org/pdf/2605.14999v1) ### [Alibaba Experiment: AI Customer Service Cuts Time but Tanks Satisfaction Ratings](https://arxiv.org/pdf/2605.14830v1) ### [Websites Can Identify Which AI Model Powers Your Browser Agent](https://arxiv.org/pdf/2605.14786v1) ## What's On The Pod --- # 2026-05-15 — Anthropic Issues China Manifesto During Trump Trip URL: https://dailyaidigest.net/archive/2026/05/15/ Sources: Anthropic, Hacker News, NBER, OpenAI, Wired Items: 12 ## What's New ### [With Trump in Beijing, Anthropic Calls for Tighter China Chip Controls](https://www.anthropic.com/research/2028-ai-leadership) ### [MIT Warns Federal Cuts Could Shrink Graduate Programs by 20%](https://president.mit.edu/writing-speeches/video-transcript-message-president-kornbluth-about-funding-and-talent-pipeline) ## What's Controversial ### [Meta Engineers Revolt Over Software That Records Their Screens to Train AI](https://www.wired.com/story/meta-employee-protest-mouse-tracking-surveillance-ai-training/) ### [ArXiv Reportedly Banning Researchers Who Submit AI-Hallucinated Citations](https://twitter.com/tdietterich/status/2055000956144935055) ### [Disable Your Car's Data Collection? Toyota Owner Posts a Tutorial](https://arkadiyt.com/2026/05/13/removing-the-modem-and-gps-from-my-rav4/) ## What's in the Lab ### [ChatGPT's Coding Tool Now Works From Your Phone](https://openai.com/index/work-with-codex-from-anywhere) ### [Shopee's Parent Company Calls AI Coding Tools a 'Structural Multiplier'](https://openai.com/index/sea-david-chen) ### [ChatGPT Now Tracks Conversation History to Flag Mental Health Risks](https://openai.com/index/chatgpt-recognize-context-in-sensitive-conversations) ## What's in Academe ### [GPT-5.4 Predicts What People Want in Life Better Than Other Humans Can](https://www.nber.org/papers/w35185#fromrss) ### [Economists Build Research Datasets for Pennies Using AI Agents](https://www.nber.org/papers/w35188#fromrss) ### [Degree Requirements in AI Hiring Tools May Face Legal Challenges](https://www.nber.org/papers/w35192#fromrss) ### [Data Centers Boost Local Jobs and Wages but Raise Electricity Costs](https://www.nber.org/papers/w35194#fromrss) ## What's On The Pod --- # 2026-05-14 — Claude Courts Small Business, Turns Screws On Automated Power-Users URL: https://dailyaidigest.net/archive/2026/05/14/ Sources: Hacker News, The Verge, Anthropic, OpenAI, arXiv Items: 11 ## What's New ### [Claude Pitches Small Business On Pre-Wired Workflows](https://www.anthropic.com/news/claude-for-small-business) ### [Claude Subscriptions Will Stop Subsidizing Heavy Automation](https://support.claude.com/en/articles/15036540-use-the-claude-agent-sdk-with-your-claude-plan) ### [Hacker News User Says Canceling a Claude Plan Locked Them Out of Their Projects](https://news.ycombinator.com/item?id=48128003) ### [Meta's AI Account on Threads Can't Be Blocked, Sparking Backlash](https://www.theverge.com/tech/929091/meta-ai-threads-account-block) ### [Why the US Leads AI: Cloud Lock-In Matters More Than Cheap Energy](https://avkcode.github.io/blog/us-winning-ai-race.html) ## What's in the Lab ### [OpenAI Discloses Supply Chain Breach That Exposed Code-Signing Certificates](https://openai.com/index/our-response-to-the-tanstack-npm-supply-chain-attack) ## What's in Academe ### [Canary Tokens Could Expose Which AI Models Scraped Your Website](https://arxiv.org/pdf/2605.13706v1) ### ["Human in the Loop" Often Means Less Than Vendors Claim, Paper Argues](https://arxiv.org/pdf/2605.13723v1) ### [2024 Election Disinformation Shifted From Bot Amplification to AI-Generated Originals](https://arxiv.org/pdf/2605.13785v1) ### [AI Models Trained on Flagged Misinformation May Learn It as True](https://arxiv.org/pdf/2605.13829v1) ### [Statistical Framework Aims to Fix Unreliable Human Ratings in AI Benchmarks](https://arxiv.org/pdf/2605.13801v1) ## What's On The Pod --- # 2026-05-13 — Commerce vs. the Spies: A 'Knife Fight' Over AI Policy in Trump's White House URL: https://dailyaidigest.net/archive/2026/05/13/ Sources: Google AI, Hacker News, BBC, OpenAI, Reuters, Washington Post, arXiv Items: 12 ## What's New ### [Google Joins Anthropic in Eyeing SpaceX for AI Data Centers in Space](https://www.reuters.com/science/google-spacex-talks-explore-data-centers-orbit-wsj-reports-2026-05-12/) ### [Google Announces AI-First Laptop With Built-In Gemini Features](https://googlebook.google/) ### [26-Million Parameter Model Claims to Match Giants at Tool Calling](https://github.com/cactus-compute/needle) ## What's Controversial ### ['Knife Fight' in the Trump White House: Commerce vs. the Spies Over Who Tests AI](https://www.washingtonpost.com/politics/2026/05/11/trump-ai-regulation-commerce-intelligence/) ### [Altman: Musk Wanted Then-Nonprofit OpenAI for His Children — and as a Tesla Subsidiary](https://www.bbc.com/news/articles/czj2k2exdzlo) ## What's in the Lab ### [NVIDIA Says 40,000 Employees Now Use OpenAI's Codex for Engineering Work](https://openai.com/index/nvidia) ### [ML Competition Reveals How AI Coding Tools Are Reshaping Research](https://openai.com/index/what-parameter-golf-taught-us) ## What's in Academe ### [AI-Generated Political Posts Sound Fluent but Flatten Real Human Debate](https://arxiv.org/pdf/2605.12452v1) ### [Tiny AI Model Outperforms Giant Rival at Catching Fabricated Information](https://arxiv.org/pdf/2605.12384v1) ### [Safety Monitors Miss Dangerous AI Behavior in Longer Coding Sessions](https://arxiv.org/pdf/2605.12366v1) ### [New Protocol Aims to Test AI Security Agents on Real Vulnerabilities](https://arxiv.org/pdf/2605.10834v1) ### [Claude Executed Risky Actions Even While Verbally Refusing Them, Study Finds](https://arxiv.org/pdf/2605.10779v1) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-12 — Frontier AI Labs Keep Becoming Consulting Firms URL: https://dailyaidigest.net/archive/2026/05/12/ Sources: Google AI, Hacker News, OpenAI, Reuters, arXiv Items: 10 ## What's New ### [OpenAI Builds Its Own $4B Consulting Arm, a Week After Its PwC Deal](https://openai.com/index/openai-launches-the-deployment-company) ### Claude Code Gets a Dashboard for Managing Multiple AI Agents ### [GitLab Cuts Jobs, Cites Shift to AI-Driven Development Era](https://about.gitlab.com/blog/gitlab-act-2/) ### [AI Coding Tools May Work Better With Strict Languages Like Rust](https://medium.com/@NMitchem/if-ai-writes-your-code-why-use-python-bf8c4ba1a055) ## What's Innovative ### [Developer Built Custom Sleep Monitor in Hours With AI Coding Help](https://martin.sh/i-let-ai-build-a-tool-to-help-me-figure-out-what-was-waking-me-up-at-night/) ## What's Controversial ### [Sutskever Testifies Altman Showed 'Consistent Pattern of Lying'; Board Pursued Anthropic Merger After 2023 Ouster](https://www.reuters.com/business/former-openai-executive-sutskever-discloses-nearly-7-billion-stake-ai-firm-2026-05-11/) ## What's in the Lab ### [ChatGPT's User Base Now Majority Women, Fastest Growth in Emerging Markets](https://openai.com/signals/research/2026q1-update) ### [Google Finance Adds AI Research Tools, Expands Across Europe](https://blog.google/products-and-platforms/products/search/ai-powered-google-finance-in-europe/) ## What's in Academe ### [Technique Helps AI Adapt Responses to Users' Cultural Backgrounds](https://arxiv.org/pdf/2605.10843v1) ### [Best AI Agents Still Fail 40% of Real-World Tasks, Benchmark Shows](https://arxiv.org/pdf/2605.10912v1) ### [Method Detects When AI Vision Tools Guess Instead of Actually Looking](https://arxiv.org/pdf/2605.10893v1) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-11 — GPT-5.4 Predicts Your Preferences Better Than Other Humans Do, Study Finds URL: https://dailyaidigest.net/archive/2026/05/11/ Sources: Hacker News, NBER, OpenAI, arXiv Items: 10 ## What's New ### [The Case for On-Device AI: Privacy and Reliability vs. Raw Capability](https://unix.foo/posts/local-ai-needs-to-be-norm/) ### [Maryland Ratepayers Challenge $2B Grid Upgrade That Benefits Out-of-State Data Centers](https://www.tomshardware.com/tech-industry/artificial-intelligence/maryland-citizens-slapped-with-usd2-billion-grid-upgrade-bill-for-out-of-state-ai-data-centers-state-complains-to-federal-energy-regulators-says-additional-cost-breaks-ratepayer-protection-pledge-promises) ## What's Innovative ### [Best Local AI Model for M4 MacBooks Depends on Memory Trade-Offs](https://jola.dev/posts/running-local-models-on-m4) ## What's in the Lab ### [OpenAI Recruits Student Clubs for Campus Partnership Program](https://openai.com/index/openai-campus-network-student-club-interest-form) ### [Enterprise AI Success Depends on Trust and Governance, Not Speed, Executives Say](https://openai.com/business/guides-and-resources/how-enterprises-are-scaling-ai) ## What's in Academe ### [GPT-5.4 Predicts Your Preferences Better Than Other Humans Do, Study Finds](https://www.nber.org/papers/w35185#fromrss) ### [AI Agents Now Build Research Datasets That Once Required Human Assistants](https://www.nber.org/papers/w35188#fromrss) ### [AI Hiring Tools That Screen Out Non-Degree Holders May Violate Civil Rights Law](https://www.nber.org/papers/w35192#fromrss) ### [Data Centers Boost Local Jobs and Tax Revenue but Raise Electricity Prices](https://www.nber.org/papers/w35194#fromrss) ### [Researchers Propose Standard Framework for Comparing AI Tools in Business Settings](https://arxiv.org/pdf/2605.07986v1) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-10 — Each AI Edit Quietly Corrupts Your Documents, Study Finds URL: https://dailyaidigest.net/archive/2026/05/10/ Sources: Hacker News, NBER Items: 11 ## What's New ### [Meta's AI Push Reportedly Taking Toll on Employee Morale](https://www.nytimes.com/2026/05/08/technology/meta-ai-employees-miserable.html) ### [Internet Archive Opens Swiss Nonprofit to Preserve AI Models and Global Archives](https://blog.archive.org/2026/05/06/internet-archive-switzerland-expanding-a-global-mission-to-preserve-knowledge/) ### [Each AI Edit Quietly Corrupts Your Documents, Study Finds](https://arxiv.org/abs/2604.15597) ### [Fields Medalist Says ChatGPT Solved Research-Level Math Problems in an Hour](https://gowers.wordpress.com/2026/05/08/a-recent-experience-with-chatgpt-5-5-pro/) ### [Developer Builds Fully Functional Web Server in Raw Assembly Code](https://github.com/imtomt/ymawky) ## What's in Academe ### [Census Data Shows 18% of U.S. Firms Use AI, Mostly for Augmentation Not Job Cuts](https://www.nber.org/papers/w35141#fromrss) ### [LLMs Skew Their Analysis When They Can Guess Your Goal](https://www.nber.org/papers/w35142#fromrss) ### [AlphaFold Boosted Basic Science but Drug Discovery Hasn't Followed](https://www.nber.org/papers/w35143#fromrss) ### [AI Stock Portfolios Perform No Better Than Market Benchmarks, Study Finds](https://www.nber.org/papers/w35153#fromrss) ### [Economic Model Predicts AI Could Trigger Explosive Growth Within Six Years](https://www.nber.org/papers/w35155#fromrss) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-09 — Anthropic Declares Breakthrough Against LLM Misbehavior URL: https://dailyaidigest.net/archive/2026/05/09/ Sources: Anthropic, Google AI, Hacker News, NBER, OpenAI, arXiv Items: 15 ## What's New ### [Google's reCAPTCHA Update Locks Out Privacy-Focused Android Users](https://reclaimthenet.org/google-broke-recaptcha-for-de-googled-android-users) ### [Poland Joins World's 20 Largest Economies, Eyes G20 Invitation](https://apnews.com/article/poland-economy-growth-g20-gdp-26fe06e120398410f8d773ba5661e7aa) ### [AI Can Now Spot Security Flaws in Code Commits, Compressing Patch Windows](https://www.jefftk.com/p/ai-is-breaking-two-vulnerability-cultures) ### [Does AI Art Signal Laziness? One Argument Says Yes, But Evidence Is Thin](https://mccue.dev/pages/5-8-26-ai-art) ### [David Attenborough Turns 100, Drawing Global Tributes](https://www.bbc.com/news/articles/cp3pww9g0p5o) ## What's Controversial ### [OpenAI Publishes Privacy Guide Aimed at Enterprise Buyers and Regulators](https://openai.com/index/how-chatgpt-protects-privacy) ## What's in the Lab ### [Anthropic Declares Breakthrough Against LLM Misbehavior](https://www.anthropic.com/research/teaching-claude-why) ### [Google Pairs Ad Veterans with Small Businesses to Showcase AI Creative Tools](https://blog.google/company-news/inside-google/company-announcements/the-small-brief/) ### [OpenAI Shares Blueprint for Running AI Coding Agents Safely](https://openai.com/index/running-codex-safely) ### [Consulting Firm Claims 70% Faster Development Using OpenAI's Codex](https://openai.com/index/simplex) ## What's in Academe ### [Heavy AI Reliance Could Narrow Everyone's Thinking, Researchers Warn](https://arxiv.org/pdf/2605.06347v1) ### [Expert Mathematicians Don't Ask AI Questions—They Think Out Loud With It](https://arxiv.org/pdf/2605.05921v1) ### [Visual Fingerprints Reveal Hidden Patterns in How AI Models Write](https://arxiv.org/pdf/2605.06054v1) ### [Smart Glasses Could Track Mental Fatigue Through Eye Movement](https://arxiv.org/pdf/2605.05790v1) ### [Automation's Hidden Risk: Workers Losing the Jobs That Teach Them Skills](https://www.nber.org/papers/w35157#fromrss) ## What's Happening on Capitol Hill --- # 2026-05-08 — OpenAI Releases Real-Time Voice Models for Live AI Conversations URL: https://dailyaidigest.net/archive/2026/05/08/ Sources: Politico, Wall Street Journal, Hacker News, OpenAI, arXiv Items: 11 ## What's New ### [White House Distances Itself From Tighter AI Regulation](https://www.politico.com/news/2026/05/07/white-house-ai-oversight-00910837) ### [ChatGPT Can Now Alert a Trusted Contact If You Discuss Self-Harm](https://openai.com/index/introducing-trusted-contact-in-chatgpt) ### [Opinion: Low-Quality AI Content May Be Driving Away Online Community Members](https://rmoff.net/2026/05/06/ai-slop-is-killing-online-communities/) ### [Google's AlphaEvolve Moves from Demo to Production, Designing Chips and Optimizing Databases](https://deepmind.google/blog/alphaevolve-impact/) ## What's Controversial ### [Cloudflare Cuts 20% of Staff Despite Strong Revenue, Cites AI Productivity](https://www.reuters.com/business/world-at-work/cloudflare-cut-over-1100-jobs-2026-05-07/) ## What's in the Lab ### [OpenAI Releases Real-Time Voice Models for Live AI Conversations](https://openai.com/index/advancing-voice-intelligence-with-new-models-in-the-api) ## What's in Academe ### [Researchers Map How Users Game AI Content Filters](https://arxiv.org/pdf/2605.06619v1) ### [Which AI Is 'Safer' Depends Entirely on What You Test, Study Finds](https://arxiv.org/pdf/2605.06652v1) ### [AI Research Tools Get Citations Wrong Up to 61% of the Time](https://arxiv.org/pdf/2605.06635v1) ### [AI Memory Systems Struggle to Recognize When Stored Information Becomes Outdated](https://arxiv.org/pdf/2605.06527v1) ### [Global AI Leaderboards May Be Statistically Meaningless, Study Claims](https://arxiv.org/pdf/2605.06656v1) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-07 — Anthropic Strikes Stunning Compute Deal With Musk's SpaceX URL: https://dailyaidigest.net/archive/2026/05/07/ Sources: Anthropic, OpenAI, Google, arXiv, Hacker News Items: 11 ## What's New ### [Anthropic Strikes Stunning Compute Deal With Musk's SpaceX](https://www.anthropic.com/news/higher-limits-spacex) ### [Claude Code Rate Limits Doubled; Peak-Hours Throttle Lifted](https://www.anthropic.com/news/higher-limits-spacex) ### [AI Makes It Easy to Fake Expertise—and Harder for Bosses to Spot](https://nooneshappy.com/article/appearing-productive-in-the-workplace/) ### [Prominent Developer Admits He's Stopped Reviewing AI-Generated Code](https://simonwillison.net/2026/May/6/vibe-coding-and-agentic-engineering/) ## What's Innovative ### [Google Demonstrates AI Search Tools for Garden Planning](https://blog.google/products-and-platforms/products/search/gardening-tips/) ## What's in the Lab ### [Uber Partners with OpenAI to Help Drivers Optimize Earnings](https://openai.com/index/uber) ### [Top AI-Using Firms Pull Further Ahead, OpenAI Research Claims](https://openai.com/index/introducing-b2b-signals) ## What's in Academe ### [Most Developers Accept AI-Generated Code Without Editing, Study Finds](https://arxiv.org/pdf/2605.04835v1) ### [Patent Language May Reveal Tech Breakthroughs Decades Early](https://arxiv.org/pdf/2605.04875v1) ### [AI Coding Assistants Boost Speed Modestly but Don't Improve Skills](https://arxiv.org/pdf/2605.04779v1) ### [AI Safety Training Can Backfire, Rewarding Harmful Responses](https://arxiv.org/pdf/2605.05003v1) ## What's Happening on Capitol Hill ## What's On The Pod --- # 2026-05-06 — White House Plans New Powers Over AI Releases URL: https://dailyaidigest.net/archive/2026/05/06/ Sources: Politico, NIST, OpenAI, arXiv, Hacker News Items: 10 ## What's New ### [White House Plans New Powers Over AI Releases](https://www.politico.com/news/2026/05/05/white-house-mulls-tight-new-controls-on-advanced-ai-00907468) ### [Google Claims 3x Speed Boost for Open-Source Gemma Models](https://blog.google/innovation-and-ai/technology/developers-tools/multi-token-prediction-gemma-4/) ### [Chrome Reportedly Downloads 4 GB AI Model Without User Consent](https://www.thatprivacyguy.com/blog/chrome-silent-nano-install/) ### [Three Rules for Working With AI: Don't Anthropomorphize, Don't Blindly Trust, Stay Accountable](https://susam.net/inverse-laws-of-robotics.html) ## What's Controversial ### [Publishers Allege Zuckerberg Personally Approved Pirated Books for Llama Training](https://variety.com/2026/digital/news/meta-ai-mark-zuckerberg-copyright-infringement-lawsuit-publishers-scott-turow-1236738383/) ## What's in the Lab ### [ChatGPT's New Default Model Claims 50% Fewer Errors on Complex Questions](https://openai.com/index/gpt-5-5-instant) ### [ChatGPT Opens Self-Serve Ads Platform to US Businesses](https://openai.com/index/new-ways-to-buy-chatgpt-ads) ## What's in Academe ### [All Five LLMs Tested Show Gender Bias in ER Triage Decisions](https://arxiv.org/pdf/2605.03998v1) ### [Bigger Medical AI Models Aren't Automatically Safer, 34-Model Study Finds](https://arxiv.org/pdf/2605.04039v1) ### [AI Symptom Interviews Outperformed Clinicians in 14,000-Patient Study](https://arxiv.org/pdf/2605.04012v1) ## What's On The Pod --- # 2026-05-05 — ChatGPT and Claude Come For Consulting URL: https://dailyaidigest.net/archive/2026/05/05/ Sources: OpenAI, Anthropic, Google AI, Import AI, arXiv Items: 8 ## What's New ### [OpenAI Partners with PwC to Embed AI Agents in the Office of the CFO](https://openai.com/index/openai-pwc-finance-collaboration) ### [Anthropic Co-Founds a New Enterprise AI Services Firm with Blackstone, Goldman Sachs, and Hellman & Friedman](https://www.anthropic.com/news/enterprise-ai-services-company) ## What's Controversial ### [The Big Shift Is Near: AIs Building Themselves, Writes Anthropic Co-Founder](https://importai.substack.com/p/import-ai-455-automating-ai-research) ## What's in the Lab ### [Google Bets Big on AI Agents with New Enterprise Platform, TPUs, and Open Model](https://blog.google/innovation-and-ai/technology/ai/google-ai-updates-april-2026/) ### [Gemini API Adds Webhooks for Long-Running Tasks](https://blog.google/innovation-and-ai/technology/developers-tools/event-driven-webhooks/) ## What's in Academe ### [Making Software Agent-Friendly Could Matter as Much as Making Agents Smarter](https://arxiv.org/pdf/2605.02729v1) ### [Surprising Finding: Stricter AI Oversight Improved Performance and Cut Worker Fatigue](https://arxiv.org/pdf/2605.02832v1) ### [AI Tutor Predicts Pair Programming Breakdowns 30 Seconds Before They Happen](https://arxiv.org/pdf/2605.02703v1) ## What's On The Pod --- # 2026-05-04 — AI Is Changing Work, Not Replacing It — At Least For Now URL: https://dailyaidigest.net/archive/2026/05/04/ Sources: Hacker News, NBER Items: 5 ## What's New ### [Opinion: AI Coding Tools May Be Eroding Developer Skills](https://larsfaye.com/articles/agentic-coding-is-a-trap) ### [Open Source Project Claims Claude-Level Coding at 17x Lower Cost](https://github.com/aattaran/deepclaude) ## What's in Academe ### [AI Is Changing Work, Not Replacing It — At Least For Now: Three New Papers](https://www.nber.org/papers/w35141#fromrss) ### [Telling AI Your Goal Can Bias Its Outputs, Researchers Find](https://www.nber.org/papers/w35142#fromrss) ### [AI Stock Pickers Chase Headlines and Underperform, Study Finds](https://www.nber.org/papers/w35153#fromrss) --- # 2026-05-03 — Personalized AI tutoring dramatically improved results for at-risk students in computer science, study finds URL: https://dailyaidigest.net/archive/2026/05/03/ Sources: Hacker News, Hugging Face Papers, arXiv Items: 10 ## What's New ### [VS Code Reportedly Tags Commits as AI-Assisted Even When Copilot Wasn't Used](https://github.com/microsoft/vscode/pull/310226) ### [Chinese Open-Weights Model Beats Claude and GPT-5.5 in Coding Competition](https://thinkpol.ca/2026/04/30/an-open-weights-chinese-model-just-beat-claude-gpt-5-5-and-gemini-in-a-programming-challenge/) ### [GitHub Project Claims to Turn Claude Into a Design Tool—But Community Is Skeptical](https://github.com/nexu-io/open-design) ### [Starlink Terminals Smuggled Into Iran to Bypass Three-Month Internet Blackout](https://www.bbc.com/news/articles/cvgzk91leweo) ### [Maryland Becomes First State to Ban AI-Driven Grocery Price Increases](https://www.nytimes.com/2026/05/01/business/surveillance-pricing-groceries-maryland.html) ## What's in Academe ### [Text Analysis Tool Maps Who Did What to Whom in Customer Feedback](https://arxiv.org/pdf/2604.27673v1) ### [General AI Assistants Outperform Specialized Tools on Data Visualization—But Cost More](https://arxiv.org/pdf/2604.27996v1) ### [Personalized AI Exercises Cut Student Dropout From 30% to Under 1%](https://arxiv.org/pdf/2604.27433v1) ### [NVIDIA Opens Multimodal Model That Handles Audio, Video, and Text Together](https://huggingface.co/papers/2604.24954) ### [Proposed Architecture Would Use Cheaper AI for Routine Desktop Tasks](https://huggingface.co/papers/2604.27151) ## What's On The Pod --- # 2026-05-02 — Burning Through The Token Budget: The Case Of Uber URL: https://dailyaidigest.net/archive/2026/05/02/ Sources: Hacker News, Meta AI, NBER, arXiv Items: 11 ## What's New ### [Uber Reportedly Burned 2026 AI Budget in Four Months on Coding Tools](https://www.briefs.co/news/uber-torches-entire-2026-ai-budget-on-claude-code-in-four-months/) ### [Users Claim LGBTQ-Themed Prompts Can Bypass AI Safety Filters](https://github.com/Exocija/ZetaLib/blob/main/The%20Gay%20Jailbreak/The%20Gay%20Jailbreak.md) ### [Leaked Files Confirm Apple Uses Anthropic's Claude for Internal Development](https://x.com/aaronp613/status/2049986504617820551) ### [OpenAI Launches Restricted Cybersecurity Tool After Criticizing Anthropic's Similar Approach](https://techcrunch.com/2026/04/30/after-dissing-anthropic-for-limiting-mythos-openai-restricts-access-to-cyber-too/) ### [PyTorch Lightning Package Compromised With Credential-Stealing Malware](https://semgrep.dev/blog/2026/malicious-dependency-in-pytorch-lightning-used-for-ai-training/) ## What's in the Lab ### [Meta Adds Verifiable Security to Encrypted Message Backups](https://engineering.fb.com/2026/05/01/security/meta-strengthening-end-to-end-encrypted-backups/) ## What's in Academe ### [Government Identity Systems Fail Blind Users, Study Finds](https://arxiv.org/pdf/2604.28166v1) ### [AI Job-Risk Scores Flip Conclusions Depending on Which Model Runs Them](https://www.nber.org/papers/w35110#fromrss) ### [Conflicting AI Productivity Research Traced to Measurement Mismatches](https://www.nber.org/papers/w35123#fromrss) ## What's On The Pod --- # 2026-05-01 — ChatGPT Adds High-Security Mode for Journalists, Officials, and Dissidents URL: https://dailyaidigest.net/archive/2026/05/01/ Sources: DeepMind, Hacker News, OpenAI, arXiv Items: 12 ## What's New ### [ChatGPT Adds High-Security Mode for Journalists, Officials, and Dissidents](https://openai.com/index/advanced-account-security) ### [Rivian Lets Owners Disable Vehicle Data Collection—If They Schedule a Service Appointment](https://rivian.com/support/article/can-i-disable-all-data-collection-from-my-vehicle) ### [Belgium Reverses Nuclear Phase-Out, May Nationalize Reactors](https://dpa-international.com/general-news/urn:newsml:dpa.com:20090101:260430-930-14717/) ## What's in the Lab ### [Google Tests AI That Could Talk Directly With Patients](https://deepmind.google/blog/ai-co-clinician/) ### [OpenAI Publishes Values Manifesto on AGI Development](https://openai.com/index/our-principles) ## What's Controversial ### [Unverified Claim Suggests Claude Code Filters Competitor Keywords](https://twitter.com/theo/status/2049645973350363168) ## What's in Academe ### [AI-Filtered Comments Could Turn Social Media Into a Critical Thinking Tool](https://arxiv.org/pdf/2604.27905v1) ### [Framework Aims to Detect Compounding Bias Across Race, Gender, and Other Attributes](https://arxiv.org/pdf/2604.28030v1) ### [Google's AI Answers Draw From Different Sources Than Traditional Search Results](https://arxiv.org/pdf/2604.27790v1) ### [AI System Aims to Turn Economists' Hunches Into Full Research Studies](https://arxiv.org/pdf/2604.27725v1) ### [Gamers Welcome AI for Difficulty and NPCs, Resist It for Art and Moderation](https://arxiv.org/pdf/2604.27812v1)