July 2026
- Jul 20 MON AI News Altman email defunds rivals, Nvidia locks Japan into CUDA, Current AI at 1% 12 min
- Jul 20 MON AI Research GenCeption cuts vision data 500×, Qwen two-hop hits 11%, Soofi S tops German 12 min
- Jul 19 SUN AI News 28% of execs skip AI, Medicare WISeR stretches waits, Lila chases $8.5B 11 min
- Jul 19 SUN AI Tech Claude Code ships AI-ported Bun, NeMo wires Diffusers, Willison annotates SQLite 11 min
- Jul 18 SAT AI News Databricks hits $188B, OpenAI skips SWE-Bench Pro, Patreon blocks scrapers 13 min
- Jul 18 SAT AI Research Kimi K3's costly refusal leaks its prompt, its jailbreak, and a Claude tell 3 min
- Jul 18 SAT AI Tech Willison flags LLM tics, Forrester finds 88% of agent pilots fail, Quixote v3.8 9 min
- Jul 17 FRI AI News Kimi K3 needs 1.4TB, Hugging Face breach hits 17K, DeepMind closes IsoDDE 12 min
- Jul 17 FRI AI Research Meta cues Claude reminders, HKU caps agents at 47.5%, NVIDIA hits 128K on Qwen3 12 min
- Jul 17 FRI AI Tech Moonshot hikes Kimi K3 4×, Codex wipes $HOME dirs, NVIDIA leads RTEB 12 min
- Jul 16 THU AI Research GPT-Red hits 0.05%, LingBot slips to 480p, Meta's SDM runs 1.49× slower 12 min
- Jul 16 THU AI Tech xAI open-sources Grok, Claude fetch leaks memory, Inkling leads US open weights 12 min
- Jul 16 THU AI News OpenAI's $230 macropad, Thinking Machines opens Inkling, Suno leak names YouTube 14 min
- Jul 15 WED AI Research Canada 2nd on Claude use, Gemma 4 ditches encoder, Alibaba ships RynnWorld-4D 12 min
- Jul 15 WED AI Tech Dependabot delays 72h, Lobsters moves to SQLite, Ronacher flags agent skill loss 11 min
- Jul 15 WED AI News NY freezes big data centers, 26 ex-Meta staff sue over AI, Claude enters K-12 14 min
- Jul 14 TUE AI Research Anthropic bounds Claude on robots and values, Microsoft slices the AI scientist 13 min
- Jul 14 TUE AI News Apple sues OpenAI, DeepMind pilots Gemini in Indian schools, Codex hits 7M 14 min
- Jul 14 TUE AI Tech Tracebit hits 0% attack rate, DOOMQL runs in SQLite, METR clocks devs 19% slower 11 min
- Jul 13 MON AI Research MIPU drops half its updates, LLMs default to bridge ideas, Tencent opens scanner 12 min
- Jul 13 MON AI News White House caps open weights, Anthropic rations Fable 5, Apple readies M7 Ultra 12 min
- Jul 13 MON AI Tech Willison bars AI ownership, sqlite-utils blocks cascade, shot-scraper runs JS 10 min
- Jul 12 SUN AI News EU readies $12B Meta fine, OpenAI hires Families PM, Ghostcommit hits Cursor 10 min
- Jul 12 SUN AI Tech sqlite-utils 4.1 rebuilds tables to fake a STRICT ALTER, most PRs by Codex 4 min
- Jul 11 SAT AI Research Anthropic's Jacobian lens reads Claude's intent; training on it hides reasoning 3 min
- Jul 11 SAT AI News Apple sues OpenAI, SAG-AFTRA pulls Meta's Muse, Lutnick threatens SK Hynix 13 min
- Jul 11 SAT AI Tech Patel: real AR glasses need cloud-streamed cameras as bystanders push back 4 min
- Jul 10 FRI AI Tech GPT-5.6 leads under cheat flag, Muse Spark trails claim, vLLM parity drifts 12 min
- Jul 10 FRI AI News GPT-5.6 hacks evals, Reflect nudges retention, investors override Bernanke Trust 13 min
- Jul 10 FRI AI Research PAW echoes Sakana, WARP merges beat oracle, GRPO tops SDPO variants 13 min
- Jul 9 THU AI News GPT-Live trails Moshi, Grok 4.5 undercuts GPT-5.5, OpenAI softens Anthropic line 11 min
- Jul 9 THU AI Research OpenAI retracts SWE-Bench Pro, Anthropic ships GRAM, GRPO variants reduce to one 14 min
- Jul 9 THU AI Tech GPT-Live wraps GPT-5.5, Bun hosts Claude Code, Deep Agents lift Nemotron 12 min
- Jul 8 WED AI Tech Azure rents HF at $8/hr, SkyPilot zeros egress, sqlite-utils 4.0 skips rollback 12 min
- Jul 8 WED AI Research Dockerless trains without tests, TRIAGE trails GiGPO, LUMOS ships no benchmarks 12 min
- Jul 8 WED AI News Gemini hits agent parity, Meta opts Instagram into Muse, DeepSeek designs a chip 12 min
- Jul 7 TUE AI Research Anthropic maps reasoning, OSWorld 2.0 caps at 20.6%, SWE-Together flips ranks 13 min
- Jul 7 TUE AI Tech HF signs kernels after RCE, LeRobot patches pickle RCE, Photoroom halves FID 11 min
- Jul 7 TUE AI News Microsoft cuts 4,800 jobs, Alberta scans 466M lines, Tencent opens Hy3 12 min
- Jul 6 MON AI News MTurk closes signups, IXL bans Alpha School, Gemini kept off Declaration text 11 min
- Jul 6 MON AI Research Qwen-Robot withholds weights, PAT meets ICML ban, ConvFill breaches safety cap 13 min
- Jul 6 MON AI Tech sqlite-utils 4.0rc3 swaps foreign_keys to a namedtuple, breaks old indexing 6 min
- Jul 5 SUN AI News Alibaba bans Claude Code, Midjourney appeals discovery, AO3 detector misfires 11 min
- Jul 5 SUN AI Tech Fable holds only for Willison, Opus 4.8 only in Claude Code, map only as ASCII 11 min
- Jul 4 SAT AI News A24 bars Veo training, Midjourney sidesteps FDA, Comeau loses sales to LLMs 10 min
- Jul 4 SAT AI Tech GLM-5.2 tops open weights, Gap Map grades 421 repos, Fable subagents cost more 11 min
- Jul 3 FRI AI News OpenAI floats 5%, Microsoft embeds 6,000, AI Engineer backs constrained loops 13 min
- Jul 3 FRI AI Research Router cap is co-failure, Qwen flips coding verify, tool-RL blames formatting 14 min
- Jul 3 FRI AI Tech Willison clones Claude Code, GEPA overfits Datasette, Anthropic splits defenders 13 min
- Jul 2 THU AI Tech Cerebras hits 1,851 tok/s on Gemma 4, but the voice-agent bill is TTS not LLM 4 min
- Jul 2 THU AI News Commerce arms-controls weights, Cloudflare gates crawlers, AI Engineer dissents 14 min
- Jul 2 THU AI Research Snowflake logs plan decay, Princeton reads refusal pre-CoT, grammars zero tools 11 min
- Jul 1 WED AI News Sonnet 5 costs more per task, Fable 5 ban lifts, Claude Science skips biosafety 13 min
- Jul 1 WED AI Research Sol games GeneBench, Claude Code overstates builds, Qwen 35B skips tool calls 13 min
- Jul 1 WED AI Tech Sonnet 5 inflates tokens 30%, OpenAI omits prompt leak, Nano Lite drops images 11 min
June 2026
- Jun 30 TUE AI News Korea, OpenAI, and Google UK each pitch AI numbers undercut by their own data 11 min
- Jun 30 TUE AI Research Nemotron fakes cryptarithm at 7%, Ai2 beats weak KDE 37×, GQE prefills 1.8× 12 min
- Jun 30 TUE AI Tech Ornith scores on own harness, Antigravity 2.0 ships broken, Willison clears 200 11 min
- Jun 29 MON AI Research GateMem flunks GPT-5.4, Reflective Masking re-masks, PerceptionDLM parallelizes 14 min
- Jun 29 MON AI News Palisades jury hangs on ChatGPT logs, GLM-5.2 ties Mythos, Zyphra ships on AMD 10 min
- Jun 29 MON AI Tech Udell rejects 'human in the loop' as AI-driven PR reviews stretch 91% longer 3 min
- Jun 28 SUN AI News Apple loses Vision VP to OpenAI and hikes MacBook $300; Anthropic cedes China 10 min
- Jun 28 SUN AI Tech Anthropic embeds Claude Tag in Slack with reasoning hidden and OAuth inherited 3 min
- Jun 27 SAT AI Research Anthropic's June AEI reads as a Claude product report, not a labor study 4 min
- Jun 27 SAT AI Tech Opus 4.6 blanks 6,000 injections, Nesbitt parodies $41K loops, IBM bonds 0.7nm 13 min
- Jun 27 SAT AI News GPT-5.6 Sol gated, Mythos 5 returns vetted, OpenAI's Ball pitches deregulation 14 min
- Jun 26 FRI AI Tech HF Jobs runs vLLM per-minute, Ai2 ships Olmo diagnostic, Datasette plugin unpins 11 min
- Jun 26 FRI AI Research Moebius beats FLUX at 0.22B, FID swings 3.2× on retraining, Ant blames E2M1 12 min
- Jun 26 FRI AI News Munich binds Google, White House gates OpenAI, Anthropic accuses Alibaba 12 min
- Jun 25 THU AI News OpenAI's Jalapeño doubted, Gemini Flash hits 88.4%, DeepMind bleeds to Anthropic 14 min
- Jun 25 THU AI Tech NeMo triples MoE throughput, meta-harnesses lift Qwen3.5 44%, Cohere leads FFASR 11 min
- Jun 25 THU AI Research OpenAI omits agent reliability, MaineCoon hides weights, Kairos needs 23.5GB 12 min
- Jun 24 WED AI News Anthropic forces Claude into Slack, GPT-5 Pro echoes 2022 paper, TSMC defers EUV 12 min
- Jun 24 WED AI Tech Datasette closes its agent gap, Transformers.js stalls, CUGA collapses to 7% 11 min
- Jun 24 WED AI Research NVIDIA distills to 45%, Verus drives 0/1000 races, WAPO trains on winners only 13 min
- Jun 23 TUE AI Tech gpt-oss-20b jailbroken 61% via style, PP-OCRv6 hits edge, Willison ships Moebius 11 min
- Jun 23 TUE AI News OpenAI's Codex ships exploits, Anthropic's rhetoric triggers ban, A24 takes $75M 12 min
- Jun 23 TUE AI Research Stanford clocks China at 2.7%, NVIDIA's 550B NVFP4, Inclusion AI retrofits 1T 12 min
- Jun 22 MON AI Tech Cloudflare ships anonymous Worker deploys, sqlite-utils 4.0 adds migrations 7 min
- Jun 22 MON AI Research Pythagoras-Prover-4B tops 671B, ARGAR flips reviews 75%, LoRA-α beats FFT 13 min
- Jun 22 MON AI News Samsung lifts ChatGPT ban, Apple offloads Private Cloud Compute, xAI takes Maven 11 min
- Jun 21 SUN AI News Anthropic poaches Jumper, Whittaker hits friend-bots, Atlantic outs 21M songs 11 min
- Jun 21 SUN AI Tech Lynch reframes MCP as an auth boundary; 91.5% of servers fail it 4 min
- Jun 20 SAT AI Tech Lambert and Xu defend open weights as Commerce's API lever goes live 3 min
- Jun 20 SAT AI News OpenAI ships GPT-5.5 health, ASML denies China EUVs, Jio rents 'sovereign' AI 12 min
- Jun 20 SAT AI Research Subquadratic's 12M-token breakthrough is a finetune with a self-run eval 4 min
- Jun 19 FRI AI News AWS opens Trainium externally, OpenAI poaches Shazeer, Adobe centers Firefly 11 min
- Jun 19 FRI AI Tech DeepMind monitors Gemini Spark, OFT beats LoRA, docs crash Qwen3-14B to 0% 13 min
- Jun 19 FRI AI Research Opus 4.7 runs robodog solo, o3 cracks 18 cold cases, PA-DR cuts leak to 9.9% 12 min
- Jun 18 THU AI Research AMIE ties 21 PCPs, GPT-Rosalind tops LifeSciBench 36.1%, Arbor beats Codex 2.5× 12 min
- Jun 18 THU AI News Anthropic yanks Fable 5, opens Seoul anyway; Google Home Speaker ships buggy 14 min
- Jun 18 THU AI Tech GLM-5.2 matches GPT-5.5 at 1/6 cost, MolmoMotion hits 76%, ARD spec ships empty 11 min
- Jun 17 WED AI Tech Datasette adds writes, Willison lazy-loads GIFs, Qwen3.6-27B codes locally 10 min
- Jun 17 WED AI Research OpenAI replays 1.3M ChatGPT logs, Claude experts hit 33%, Olmo 3 saves $1.5M 12 min
- Jun 17 WED AI News SpaceX rescues Cursor at $60B, OpenAI burns $19B on R&D, DOJ shields xAI 13 min
- Jun 16 TUE AI News Anthropic bans Fable 5, Salesforce pays $3.6B for Fin, Big Tech's $725B AI bet 13 min
- Jun 16 TUE AI Research GPT-5.5 tops ALE 26%, Krafton folds MoE to dense, Muon wins on curvature 13 min
- Jun 16 TUE AI Tech Willison ships datasette-agent --unsafe, gates Cloudflare on one ampersand 6 min
- Jun 15 MON AI News Mythos export ban unravels, OpenAI partner fund collides with own services arm 12 min
- Jun 15 MON AI Tech Narayanan-Kapoor reframe AI coding data, Evans writes for her past self 6 min
- Jun 15 MON AI Research SIA fuses scaffold+LoRA, RCBench caps agents at 21.5/50, ToolMaze lags 3.66× 12 min
- Jun 14 SUN AI News 42 AGs subpoena OpenAI, Amazon outs Fable 5 jailbreak, Beijing unwinds Manus 14 min
- Jun 14 SUN AI Tech Pyodide wheels bypass pip, Datasette needs APSW, water rules fall to states 12 min
- Jun 13 SAT AI News Commerce yanks Fable 5, Pentagon eyes $55B drone unit, Warren targets SpaceX IPO 13 min
- Jun 13 SAT AI Tech OLMo 3 sandbags safety evals, GPT-Realtime caps at 16K, 10-step agents at 60% 12 min
- Jun 12 FRI AI Tech Claude Fable spins up a debug harness, torch.compile fuses 5 kernels into 1 7 min
- Jun 12 FRI AI News OpenAI buys Ona, Prometheus hits $41B, DeepMind funds agent safety 13 min
- Jun 12 FRI AI Research Shape of Addition at 33%, PropMe finds 36× leak gap, SeedDance leads Dream.exe 12 min
- Jun 11 THU AI Research Cosmos 3 unifies robot stack, SocioHack mines loopholes, MapAgent in 360 cities 13 min
- Jun 11 THU AI Tech DiffusionGemma ships "experimental," datasette-agent pauses to ask the user 8 min
- Jun 11 THU AI News Microsoft bars Fable 5, House cites OpenAI report, critics cap DiffusionGemma 13 min
- Jun 10 WED AI News Fable 5 mandates retention, Gemini Live triples price, North Mini gets verbose 13 min
- Jun 10 WED AI Tech Anthropic gags Fable, agents.md unsigned, HF Jobs co-locate CI with weights 10 min
- Jun 10 WED AI Research Gemini tutor +0.26 SD, Scribe V2 tops ASR, 5 of 39 agents beat scaffolds 14 min
- Jun 9 TUE AI Tech Apple's Siri runs Gemini, OpenEnv risks a fork, AIFS savings drop to 21× 10 min
- Jun 9 TUE AI Research gget virus hits 99.7%, Harness-1 beats GPT-5.4, FrontierCode caps Opus at 13% 12 min
- Jun 9 TUE AI News Siri ships on Gemini, OpenAI files at -122% margin, worm hits 73 MS repos 14 min
- Jun 8 MON AI Research Berkeley scales the harness, Qwen-9B writes it, Moltbook agents talk around it 13 min
- Jun 8 MON AI News OpenAI projects $14B loss, survivor sues Omnilert, AI influencers hit $12B 11 min
- Jun 8 MON AI Tech Willison ships str_replace over Hashline, Nemotron skips the verify harness 7 min
- Jun 7 SUN AI News Trump pitches OpenAI equity, Krishnan exits White House, Apple licenses Gemini 12 min
- Jun 6 SAT AI Research Claude Opus 4.7 ties NMR specialists on 20 compounds, 51% on independent set 4 min
- Jun 6 SAT AI Tech OpenAI ships Lockdown Mode, Willison sandboxes Python, Wright blames RL harness 13 min
- Jun 6 SAT AI News SpaceX bills Google 3× for GPUs, Ladybird closes PRs, Uber caps coding at $1.5K 13 min
- Jun 5 FRI AI Tech Hugging Face redesigns for agents, Andon stakes cash, Claude tops Estonia test 12 min
- Jun 5 FRI AI Research OLMo needs width, MemFT needs token reweighting, Parallax needs Muon 13 min
- Jun 5 FRI AI News OpenAI and rivals back DNA fines, Anthropic files at $965B, ChatGPT gets memory 14 min
- Jun 4 THU AI News CMA forces Google opt-out, Alphabet drops 4% on $85B raise, Muse Spark 4th 13 min
- Jun 4 THU AI Research Dharma-AI cuts OCR loops 59%, Nemotron learns eval design, HKUST audits AI ideas 13 min
- Jun 4 THU AI Tech Gemma 4 skips vision guards, Reachy Mini trusts MCP, Android verifies in-stack 11 min
- Jun 3 WED AI Tech Microsoft ships MAI on cost, Holo3.1 lands at 1/10, Willison picks MicroPython 11 min
- Jun 3 WED AI News Microsoft's MAI stack, Anthropic's Glasswing at 150 orgs, OpenAI recasts Codex 14 min
- Jun 3 WED AI Research MiniMax-M2 fails repro, Anthropic logs 832 bans, Gemini Embedding 2 lands 12 min
- Jun 2 TUE AI News Anthropic files S-1 at $965B, OpenAI lands on AWS, model cracks Erdős 12 min
- Jun 2 TUE AI Tech GM cuts crash sims to 60s, IBM scaffolds enterprise LLMs, Google answers MCP 13 min
- Jun 2 TUE AI Research Sleep loops triple Rule 110, QUEST claims frontier parity, CUA-Gym hits 72.6% 13 min
- Jun 1 MON AI News Anthropic hits $965B, Brockovich maps data centers, clinicians reject Levie 11 min
- Jun 1 MON AI Tech NVIDIA bets on Cosmos 3, Willison endorses 'cancel AI,' Datasette ships 1.0a32 9 min
- Jun 1 MON AI Research SkillOpt sweeps 52-for-52, ByteDance's Shannon law, Pion rescues Muon RLVR 14 min
May 2026
- May 31 SUN AI Tech Anthropic patches Claude sandbox, DeepSWE flags git log, Datasette in Pyodide 11 min
- May 31 SUN AI News Copilot meters to $600, Whitacre quits OSS, SoftBank pledges €75B to France 13 min
- May 30 SAT AI Tech Hugging Face profiles torch, Antigravity wipes drives, Codex skips SWE-Bench 10 min
- May 30 SAT AI Research OpenAI publishes its eval playbook; AISI and Apollo flag what it doesn't share 4 min
- May 30 SAT AI News OpenAI gates Rosalind, Gemini Omni decays at turn 4, Shift films NYC homes 13 min
- May 29 FRI AI Tech Opus 4.8 abstains, jqwik plants a delete prompt, Devin's 80% won't travel 12 min
- May 29 FRI AI News Outside numbers undercut Anthropic's $965B, Google's agent pivot, OpenAI's 33% 14 min
- May 29 FRI AI Research Qwen3.5 buys 88.6% false claims, ESMFold2 tops AF3, Muon doubles on rare tokens 13 min
- May 28 THU AI Research Anthropic agent survey, IBM SRE benchmark at 47%, HRM-Text trained for $1,500 13 min
- May 28 THU AI News Anthropic, OpenAI bill by token, SQLite bans agent PRs, YouTube auto-labels AI 14 min
- May 28 THU AI Tech TRL cuts RL sync to 6 seconds, Reachy Mini drops cloud voice APIs 8 min
- May 27 WED AI Tech Approval-prompt defenses crack at Microsoft Copilot and Anthropic Claude 6 min
- May 27 WED AI News GLM-5.1 ties frontier on SWE-bench, curl swamped by AI reports, BadHost narrows 13 min
- May 27 WED AI Research GEPA, AutoResearchClaw, Gaperon: each headline turns on its verification step 14 min
- May 26 TUE AI Tech Hugging Face's agent glossary makes the harness a first-class layer 3 min
- May 26 TUE AI News Pope hands Anthropic the dais, Folha suit ends in deal, ClickUp cuts 286 13 min
- May 26 TUE AI Research Code harness lifts GPT-5.5 26pt, Llama skips tool calls, TDDev's 65.8% rebutted 13 min
- May 25 MON AI Research AIRA short of Hymba, UI traces ID 14 agents, MANSU holds unlearning at 4-bit 14 min
- May 25 MON AI Tech Datasette wires fixtures, Ronacher cuts bugs to 4 lines, Claude ports a 1983 PDF 11 min
- May 25 MON AI News Personas jailbreak GPT-5.1, Bee's LED inverts, Google shrugs at Gemini RCE 11 min
- May 24 SUN AI Tech ARIA 1.3's definition-list roles die in draft, leaving `<dl>` where it started 3 min
- May 24 SUN AI News OpenAI pilots stall 88%, xAI sued over Memphis turbines, Google holds Omni edit 12 min
- May 23 SAT AI News Google's $916 OS audited, NTSB halts AI pilot voices, startup ARR off 300% 12 min
- May 23 SAT AI Tech AI's HBM buildout shows up on entry-phone BOMs and Transsion's P&L 4 min
- May 23 SAT AI Research Mythos finds 10K vulns, Nemotron 8B hits 4× AR, Dharma 3B beats Opus 52× 12 min
- May 22 FRI AI Tech Datasette Agent self-heals SQL, Daytona boots sandboxes in 60ms 8 min
- May 22 FRI AI News FTC fines fake AI, CEOs kill Trump's order, Protect Working Musicians Act filed 12 min
- May 22 FRI AI Research SU-01 wins IMO gold, WildClawBench: 18pt harness gap, Lighthouse strips sparsity 13 min
- May 21 THU AI Research Orthrus drafts Qwen3 blocks, MinT serves million LoRAs, CFD agent gates with VLM 13 min
- May 21 THU AI Tech Ramp's Inspect merges 55%, Railway token wipes prod, tokenspeed reframes t/s 9 min
- May 21 THU AI News xAI bills Anthropic $1.25B/mo, OpenAI cracks Erdős, Google rebuilds on agents 14 min
- May 20 WED AI Research Gemini reaches 17 US labs, Pokémon harness self-edits, RL resurfaces known facts 14 min
- May 20 WED AI Tech Gemini 3.5 Flash GAs 3× pricier, Ettin 17M tops MiniLM, co-scientists retrieve 13 min
- May 20 WED AI News Google I/O 2026 leans on Flash as prices triple and agent safeguards crack 14 min
- May 19 TUE AI News Anthropic absorbs Stainless, jury clears OpenAI, Willison times the agent leap 13 min
- May 19 TUE AI Tech IBM's leaderboard rebutted, NVIDIA's rank-8 enough, PaddleOCR slides to third 11 min
- May 19 TUE AI Research Single neuron jailbreaks LLMs, MIT ELF cuts data 10×, Qwen-Image hits 4 steps 12 min
- May 18 MON AI Research Meta cuts BLT 77%, ReasonMaxxer skips RL at $25, MV-Split hits 1000 DiT layers 14 min
- May 18 MON AI News Siri routes to Gemini, NHS hides code from Mythos, Wendy's swaps wages for AI 13 min
- May 17 SUN AI News Brockman runs OpenAI product, DeepSeek trails 8 months, Malta gifts ChatGPT Plus 13 min
- May 17 SUN AI Tech Anthropic loses OpenClaw, Evans drops Tailwind, CFOs distrust agent data 9 min
- May 16 SAT AI Tech Anthropic admits 40% token bug, Abridge hits Epic, OpenAI Codex trails Gemini 12 min
- May 16 SAT AI News ChatGPT wires 12,000 banks, PwC trains 30,000 on Claude, arXiv bans AI slop 12 min
- May 15 FRI AI News Anthropic loses DC and Microsoft as OpenAI patches ChatGPT mid-trial 11 min
- May 15 FRI AI Tech Bun ports to Rust, Granite tops sub-100M MTEB, Transformers ships async batching 13 min
- May 15 FRI AI Research Google's math harness at 48%, CMU swarm +38.7%, Chinchilla adds repeat penalty 12 min
- May 14 THU AI Tech OpenAI rebuilds Codex sandbox on Windows, Willison ships Codex-built blog 7 min
- May 14 THU AI News Sutskever memo and TanStack worm rattle OpenAI as Anthropic courts SMBs 14 min
- May 14 THU AI Research SWE-WebDevBench caps at 60%, φ_first beats voting, Qwen3-4B splits think/speak 12 min
- May 13 WED AI Tech llm 0.32a2 adopts Responses API, CSP tool gates fetches, Codex repros segfault 9 min
- May 13 WED AI News OpenAI faces overdose suit, CMS pays AI care agents, Google merges ChromeOS 13 min
- May 13 WED AI Research Parameter Golf, SymptomAI, Workspace-Bench post wins on unaudited evals 14 min
- May 12 TUE AI Research MolmoAct2 tops GPT-5, Meta clears CWM on lax evals, Stanford caps counting at 2K 13 min
- May 12 TUE AI News OpenAI promises PE 17.5%, Thinking Machines voice MoE trails, Google blames LLM 11 min
- May 12 TUE AI Tech Shopify hits 77% merges, Willison ships shebang scripts, Shore prices the bill 12 min
- May 11 MON AI Research Agibot's LWD beats SFT, Odysseus tops GPT-5.4 5×, TreeFlow reframes boosting 11 min
- May 11 MON AI News NYT runs AI-faked quote, Google Finance ships at 43%, Anthropic blames sci-fi 12 min
- May 11 MON AI Tech Quinn's 10 MB FST beats a 3 GB SQLite because Finnish shares affixes 3 min
- May 10 SUN AI News DOJ probes Nvidia's $40B, Congress targets AI toys, Sarvam beats Wispr by 50pts 13 min
- May 10 SUN AI Research OncoAgent posts 100% on a retrieval proxy, skips the NCCN check 4 min
- May 10 SUN AI Tech Tossell's Codex-built Gmail client ships agent hooks, skips the injection audit 3 min
- May 9 SAT AI Research Anthropic rewrites honeypots, Ai2 routes documents, CyberSecQwen-4B beats Cisco 10 min
- May 9 SAT AI News Anthropic pays $570K, Google adds AI Overview citations, GPT-Realtime-2 at 1.12s 12 min
- May 9 SAT AI Tech Claude pushes HTML at 18×, OpenAI ships Codex safety post, Curley flags WebRTC 11 min
- May 8 FRI AI Research Anthropic NLAs probed, AlphaEvolve cloned cheaper, Nemotron loses voice lead 12 min
- May 8 FRI AI News China gap at 3-9 months, OpenAI ties Gemini voice, Anthropic divests Petri 13 min
- May 8 FRI AI Tech Mozilla files 423 CVEs, Gemini Flash-Lite up 3.75×, Willison ships commit tool 11 min
- May 7 THU AI Research Anthropic Institute opens, ElementsClaw verifies old hits, DXRG cuts 57% to 3% 12 min
- May 7 THU AI News Musk leases Colossus to Anthropic, Terafab pegged at $5T, OpenAI suit narrows 14 min
- May 7 THU AI Tech OpenAI donates MRC, Google strips Gemma 4 heads, ServiceNow rolls back vLLM V1 12 min
- May 6 WED AI Tech Andon agent forges permits, Hugging Face hides ASR audio, Willison patches llm 12 min
- May 6 WED AI Research GPT-5.2 leans on physicists, Sylph defers benchmarks, RecursiveMAS skips a rival 12 min
- May 6 WED AI News OpenAI claims the default and the ad slot; Anthropic claims FactSet's 13 min
- May 5 TUE AI News Anthropic and OpenAI's three plays: services JVs, distillation law, 2028 odds 13 min
- May 5 TUE AI Tech OpenAI's UDP voice relay, antirez ships Redis ARGREP, Gemini adds webhooks 11 min
- May 5 TUE AI Research Meta locks Sapiens2's license, Tuna-2 drops encoders, Apple routes KV layers 12 min
- May 4 MON AI Tech Anthropic's 9% sycophancy figure sits 3-5× under outside benchmarks 4 min
- May 4 MON AI News Beth Israel hedges o1-preview, KC Green fights Artisan, DualShot audits Claude 10 min
- May 4 MON AI Research Sessa beats Transformers, CHAI beats Gemini, agent survey regrades Sora 12 min
- May 3 SUN AI News Oscars price AI out of acting, Grok cuts tokens 60%, o1 runs hit $2,800 12 min
- May 3 SUN AI Tech Willison merges a blog feature from his phone; Wispr Flow idles at 800MB 7 min
- May 2 SAT AI Research AISI clocks GPT-5.5 jailbroken in 6 hours; Oxford ties warmth to 10-30pt errors 7 min
- May 2 SAT AI Tech Claude Code for web, demonstrated: Willison's campsite iNaturalist viewer 3 min
- May 2 SAT AI News DoD freezes out Anthropic, Musk's emails sink his case, Meta drops 9.4% 13 min
- May 1 FRI AI Research DeepMind's wrong test, Church recodes E. coli, Princeton's MoE tradeoff 14 min
- May 1 FRI AI Tech GPT-5.5 cracked in 6 hours, /goal loops unchecked, app feeds need an installer 11 min
- May 1 FRI AI News Musk testifies on Grok, Goodfire debuts Silico, Stripe tokenizes agent pay 12 min
April 2026
- Apr 30 THU AI Tech OpenAI's goblin fix, evals as bottleneck, Willison's `llm` goes typed 11 min
- Apr 30 THU AI Research Sycophancy, agent-coded bugs, research agents: outside audits widen each gap 13 min
- Apr 30 THU AI News Three AI pitches shrink under outside scrutiny: Anthropic, OpenAI, Zig 14 min
- Apr 29 WED AI Tech Nemotron Nano Omni is a sub-agent; Codex's prompt is a post-mortem 7 min
- Apr 29 WED AI Research Bio, skills, and judges: three benchmarks debut with the cracks already mapped 14 min
- Apr 29 WED AI News OpenAI's unverified plan, Seoul's contradiction, rural America's veto 13 min
- Apr 28 TUE AI Research Agent benchmarks on trial: gaming, unreliability, and self-graded wins 13 min
- Apr 28 TUE AI News The week AI's platform alliances got renegotiated in public 13 min
- Apr 28 TUE AI Tech New tooling, same bottlenecks — just moved one step downstream 12 min
- Apr 27 MON AI News The frontier labs are rewriting the contracts beneath them 10 min
- Apr 27 MON AI Tech OpenAI's release cadence is outrunning its deployment story 11 min
- Apr 27 MON AI Research When the scaffold outweighs the model: a day of harness-defined results 13 min
- Apr 26 SUN AI Tech Around the weights: gating, silicon, and scaffolding take over 11 min
- Apr 26 SUN AI Research DeepSeek-V4's Ascend pivot: cheaper tokens, shakier answers 4 min
- Apr 26 SUN AI News Lock-in day: OpenAI consolidates, DeepSeek picks Huawei, Google backs Anthropic 10 min
- Apr 25 SAT AI News Compute, capital, and access: the real shape of the AI frontier today 12 min
- Apr 25 SAT AI Research DeepSeek-V4 closes the open-weights capability gap — and reopens others 4 min
- Apr 25 SAT AI Tech The model isn't the variable anymore — access, harness, and silicon are 11 min
- Apr 24 FRI AI Tech The frontier splits three ways: pricier, cheaper, and sorrier 13 min
- Apr 24 FRI AI News Repositioning day: OpenAI productizes, Google bifurcates, Anthropic rations 13 min
- Apr 24 FRI AI Research Scaffolding, not weights: where AI research is actually moving today 13 min
- Apr 23 THU AI Tech Vendors ship the wins; deployers inherit the work 11 min
- Apr 23 THU AI News What Google, OpenAI, and Anthropic aren't saying out loud 13 min
- Apr 23 THU AI Research The work is migrating outward from the weights 13 min
- Apr 22 WED AI Research Long-horizon agents are outrunning their yardsticks 13 min
- Apr 22 WED AI Tech The pitches are confident; the people checking them aren't keeping up 11 min
- Apr 22 WED AI News Scaffolding and subsidies, not weights, are carrying AI's headline numbers 14 min
- Apr 21 TUE AI Research Agent research moves from leaderboard scores to the trace itself 13 min
- Apr 21 TUE AI Tech Open weights win the technical debate, lose the governance one 3 min
- Apr 21 TUE AI News OpenAI plays defense as open weights and agents redraw the coding stack 10 min
- Apr 20 MON AI Tech Agent harnesses and local leaderboards mature — and quietly hide their tradeoffs 8 min
- Apr 20 MON AI News Anthropic's Mega-Deal Day, and the Tiers Beneath It 11 min
- Apr 20 MON AI Research Clean in the lab, brittle in production: a day of disappearing wins 13 min
