Sottiaux on time dilation (9336 likes, top of the day): 'A few weeks at OpenAI feel like years at other companies in terms of how much gets done. Crazy how much you can age in a few days.' And the reset tease (6582 likes): 'A good thing about having aged is that I feel that it's been 20 years since I've pressed the reset button. Intrigued to see if I can find it tomorrow and dust it up.' Altman's double: the next release party (9236 likes) — 'i think we should do another party for our next model release, the 5.5 party was a lot of fun. what would make the next one awesome?' — and 'this is a good report about a bad thing' (3248 likes).

Dan Shipper's trio: 'THEY DID IT FOLKS' (1367 likes); 'Codex native apps are here!' (258 likes); and 'definite golden age for polymaths and philosophers (luckily, a good antidote for epistemic despair) incredible time' (259 likes). Steipete's pair: 'codex' visualization feature got really good.' (492 likes) and the bubble question (463 likes): 'maybe it is a bubble?'

Claude ships its own browser in Cowork (7149 likes): 'When your task involves a website, a browser opens in Cowork's side panel, and Claude navigates, fills forms, and finishes the job.' Nothing to install — built into the desktop app, stays separate from your own browser and logins (455 likes); Claude in Chrome is now GA on all paid plans (247 likes). Thariq's double: the SendFeedback tool (944 likes) — 'instead of you guys having to hit /feedback and write up a report, you can just tell Claude to draft and approve it' — and the Stripe fraud fight (347 likes): 'We're seeing several of our customers be targeted by fraudulent requests and glad for all the work Stripe is doing to help them.'

Rauchg's pair: the security dashboard + 𝚟𝚎𝚛𝚌𝚎𝚕 𝚜𝚎𝚌𝚞𝚛𝚒𝚝𝚢 𝚌𝚑𝚎𝚌𝚔 CLI (261 likes): 'Much like 𝚒𝚜-𝚊𝚐𝚎𝚗𝚝𝚒𝚌, you can call on agents to improve your security posture with human-in-the-loop, or run in crons. Who would have thunk CLIs would save the world 😁' Plus global compute (181 likes): 'Multiple regions, failover, up to 10,000 concurrent sandboxes & 5,000 vCPUs/min ramp by default.'

Matt Turck's NVIDIA × Hugging Face analysis (119 likes): 'Truly a win-win-win: NVIDIA is officially the center of open source AI w/ Nemotron and now HF; Hugging Face scores huge + gets the perfect home (no more nagging re business model!); Open source AI wins!' Plus his guest-chair appearance (12 likes): why the best venture firm in the world might be NVIDIA; the hyper power law; why apps are becoming labs faster than labs becoming apps; RSI and 'building God'; AI bubble vs duration mismatch; learning to stop underestimating China. Levie's Box Q2 (188 likes): revenue of $321.1M, up 9% (11% cc) — highest cc growth in 14 quarters, full-year target raised to $1.290B, driven by enterprise AI demand. Josh Woodward experiments with multiplayer AI (157 likes).

The mosaic: Aditya announces DeepCogito's $43M Series A (59 likes) — a post-training research lab focused on large-scale RL and recursive self-improvement via iterated distillation and amplification (IDA), proven in public from 3B to 600B+ parameters. Zara's pair: PR teams at large companies exist to block real marketing (22 likes); the AI-writing double standard — 'it's fine if I use it; it's not fine if YOU use it' (71 likes). Nikunj's link-only post (325 likes), Amjad's cryptic 'Interesting' (133 likes), Nan Yu's trio (rule of 2.5, scrum, watermarking comms anger at Gruber), and Peter Yang's trio (who to tag for X bugs, newsletter plug, Linear sauna perk).

AI & I's 'Best of the Pod' reissue: Every's Ben talks with the CEO/CIO/managing partner of Walleye, a ~$10B AUM hedge fund (and Every consulting client). The episode opens with the legendary all-hands memo: 'Using ChatGPT is not cheating. That's a non applicable idea from academia. I use ChatGPT to write this email. You should be using it too and be proud of it. As a hedge fund, we should be ashamed to leave money on the table by ignoring tools that make us faster, smarter, and more effective. Not using these tools is like refusing to use the Internet in 1995 because it wasn't perfect.' Core threads: the owner-operator's conviction — 'if we ultimately get disrupted by AI, that's on me'; the origin moment in March 2023 when a TMT analyst demoed tools he'd built to replace his own job (became Current, the internal fundamental-investing agent platform every team now uses, spiking during earnings); the Sword in the Stone / Merlin analogy — living backwards in time, seeing the endpoint more easily than the steps; leading from the front: weekly demos with a third of the 400-person firm joining, tolerating imperfect demos, flipping fear from 'afraid to make mistakes' to 'afraid to fall behind'; adoption stats: ~75% of the firm are active ChatGPT(-like) users weekly, ~1/3 use AI coding tools like Windsurf; LLM-driven sentiment analysis on unstructured data at scale as pure alpha; recording every Zoom/call into a data lake, morning risk calls processed by LLMs for memory and prediction; writing as thinking — bullet points + context + own-voice prompting turning 4-5 hour memos into 15 minutes ('I don't think we need to waste time trying to be clever anymore'); the tools don't negate thinking — they give you more time to think; the robber-barons/Gilded Age analogy — people went from relevant to obsolete skills within their own lifetimes, faster this time; 'I'm not a VC, thank God' — VCs provocate but someone still has to build; cowboy/barbed-wire history; human nature as the stationary constant ('history doesn't repeat itself, it rhymes'); quant vs human intuition — humans keep an edge in low-large-numbers fuzzy situations; first principles as incentives alignment, hating fluff, intellectual honesty, measuring everything; daily AI journaling and tracking every workout; and the firm as microcosm — 'you have to learn these tools or you're just not gonna be competitive. At some point, if you don't, that's your choice.'