Cherny on Opus 5 (5042 likes): 'Opus 5 is a great model for coding, data analysis, design, biology, knowledge work. More than any of these eval scores, what is most exciting to me is something else: Opus 5 is our least prompt injectable model.' Security as a headline feature.

Thariq's system prompt revelation (10199 likes): 'We removed ~80% of the Claude Code system prompt for our newest models, this is what we've learned about writing system prompts, skills and Claude.MDs for them.' — simpler prompts for smarter models. Plus his daily-driver endorsement (3713 likes): 'Opus 5 rounds out our Claude 5 family beautifully. Pair it with Fable for planning, brainstorming or fixing the hardest bugs.'

Alex Albert's 6-month comparison (2285 likes): 'Just over 6 months later, Opus 5 now produces near-superhuman level spreadsheets and slide decks that match what a consultant would make.' Plus token efficiency (607 likes). Cat Wu (626 likes): 'Claude Opus 5 is great at long-running autonomous work!'

Dan Shipper's honest review (2221 likes): 'BREAKING: Claude Opus 5 is OUT NOW! And…it's a hard model to love. We've spent the last week @every testing it across coding, writing, knowledge work, and our internal agent. It argued with instructions.' — not every launch is universally loved, and that's worth reporting honestly.

Claude official's three launch tweets: pricing (2665 likes) — 'Opus 5 is available today on all paid plans and the Claude API, priced the same as Opus 4.8. It's the default model on Claude Max.' Cybersecurity (2310 likes) — 'Opus 5 is stronger than Opus 4.8 on cybersecurity tasks. But it remains substantially behind Mythos 5 at developing exploits.' Alignment (2162 likes) — 'most aligned model to date, lowest rates of reckless or deceptive behavior.'

Levie's enterprise confirmation (281 likes): 'Claude Opus 5 is out now, and it's a huge jump over Opus 4.8 on a wide number of capabilities. At Box, we've been testing Claude Opus 5 with the Box AI Agent on Box's Complex Work Eval.'

Altman's open weights endorsement (11498 likes): 'i want the US to win in AI both in open source and proprietary models, and i am glad to see this.' — the highest-engagement tweet of the day, signaling a major policy alignment. Plus his new tier teaser (3132 likes): 'let us know what you think of pro-ultra-superhard.'

Levie's three tweets on open weights: broad support (271 likes) — 'Not sure if anything in tech has ever gotten as much broad-based support and alignment as this post.' Box signing (216 likes): 'Very happy to have Box sign this letter. We're big believers in the power of open weights AI.'

Amjad Masad's challenge (614 likes): 'Will Anthropic sign? If you work at Anthropic worth asking your leadership to sign or make their position clear: Are they for banning open weight models?' — putting direct pressure on the closed-weights camp.

Steipete's graph engineering moment (2453 likes): 'am I a graph engineer now' — his loops vs graphs question from last week finally resolved into identity. He also hit a new autoreview record (135 likes): '66 rounds on a gnarly refactor.' The iterative depth possible with AI-powered code review is staggering.

Josh Woodward's Gemini Spark demo (768 likes): 'Gemini does more than talk. It gets things done. Drop in a school calendar PDF. Tell Gemini to add every No School day to your Google Calendar. Done.' — practical agentic AI for parents.

Rauchg's Figma2React endorsement (361 likes): 'Figma2React. It's good.' Plus his challenge to the community (396 likes): 'Ok let's see what you can do.' And his ambition complaint (38 likes): 'Where's the ambition!'

Sottiaux's global launch (1167 likes): 'ChatGPT Work is available globally for all paid plans across mobile/web/desktop. Probably already on your phone. Puts a jetpack on your ChatGPT.' Plus his pet show (1416 likes): 'Show me your pet' — building community one adorable photo at a time.

Garry Tan's YC rehearsal (598 likes): 'YC Startup School 2026 rehearsal at Chase Center. Welcome to San Francisco, future founders!' His productivity reality check (283 likes): 'To get macro productivity gains, managers and CEOs have to greenlight radically different staffing and workflow plans. Be prepared for this to take 10 years not [less].' His tech adoption history lesson (181 likes): 'How fast a country adopted new technology over the last 200 years accounts for at least 25% of why some nations are rich and others aren't today.'

Zara on speed vs intelligence (128 likes): 'The #1 thing I want from any model right now: speed. Intelligence is already good enough. But waiting 1-5 min per task is the worst possible window.' Peter Yang's workflow (88 likes): 'Lying on the bed and talking to ChatGPT Voice to do work in codex is pretty great.' Madhu Guru on the real opportunity (282 likes): 'There is a massive opportunity for people who know how to take messy real-world workflows and adapt foundation models to them.'

No Priors hosts Andy Fang and Stanley Tang, co-founders of DoorDash. Topics: how DoorDash has been a robotics company for 8 years, the data advantages of their logistics network, what autonomous delivery means for 9M dashers and 3B deliveries/year. Full transcript (58049 chars) covers their autonomous vehicle strategy, the transition from human to hybrid to autonomous logistics, and how AI is reshaping last-mile delivery at scale.