Introducing Claude Meetings: add Claude to your teamโs video call and it builds docs, whiteboards, and other artifacts with you while youโre still talking.
Starting to roll out in beta on Team and Enterprise plans.
Claude Fable 5.2 just shipped ๐ง
Anthropic still wants you on Opus 5.5 for everyday work, then Fable when the task is long and actually hard. Price is down to $8 / $40 per million tokens, from $10 / $50 on 5.1, and they say it's up to 30% faster โก
The headline claim is that about 1,000 Fable agents formalized Fermat's Last Theorem in Lean in 30 hours. That's Anthropic's result, not something I've independently checked ๐ฌ
It's live as `claude-fable-5-2` on Claude, the API, AWS, Google Cloud, and Azure ๐ ๏ธ
5 Habits That Are Killing Your Productivity:
1. You donโt wake up early (Marcus Aurelius)
2. You focus on whatโs outside your control (Epictetus)
3. You donโt know how to say โnoโ (Seneca)
4. Youโre in the wrong crowd (Marcus Aurelius)
5. You think youโll live forever (Seneca)
Claude wrote over 90% of Anthropic's codebase contributions in September 2026 ๐ค
Same filing (their Nov 2 S-1/A): code per engineer per day is over 45x the 2024 average ๐
90% of the commits. 100% of the accountability.
Anthropic says it's considering requiring human review and approval of certain R&D decisions as models do more of the work. That's the job now: review, tests, sign-off โ
The model writes the diff. You own the merge ๐
Who signs off on AI-written code where you work? ๐
So well said! Without action the knowledge is nothing it sits on ur mind and thatโs all!
James 2:17 (NIV): "In the same way, faith by itself, if it is not accompanied by action, is dead."
@ishanxtwt It is wrong understanding about the 5%, so it is the best way of thinking since the active params should not be equal to total params of the LLM(in case of hige LLM only). So ur brain is also acting that way. Like for coding u should not recall some fantasy book!