"Pleading prompts produce Overfit code" - Hang this quote placard near your Devs or set this as wallpaper.
"Please fix this case" == LLMs create code just to fix that case with lot of "if" conditions.
At the current state, there is no way LLM is replacing an quality engineer.
Leaderboards never work. Meta even famously did "something" to make it to ML leaderboards.
Who ever suggested a token leaderboard at Anthropic should be at Sales :)
I can now probably say this:
Two months ago, inside Anthropic someone suggested building a token leaderboard.
A heated internal debate followed and the decision was made to *never* ever do it… because several people inside Anthropic simply thought ahead of the consequences
"Attention is all you need" is a seminal paper for LLMs.
For any one now that line applies more than ever. With so much AI Slop around. If you have some ones attention, thats the best moat. We can call this distribution or access or what ever. But attention is that matters.
AI doing its magic , board games have become new passion. What i learnt in this process. Never Never evey say to AI.
/goal "recreate Ticket to ride in UAE, similar to EU, Dont make mistakes".
This 100% never works.
Its just for personal fun. Dont throw copyright on me :P
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
“If you're so smart, why are you broke?”
That’s the regular quote .
“If you're so smart, why are you not shipping ?”
That’s for the new Claude architects. People just love to talk . Show me the output.
What do I fear the most now
‘npm i <something ai > ‘
Supply chain attacks are crazy .
If you don’t upgrade, known SVE will kill you . If you upgrade an unknown SCA will kill you.
Both at speed.
Wild times …
No matter what kind of company you are...start making your internal company data legible to AI. Today.
As a founder, you are essentially building two versions of your company: the one humans work in and the digital twin that AI agents navigate to do the heavy lifting for you.
If you’re still reading engineering blogs on
“how to scale cache”
“design URL shortener”
…you’re already behind.
2026 is:
AI-native
agent harness
Engineer cost vs AI cost
Managing agents not developers
@awnihannun You’re right to call that out. The spec clearly called for “clean dishes”. You even said “make no mistakes”. Next time I’ll make sure to clean the dishes and make no mistakes. Ready to clean?