Anthropic co-founder, Jack Clark:
"Stop improving your agent by hand. Learn to build one that improves itself."
30 minutes on recursive self-improvement: AI that engineers its own upgrades, and what it means for everyone who builds.
Watch it, then save the guide below 👇
KARPATHY JUST KILLED THE PROMPT ERA WITH A SINGLE DOCUMENT
prompts are easy. loops are hard. and writing fifty prompts a day is the work nobody does twice.
he shifts the burden to the harness.
you define the contract once. the model writes, reviews, restarts, and reconciles. you keep judgment. it keeps the loop.
the throughline is the same in every rule: the human owns the spec and the boundary. the model owns the execution and the bookkeeping.
planner never touches code. generator never grades itself. state lives on disk, not in context.
9 rules. start with one feature, not ten. most people are still typing prompts. this turns Claude into an agent that finishes the job on its own.
here is the official document from Karpathy explaining the architecture
🕵️♂️ Reverse-Engineering a Free Streaming Platform
Understanding how things actually work under the hood is more rewarding than the end product itself.
https://t.co/plY8ludGhZ
#Engineering#streaming
Docs fetched from the source’ ≠ bank APIs.
Some platforms use locked-down virtual browsers where you enter credentials you don’t control. Legal with consent, useful for low-risk checks, but it’s a trust workaround, not source integration.
https://t.co/6uelVQQ7Pt
#Engineering
Anyone integrated Deepseek API and figured out how to send files (like PDFs) to the model? Docs don’t mention support for file uploads. @deepseek_ai — is this possible via API? 🤔📎 #AI#API
China just destroyed the $400B global BIG luxury Brands market in a week.
Sales are plummeting for LV, Gucci, Hermès, Chanel
Chinese creators are exposing the truth...
Thread🧵
Do watch this brilliant response by @_YogendraYadav when the Anchor said "BJP itself is more than the entire INDIA block put together, What do you have to say about that".