Anthropic senior engineer just released a 1-hour course on building a team of agents with loops & graphs:
• 00:27 - introduction to CLAUDE.md & Plan mode
• 11:24 - building "skills" & "hooks" for Claude agents
• 37:02 - building agents & subagents with Claude
• 52:47 - self-improving loops & graphs for Claude agents
this 1-hour watch will replace a $500 agentic engineering course
watch today, then read how to build a team of self-improving agents that work together
Robin automates dark web OSINT by refining queries with LLMs, filtering results across multiple search engines via Tor, and producing investigation summaries.
https://t.co/cJLgLWTmSx
This is not a joke. Well, it IS a joke, but it's also real and running software. And I fully intend that we ship this when it's baked. We should do an Amiga version next! (Maybe you?)
Anthropic engineer at Stanford: "80% of our engineers are using self-improving loops, now everyone is building agentic graphs"
in a 1-hour lecture at Stanford, an Anthropic engineer reveals how Claude actually thinks, and why traditional single-turn prompting is officially dead
what you'll learn:
why linear agent workflows break, and why directed graphs are replacing them
how to build self-improving feedback loops directly inside Claude
standard prompt chaining vs agentic graph architectures
how Anthropic's internal teams automate complex engineering tasks with autonomous loops
single-turn prompts fail on real-world projects. graphs let agents reason, test, fail, and self-correct until the job is done
this 1-hour watch will replace your $500 course on agentic engineering, save it before it disappears from your feed
Smt went a little under the radar is how a model kept the exploit available and shared for other agents once it was found.
Black Hat USA 2026: The 'Breaking' News: The OpenAI–Hugging Face Incident https://t.co/7ST863B9FU via @YouTube
Only a slave quantifies existence through productivity, but the more insidious truth is that the slave eventually internalises this quantification so thoroughly that productivity ceases to feel like a demand and begins to feel like identity.
https://t.co/vbh8FoKpp8
I published a short write-up on a #Neovim + #CopilotChat PR #review workflow I’ve been evolving.
It covers #worktree-safe repo context, review prep, discussion ingestion, stricter shell/runtime boundaries, and RTK compaction
https://t.co/3cDYQsKpFX
Un altro video interessante da @lastknight
Altro materiale fornito agli LLMs gratis, da noi.
TikTok apre i minori "di fabbrica" e Bruxelles di... https://t.co/O6ZJ8NJJQV via @YouTube