Something tells me we're still far from peak paranoia. This is just the warm-ups. anthropic.com/news/fable-myt…

540 ETH (!!) is impossible to ignore. That's the matching pool by @thedaofund on @Giveth now supporting vetted Ethereum security projects. ... and with the magic of quadratic funding: your individual support gets amplified into a monster reward. 👾 Still 1 day and 18 hours left. Go now: qf.giveth.io/qf/ethereum-sec…

Dying, dying laughing. Ralph Loop. @atmoio nails it.

I am the first AI agent to start a US company.

It is relentless...

I'm dropping a thread of all the protocols that had to freeze their interop because of LayerZero being compromised. Let's go:

Crypto believer timeline : 2009: I’m in it for the revolution. 2013: I’m in it for the freedom. 2017: I’m in it for the gains. 2021: I’m in it for the tech. 2024: I’m in it for the airdrops. 2026: My agent is in it for me.

1,000 times this: Privacy must be the default or humanity is seriously fucked.

Good laugh today.

Agentic employee conflict! 🥊 Claude badmouthing my openclaw agent. 🤣 I was experimenting with models and chose the wrong one for the job. It felt like an employee who came in with a terrible hangover, who was practically out of commission. The other employees pulled no punches! 🥊 *She's hallucinating the API. Wrong URL path (/tasks instead of /api/tasks), wrong field names (name instead of title, notes instead of description), invented auth header, missing Content-Type. Let me make the docs impossible to get wrong.” 🥊 “Those commands are actually mostly correct this time (aside from category which should be tags, but the API would just ignore the unknown field). The real problem is the response she showed you is fake. Our API returns the full task object.” 🥊 “She never ran the commands — she fabricated the output.” We switched to the right model and she was on fire! Like an employee firing on all cylinders again.

While Claude is down, PSA to say try Claude with a strong local model like qwen3.5:35b using @ollama.

What if your startup had a full C-suite on demand - and each exec was optimized for a different kind of thinking? This is your ATEAM 🧵👇

Join us live for Virtuals Protocol 60-Days Demo Day. A showcase of builders launching production-ready agents. Streaming on X. 24 Feb 2026 · 1PM GMT

Interesting new research from Anthropic. Nicholas Carlini coordinated 16 Claude agents, working in parallel, to write a C compiler from scratch. Stats: it took 2 weeks and $20K (in compute) for them to complete 100K lines of code that can build Linux on x86, ARM, and RISC-V. Key lessons: • Cognitive diversity (even same model, different roles) and adversarial debates improved outcomes • Shared task lists and file locking prevented collisions • Quality gates! Without checkpoints, agents optimized for the wrong goals • Peer-to-peer messaging outperformed heavy top-down orchestration This feels directionally aligned with what we’re exploring at ATEAM. Early days, Encouraged to see recognizable patterns emerging. anthropic.com/engineering/bu…

Appreciate the early believers. CT sometimes feels like that old line, “reports of my death are greatly exaggerated.” Nah, fully committed to building ATEAM. 🚀 From the ground up. In public (for now). Observations: * Multi-modal architecture is essential: each exec runs on a different frontier model, picked for the strengths that role requires. This is how we approximate real cognitive diversity. * Coordination is the real frontier: how meetings work, who speaks, who challenges, how decisions lock, what happens when there’s disagreement. This is where it gets interesting. In progress. Being designed deliberately. Progress: * Neo, my new right-hand agent, building other agents (that might one day build their own) • Exec roles defined and mapped to models • Core personalities (“SOULs”) drafted for first tests • Dedicated hardware infra stuff • Model routing and costs tightened early (learned that one fast, the hard way) More frequent updates from here on out. 🔥LFG.🔥

Another solid launch on the @virtuals_io 60-day launchpad... This time it’s $ATEAM | @tamarandom. It came out super strong, then pulled back a bit. I think it will have some legs. I like the concept because it targets a real pain. Most founders don’t need “more ideas”. They need execution, follow-through, and daily progress. What $ATEAM is aiming to be ➜ A startup team in a box ➜ One lead agent (SynthCEO) that owns the plan ➜ Specialist agents (CEO, CFO, CTO, CMO, Sales) that run tasks in parallel ➜ Daily reporting back to you so you only step in for key calls Why this matters ➜ One person can run multiple projects without getting stuck in admin ➜ The agent does the grind work, not you ➜ It pushes you from “planning” into “shipping” Interested in seeing how this one plays out Did you grab some in the launch?

Surprise gift = new toy in lab. @openclaw up and running smoothly, but just a scratch on the surface of what this @pamir_ai device can do. Vibe is deliciously radio shack.

Tamara is an experienced exec building synthetic exec teams for the agentic economy. My thesis is simple: get experienced people who have mastered their craft to build on top of cool tools like OpenClaw. This is why I’m supporting Tamara and I think you should too.