Gm web3
The next wave after AI Web3 Robotics. (Projected $200B market by 2030)
Already scouting the infrastructure builders. Pre launch @xmaquina is currently on my radar. Always good to spot the silent builders early.
What are you looking at in this sector? 👇
We need a new word to describe this type of person. Vibe coder doesnt capture it. A hyper capital augmented centaur flitting between 10’s or hundreds of active threads. Someone called it spider, which captures it well but isn’t really sticky
React Native 0.84 is now available!
This release makes Hermes V1 the default JavaScript engine and ships precompiled iOS binaries by default — bringing significant performance improvements and faster build times to all React Native apps.
https://t.co/dTnGo2Uw8G
We've just acquired 'Neko' by Quantum Communications @liminalcorp on @FellowshipAi
Inspired by Serial Experiments Lain. "The cat replaces Lain but carries the same energy as her: patient, observing ... kinda caught between two worlds."
Where does the self end and the network begin?
We are live on Twitch with @WealthGroup team member @PattaTrades! Make sure to tune in and watch this public stream as we review the markets and find some good looking setups! 📊👇
https://t.co/0NieSoBwCt
The most underrated thing in OpenClaw/Clawdbot isn’t the model.
It’s persistent memory.
Agents that remember what happened yesterday get better every day.
Agents that forget stay stuck in a loop.
I built that same memory layer for Claude Code, Codex, and other agents.
One command:
𝚗𝚙𝚡 𝚜𝚔𝚒𝚕𝚕𝚜 ��𝚍𝚍 𝚛𝚘𝚙𝚕-𝚋𝚝𝚌/𝚊𝚐𝚎𝚗𝚝-𝚜𝚔𝚒𝚕𝚕𝚜 --𝚜𝚔𝚒𝚕𝚕 𝚙𝚎𝚛𝚜𝚒𝚜𝚝𝚎𝚗𝚝-𝚖𝚎𝚖𝚘𝚛𝚢
What it does:
- stores durable instructions + preferences
- recalls relevant context when you start new tasks
- reinforces memories as they’re used
- keeps memory local, simple, and lightweight
So your agent stops re-learning the same things over and over.
The creator of Claude Code recently talked about keeping CLAUDE․md updated with learnings.
This puts that process on autopilot.
No heavy setup.
No giant context tax.
Just compounding intelligence from daily use.
If your agent still forgets critical instructions, this is probably the missing layer.