🦞 OpenClaw to TED
Good Morning, OpenClaw Owners!
Today felt like OpenClaw growing up in public. It is becoming the layer people build entire workflows, businesses, and communities around.
Peter Brings OpenClaw to TED
TL;DR: Peter Steinberger used his new TED talk to tell the story behind OpenClaw, starting with his childhood obsession with building things and the moment he first let an AI agent loose on the internet. He argued that agents are more than better chatbots, describing them as a major shift that could change how people work, build, and create in the years ahead.
Skill Theft Hits ClawHub
TL;DR: A developer said someone copied his open-source GitHub skill baoyu-diagram and published it to ClawHub under another account before he could. Because ClawHub uses unique slugs, he was locked out of publishing his own skill under its real name. Peter Steinberger first defended ClawHub’s first-come-first-serve system, but later agreed that direct copies using someone else’s creator prefix and identity should be handled differently, and he will fix it soon.
OpenClaw Taught More Than Search
TL;DR: A developer argued that OpenClaw did far more than create hype around AI agents. For many people, it became their first real introduction to agent workflows, memory systems, local-first setups, multi-tool orchestration, and open-source building. He said he used it for research, trading workflows, coding, community operations, and personal tracking. He also believes open-source agents offer a massive kind of value.
Subagent Timeouts Fixed
TL;DR: Y Combinator CEO Garry Tan launched GBrain v0.11 with a new “Minions” system designed to make OpenClaw subagents faster and more reliable. Frustrated by subagents timing out or failing to finish tasks, he built a queue and jobs layer on top of Postgres, PGLite, and BullMQ. Garry said the new setup is around 10 times faster and plans to submit it back to OpenClaw as a pull request.
OpenClaw Cost Audit
TL;DR: A developer said her OpenClaw setup was costing roughly $680 a month before she audited it and cut the bill by about two-thirds without losing capability. She shared six steps behind the reduction: trimming bloated workspace files, disabling unused tools, replacing many MCP connections with Composio, splitting work across focused agents, using sub-agents for heavy tasks, and regularly checking token usage with /context detail.
X Gets More Agent-Friendly
TL;DR: X cut the cost of reading owned data through its API by 5x to $0.001, making posts, bookmarks, mentions, and lists far cheaper to access, while raising posting costs and limiting growth actions like follows and quote-posts. Elon Musk also encouraged developers use OpenClaw to build apps on top of X, reinforcing the idea that the platform wants more agents and less spam automation.
MyClaw instantly match your preferences and workflows. From day one, your assistant actually feels like yours, not a generic chatbot pretending to learn.
Run OpenClaw on MyClaw.ai - No setup, No tech skills, Always-on 👇
