@belbix

Founder of Tetu and Sacra

 Earth
Joined August 2017
  • Tweets349
  • Following87
  • Followers520
61Media
1/8 Your AI agent takes a task, goes silent, and leaves it unfinished I got burned by this with OpenClaw’s native harness and switched back to Codex. But even with Codex, agents still stop when I work through Telegram.
1
38
more replies
7/8 My default: avoid subagents and background execution unless they bring a real benefit. Parallelism can speed up a large job; otherwise it adds supervision. This is my OpenClaw/Telegram experience. I haven’t benchmarked other harnesses.
1
15
8/8 Put it in the task: “Work sequentially, without subagents or background execution. Wait for operations to finish, verify their results, and continue until the task is complete. If background execution is necessary, explain why and how completion will be checked.”
6
Are you sure your AI agent’s memory actually works? “Continue the task.” “Which one?” 😅 Or it remembers the task but forgets every decision behind it. My eight-harness comparison on LinkedIn: linkedin.com/feed/update/urn…
2
2
100
more replies
7/8 “Remember everything” costs latency, storage and money. Reindexing and retries add up. Paid embedding APIs are generally separate from Codex and Claude Code subscriptions. Worse: the agent may silently continue when retrieval fails.
1
8
8/8 You think the agent is using accumulated knowledge. It’s doing yesterday’s work again. Memory isn’t a five-minute setup. Don’t dive down that rabbit hole expecting one. How do you handle agent memory—and have you caught it failing?
10
An AI agent that starts brilliantly but stops halfway is just a more expensive to-do list. I tried to simplify my agent setup. It cost me almost a week getting everything working again.
2
1
72
more replies
Two memory systems existing does not prove they conflict or double token usage. Check what actually enters a run's context before migrating a whole fleet on a suspicion. A harness controls tools, context and task flow. Model intelligence alone is not enough.
1
10
My setup for now: OpenClaw for infrastructure, Codex as the execution harness, Telegram as the control panel. “Native” is not my quality criterion. Did the agent finish the job—or am I running the last few metres again?
34
Load more