After 250 rejections, an engineer manages AI agents that build themselves
Arindam_1729 · x · 2026-10-10
- SpaceX AI engineer Lauren Tan shares how, after 2 years and 250 job applications with zero offers, he switched to managing AI agents instead of doing everything himself.
- His workflow: GrokBot receives feedback, agents triage issues and write fixes, Cloud Agents run the code, and PRs come back with videos and screenshots as verification. GrokBot now helps build GrokBot itself.
- He frames this as "agentic engineering" — treating AI as a team of digital colleagues, with a Research → Build → Launch → Improve loop.
- Includes a 24-minute talk and a full walkthrough showing GrokBot browsing sites on its own and learning a daily news workflow from watching him do it once.
More from coding & agent
- Steinberger's DevDay talk coins 'jellyware', a preview of team-Agent collaboration in 6-12 months — steipete · 2026-10-10
- AI agents rewrote attack economics: 17,600 attempts in 4.5 days, a 2-week intrusion done in 10 hours — bigdata · 2026-10-10
- When 2,000 agents share one CPU: Daytona on scaling isolated agent environments — mattturck · 2026-10-10
- 20 Grok Bot morning automation routines with exact prompts, after its search and email upgrades — socialwithaayan · 2026-10-10
- Hidden two-file test: 3 of 8 agent fixes closed the reported bug by breaking another — lulzxdxdxd · 2026-10-10
- MailCheap goes open source: self-hosted newsletters on SES, 92% cheaper than Beehiiv — marclou · 2026-10-10