Story

i work in AI security at Meta. OpenClaw is building an open-source AI video project for me and autonomously drove 100+ signups in 5 days. meta and other big tech let you do this as long as your OSS project doesn't *directly* conflict with your work.

not helping me build it - building it. autonomously writing the pipeline, finding cost optimizations, building tooling i didnt ask for. i mostly just review PRs now.

heres why Im doing this: around 250k tech jobs were cut last year. ~70k directly because of AI. Amazon cut like 30k roles while growing. Microsoft dropped around 15k while posting record revenue. Salesforce replaced half its support org with AI agents. software engineers were among the top roles eliminated.

If you're in big tech and think you're safe because you're technical, an AI agent is literally building production software on people's laptops/mac minis right now with minimal oversight. that's your future coworker or your replacement.

IMO the move now is open-source: the biggest AI success stories right now are disproportionately OSS: ComfyUI, OpenClaw, AutoGPT, MCP. none are VC-backed SaaS - theyre repos with READMEs. There's a structural reason for this: AI moves too fast for closed products, every month there's a new model that obsoletes the last. closed platforms cant swap components fast enough, open-source with modular architecture just lets you plug in whatever's best.

and now that anyone can vibe-code a competing tool in a weekend, the barrier isnt code anymore - its trust. companies charging $50-100/mo for API wrappers are one good README away from losing their users.

You dont need to write most of the code yourself anymore. point an agent at a problem, review what it builds, open-source it.

build reputation, community, something that's yours. worst case your job is fine and you have a cool project- best case you have a lifeboat when the music stops.