Software-for-agents will concentrate into ~3 unchallengeable super-moats around kernels like Linux and ffmpeg
wordgrammer · x · 2026-11-10
wordgrammer argues that complexity-induced selection pressures in "software for agents" run opposite to coding harnesses: only about 3 pieces of software will become so industry-critical that virtually nobody will challenge them — super-moats.
His feedback loop: agents want two things — software with zero bugs and software with unique mathematical properties, because that lets them collapse 3,000 tool calls into one complex call, lowering token costs, tool call counts, and per-call latency. Failures cost money.
On market structure: the entire software-for-agents market is just ffmpeg + the Linux kernel plus 10,000 vassal startups making wrappers — "your microVM wrapper is a Linux kernel wrapper." He caps it with a Deleuze quote about only extreme forms returning, mocking pedigree founders raising from pedigree VCs to wrap Anthropic in "the greatest sales→brand→sales feedback loop the world has ever seen."
More from AGI Musings
- Top capabilities researchers forming alignment reading groups, says it's now P0 — tszzl · 2026-11-10
- Frogdie exec: compiling code once needed a datacenter — few saw desktop power coming — draginol · 2026-11-10
- Remote work dying the same year everyone got an agent is 'the dumbest timing' — StewartalsopIII · 2026-11-10
- Duvenaud: fearing 'loss of meaning' from AI job loss is really fearing ostracism — zetalyrae · 2026-11-10
- Redditor argues frontier labs should prove AI power by curing cancer — ProxyLumina · 2026-11-10
- Kai-Fu Lee tells Columbia Business School top AI agents now do a workday in one coffee break — kaifulee · 2026-11-10