Cheap code generation exposes a harder bottleneck: most organizations have not written down how the work actually happens. The rules, exceptions, preferences, customer promises, and judgment calls live in people’s heads, inboxes, spreadsheets, and informal habits.
For AI builders, that means the defensible product is not a generic agent. It is context infrastructure for a specific workflow. The system has to capture expert knowledge, structure it, keep it current, and make it usable at the moment a decision or action happens.
In overlooked industries, this is where durable advantage can compound. Implementation may become faster for everyone, but trustworthy context is earned through production use. Products that learn from real operators and preserve that learning will be harder to copy than products that merely generate more software.
Grey Haven read: Cheap code makes context more valuable. When implementation speed rises, the scarce asset is a reliable map of rules, exceptions, preferences, and operational judgment inside the business.
Reader guidance: Treat the item as a signal about operating readiness rather than headline momentum. The useful question is which control, workflow, or evidence requirement must be strengthened before the same pattern reaches a production system with customers, regulators, or internal teams depending on it.
Operator guidance: Invest in capturing domain context as a maintained system, not a one-time requirements document. Watch workflows where agents generate more software but still rely on undocumented human knowledge to be useful.
Source/provenance: Grey Haven Signal update linear-ASM-101-3, based on the existing backfill draft and source detail Hacker News discussion: https://news.ycombinator.com/item?id=48006967.
