Topic: tech-labor
2-Body Problems and the Intelligence Illusion
Why Language Models Fail When Truth Is Co-Defined
Abstract
Most commentary on large language models still frames their limitations in terms of intelligence, hallucination, or benchmark performance. That framing is too coarse. The more…
The Ledge & The Sludge
Practical categories where AI still struggles
1. Ownership and lifecycle design
- Deciding who owns a value, not just how to wrap it
- Borrowing versus retention across module boundaries
- Lifetime management across async work
- Destruction as…
Number One AI Skill
The New Abundance
Artificial intelligence has changed the experience of building things. Tasks that once required days of work can now begin in minutes. A prototype can appear quickly. A system can be sketched out, coded, and refined with assistance…
CE & the Zombie SaaS
There was a time when intelligence was scarce in the economic sense.
If you wanted a marketing strategy, you hired a strategist. If you wanted distributed systems, you hired engineers. If you wanted legal reasoning, you paid billable hours.
The…
Waterfall revisited
Agentic Development as the Return of Big-Batch Software
Abstract
Agentic development promises autonomy, speed, and exponential leverage. In practice, many implementations resemble a familiar pattern: large upfront specification, extended…
Archive
🔒 RegisteredOlder posts are grouped by date.