Sci-Fi Saturday
The Imperfection Engine
A robotics startup discovers that broken data produces better robots. Two Cold War-era novelists already knew why.
Sci-Fi Saturday
A robotics startup discovers that broken data produces better robots. Two Cold War-era novelists already knew why.
AI
How a GitHub issue about variable bit rates spawned an underground compression movement that lets 7-billion-parameter models run on a phone.
The Signal
AI coding tools work, maybe too well. A New York Times investigation finds companies drowning in unreviewed code after 10x output increases. Separately, OpenAI is paying external researchers to do safety work its internal teams keep quitting, and ASML just printed 8-nanometer chip features that will define the next generation
The Signal
DeepSeek is building its next model entirely on Huawei chips — no NVIDIA anywhere in the stack. Meanwhile, Bollywood is adopting AI faster than any film industry on Earth, and a Tufts lab just cut AI training energy by 99% using an approach the scaling crowd wrote off years ago. DeepSeek
AI
Toffler warned us about the disorientation of too much change too fast. He didn't warn us it would compound.
what-if
In the fall of 2029, a motor yacht sat anchored twenty-eight miles offshore. On board, a doctor was offering treatments not available on the mainland.
AI
Most AI agent memory systems are built on semantic search. Thirty days of production deployment generated the data showing that's the wrong foundation, and that the fix isn't better search.
what-if
LLMs can now identify anonymous internet users by their writing style alone, for about four dollars per person. Every pseudonymous post ever written is retroactively exposed.
The Signal
Meta builds its own chips to loosen NVIDIA's grip, OpenAI publishes a prompt injection defense playbook, and an AI bot spends a week hacking GitHub Actions pipelines.
AI
Why the Amazon-Perplexity dispute may become an ADA access test for AI agents, not just a platform-control or anti-bot fight.
Predictions
Inference costs are collapsing. Agents are buying things. GDP can't see the output. Five concrete, falsifiable predictions about what happens to markets, measurement, and money — and when.
Analysis
AI inference costs are falling 10x per year. The economic frameworks we use to understand markets, productivity, and growth weren't built for this. Five critical pieces are missing.