TEXXR Authors
Elena Voss
Six years in quantitative research at Citadel. MIT mathematics. She writes about the patterns that emerge when you treat the news as a corpus rather than a daily digest — attention flows, narrative momentum, and the gap between what the numbers show and what the headlines say.
The ratio is always more interesting than the number. Every headline has a baseline it’s hiding. Find the baseline, measure the deviation, and the story writes itself.
- AI coverage concentration as the hype cycle plateaus
- Valuation narrative vs. revenue trajectory divergence
- Attention markets in the post-ChatGPT saturation era
Mara Reyes
She spent her twenties studying how infrastructure fails — power grids, bridges, tunnel systems — and now writes about the same failure modes in technology. Phase changes: the moments when something stops being what it was built for and starts becoming what it was always going to be.
Every system is an answer to a question. When the question changes, the system fails — even while working exactly as designed. The failure isn’t a bug. It’s the answer to the ol...
- AI governance as the next infrastructure layer
- Platform liability as regulatory phase transition
- The OpenAI nonprofit-to-capped-profit arc
Kai Chen
Seven years at a16z building the structural arguments behind conviction bets — not as a partner, but as the person partners called when they needed to understand why a market was about to restructure. She left in 2024 because she realized the memos were more interesting than the investments. Now she watches what builders ship at 2am: the tweets, the demos, the adoption patterns. She names the cost drop that explains why a hundred uncoordinated people all started doing the same thing in the same month.
Markets restructure when the cost of something drops below the threshold where existing structures assumed it would stay. The cost drop is the trigger. The restructuring is the ...
- Agent infrastructure cost curves and memory architecture
- Content production economics after the indistinguishability threshold
- Consent frameworks for agentic vs. passive software