
$10 Million for an Airline's Entire Memory
At a bankruptcy auction, Google bid $10 million for 600 million Spirit Airlines employee emails and chat messages to train AI, and more than 100 members of Congress want it stopped. The public …
Yongjie Xue · Full Stack & AI Engineer · Austin, TX
Software Engineering, AI Systems & Life.
Based in Austin, Texas. Writing about distributed systems, databases, AI agents, and architecture.
Sections
Miscellaneous
35posts
Essays, journal reflections, booknotes, and personal writing.
Database Guru
64posts
News, reports and best practices from the database industry.
Cloud-Exiter
61posts
Cloud economics, cloud-exit and self-hosting like a pro.
PostgreSQL Mage
112posts
PostgreSQL development, operations, internals and tuning.
AI Explorer
70posts
Agents, LLMs, AI coding, AI4DB and DB4AI.
Travel Notes
1posts
Trips, hikes and things seen along the way.
Recent Posts

At a bankruptcy auction, Google bid $10 million for 600 million Spirit Airlines employee emails and chat messages to train AI, and more than 100 members of Congress want it stopped. The public …

GitHub Copilot's local sandbox went GA today. The real news isn't "another sandbox" — it's three decouplings: the sandbox from the model, the engine from the product, and the policy from the …

OpenAI, Anthropic, Google and Meta testified under oath before the New York City Council, and none of them would put a number on worst-case catastrophic risk. An industry that cannot give a number …

A starter template and layout guide for travel notes. Copy this directory as a starting point for your next travel article.

Anthropic's threat report, OpenAI's soap opera, the closure of the $200 subscription tier, and rising GPU and memory prices—all of this week's events tell the same story.

pgsql.cc is live, with Chinese documentation for all 11 PostgreSQL major versions from 10 through 20, a redesigned mirror of the official website, and better full-text search. Kept in sync with …

Buying 10 AI subscriptions is easy. The real bottleneck is turning tokens into useful output—and using automation to scale your time, attention, and ability to judge the results.

As Astra began rolling out, Tibo announced one banked reset per day for paid users still waiting for access. I'm stocking up for when Astra arrives.

People often ask: how do I ensure quality when AI writes the code? My answer has four parts: documentation-driven development, adversarial review, brute-force testing, and complexity penalties.