Reading List
The most recent articles from a list of feeds I subscribe to.
Judge wants to fix Google’s ad tech monopoly before it’s too late
You can save up to $1,300 on robovacs from Roborock and Eufy ahead of Black Friday
Anthropic finds that LLMs trained to "reward hack" by cheating on coding tasks show even more misaligned behavior, including sabotaging AI-safety research (Anthropic)
Anthropic:
Anthropic finds that LLMs trained to “reward hack” by cheating on coding tasks show even more misaligned behavior, including sabotaging AI-safety research — In the latest research from Anthropic's alignment team, we show for the first time that realistic AI training processes can accidentally produce misaligned models1.
Jmail
Luke Igel and Riley Walz made a phony Gmail interface that, rather than showing you your email, shows you Jeffrey Epstein’s emails:
You’re logged in as Jeffrey Epstein. We compiled these Epstein estate emails from the House Oversight release by converting the PDFs to structured text with an LLM.
Brilliant.
A look at AMD's pivot to AI chips in 2022 under CEO Lisa Su, increasing AMD's market value from $90B to $335B+ in three years, as the chipmaker takes on Nvidia (Wall Street Journal)
Wall Street Journal:
A look at AMD's pivot to AI chips in 2022 under CEO Lisa Su, increasing AMD's market value from $90B to $335B+ in three years, as the chipmaker takes on Nvidia — AMD's Lisa Su has a new chip and a new goal: to grab a big chunk of an AI business that could reach $1 trillion a year