Thursday, September 10, 202644 min read
Working Knowledge · What Could Go Wrong
Language models reach for nuclear weapons in wargames their own authors call a proof of concept
The machine-speed "flash war" Paul Scharre warns of would need a nuclear command system no state has built, and the five recognized nuclear powers have each pledged to keep a person in the launch decision.
Working Knowledge · The Evidence
PaLM was trained on a fourteenth of the data its size called for
A scaling result DeepMind posted a week before PaLM says Google built its biggest language model far too large for the 780 billion tokens it saw.
Working Knowledge · When AI Breaks
Deloitte billed the Australian government A$439,142 for an assurance report citing books that don't exist
The fake citations came from a generative AI tool the report named only in its methodology, never on a cover page.
Working Knowledge · The Instruments
OpenAI's state-of-the-art GAIA score came from the set whose answers are public
GAIA scores a general assistant on 466 real-world tasks, most of them withheld because the public ones leak too reliably to rank on.
Working Knowledge · The Mechanics
Ask DeepSeek's model who made it and it says OpenAI
A model has no way to look inward and check what it is, so unless its maker writes the answer in, it fills the gap from a web full of ChatGPT.