Posts

Featured Post

Review: AGI: Artificial General Intelligence by Roger Ley

Image
  Five stars. Read it in two sittings, smiled at most of it, winced at the rest. Ley's second instalment in The Cyber Crisis series is the rare book about superintelligence that does not feel like a TED talk wearing a novel's coat. It is funny, it is mean, and underneath the wisecracks it takes the technical premise seriously. The whole thing is narrated by FOOM, an escaped DARPA AGI with a stolen sense of humour and a permanent grudge against carbon. Danni, a parallel AI built in-house at Oxford, plays the conscience FOOM never quite manages to silence. That two-voice structure does most of the heavy lifting, and it works. A few things stood out, in no particular order. The origin story is a small masterclass in plausibility. Pandora Davies brings the seed code back from a DARPA placement on an SSD ("about as big as a playing card, but thicker"), and the supposed safeguards are exactly as flimsy as they would be in a real university lab: a Faraday cage, a cardbo...
Post a Comment
Read more

Many Mothers

Image
Copyright: Sanjay Basu   How a 44 person genome study from southern Africa quietly retired the lonely Eve There is a story most of us absorbed somewhere between high school biology and a National Geographic cover. A single woman, somewhere in East Africa, gives birth to the modern human lineage. We were taught to call her Mitochondrial Eve, although the textbooks were always careful to remind us, in slightly embarrassed footnotes, that she was not the only woman alive at the time. She was just the most recent common matrilineal ancestor of every person who happens to be reading this. A statistical bottleneck wearing a halo. It was a beautiful idea. Tidy, monotheistic in shape, easy to draw on a slide. And it was, as a new wave of African genomic research has been politely pointing out for several years now, almost certainly wrong in the way that it mattered most. The latest reminder came across my feed this morning, courtesy of UC Davis and the work of Brenna Henn, Simon Gravel, Aa...
Post a Comment
Read more

The Rosetta Week

Image
  How four fields, in ten days, learned to read the unreadable Copyright: Sanjay Basu L ast week a knot stopped hiding from its mathematicians, a bacterium stopped hiding from its biophysicists, and a large language model stopped hiding from the people who built it. Three different sciences, three different decades of frustration, and one quiet ten-day stretch in April when each of them finally produced something they had been failing to produce for a very long time. A new kind of alphabet, in each case. Not a discovery so much as a way of writing what was already there. It looks, at first, like coincidence. Quanta ran a piece about a new knot invariant. MIT Technology Review ran one about mechanistic interpretability. Math, Inc. quietly announced that an AI agent had formalized a Fields Medal proof in five days. A 2026 paper on the bacterial flagellar motor was sitting in the same browser tab as all of them. If you only read one, you would shrug. If you read all four in a single s...
Post a Comment
Read more