If Thoughtworks devs aren’t trusting AI generated code, you probably shouldn’t either. martinfowler.com/articles/explo…

AI research is great, until it’s successes are over-interpreted as the evolution of emergent properties, then it becomes anti-scientific dogmatism that cuts off other promising directions of AI research that don’t fit into this metaphysical straightjacket.

But if your brain is aware of this illusion, then it must be comparing this illusion to the world, which means your brain actually understands the world.

Great post that prompts perhaps the most urgent question of our time: How are people to navigate debates on climate, or AI, or pandemic response, when they concern sophisticated modeling techniques lacking any clear basis in intuition?

Or maybe it’s precisely the well-known critique of standardized tests - reducing intelligence & skills to regurgitation of factoids memorized w/o understanding - that applies to inflated LLM claims, bc both emerge from the same culture that devalues real intelligence & skills.

Arendt’s analysis of *thinking* in Life of the Mind, where she grounds the banality of evil in the nature of thinking, is one of the greatest works in 20th cen philosophy.

It can all be true that (1) LLMs produce lucid text, (2) bc they reveal meaning to be highly structured, and language to reflect that structure (no need for emergent properties), however (3) *new* meanings, new insights, require new structures which LLMs can only fabricate OOD.

Learning-to-rank (LTR) models learn to order query results by doing pairwise or listwise comparisons. At SIGKDD 2025, Amazon researchers show how to improve LTR models by using transformers to factor in absolute, not just relative, utility. #KDD2023 amazon.science/blog/leveragin…

Gen AI, like any tech, isn’t some neutral tool. It alters your relation to your future possibilities, either downskilling or upskilling your future self, and your kids.

The Official Trailer for 'Iso Kern' is now live! Click the link below to watch a moving story of the life of a modern philosopher and how he faithfully practiced his ethical and theoretical principles. #Iso #Kern #documentary #filmmaking #officialtrailer youtube.com/watch?v=UMqe-D…

One day this will be a classic. The most critical book for grasping current crises of science, from the replicability crisis to quantum theory, from ML to climate science. I’m reading it for the 3rd time. We must question what we *mean* by probability. a.co/d/33Cy7V9

The most critical aspect to understand what’s new about LLMs is that they are *probabilistic* models, and yet overcome what were thought to be central constraints around over-parameterization, sparsity and generalization in ways we don’t yet mathematically understand.

ChatGPT doesn't know the foundational text of its own design, which is widely cited.

Bay Area Friends, Join me and Matthew B Crawford in conversation on The Entanglement at Clio's in Oakland (353 Grand Ave, Oakland) next Tuesday the 23rd of January at 7 pm. I hope to see you there! eventbrite.com/e/life-art-and…

It’s interesting most U.S. grad philosophy programs no longer require foreign languages, on the idea philosophy progresses like math, and yet many math grad programs *do* require foreign languages.

One advantage in studying philosophy is that it helps you to see that you can regard another person’s views as utterly crazy and still respect his intelligence and regard him as a friend. It’s good preparation for Twitter.

Saying we should let Putin colonize Ukraine to avoid World War III is like saying we should let Hitler colonize Czechoslovakia to avoid World War II.

Washington, DC is a beautiful city. DC is home to 700K people and welcomes millions every year. We have the #1 park system, fantastic public schools, and a tremendous public transportation system. And we are at a 30-year low in violent crime. It's important for all who live