Our lab investigated whether frontier models can write optimized GPU kernels for Apple devices to speed up inference. We found that they can: our AI-generated Metal kernels were 1.24x faster across KernelBench v0.1 problems, and 1.87x faster across KernelBench…

Microsoft BASIC for 6502 Microprocessor - Version 1.1 - microsoft/BASIC-M6502

The show captures disastrous custard-making, quintessentially British faux-modesty, and the blistering hubris of bakers—including me.

On trying to mash up SQLite with ideas stolen from Accountants, Clojure, Datomic, XTDB, Rama, and Local-first-ers, to satisfy Henderson's Tenth Law. Viz., to make a sufficiently complicated data system containing an ad-hoc, informally-specified, bug-ridden…

Visualizing the internal state of a Transformer model

Blog about linguistics, programming, and my projects

We show that the scaling laws which determine the performance of large language models (LLMs) severely limit their ability to improve the uncertainty of their predictions. As a result, raising...

Exposing vulnerable people to vaccine-preventable disease is just part of life, Ladapo said.

Alex is the ultimate tool for iOS and Swift app development, empowering developers with AI for Xcode to streamline workflows, tackle complex coding challenges, and boost productivity. Discover what makes it an essential asset for modern app creation.

78% of developers claim AI makes them more productive. 14% say it's a 10x improvement. So where's the flood of new software? Turns out those productivity claims are bullshit.

<p>There's a new offline mapping program for smartphones, so I thought
I'd see how it stacks up against the one I'm already using.</p>

Together, the Bitter Lesson and Scaling Laws reveal that the god of Compute we worship is yoked to an even greater one — the god of Data.

Please forgive me: 1,100 words about eels

A big nerd debate with bigger implications for the future of work, technology, and the economy