Random annoyance: tech reporters uncritically parroting the "we've proven that the universe isn't a simulation" line. The actual proof relies on definitions of computability which fundamentally assume they are contained *within our universe*. This in turn stems from a misunderstanding of what […]

Achievement unlocked: I pushed a fish up through an existing wall cavity and caught it in the open box on the receiving side *on the very first try*. My life has officially peaked.

I wrote a long-ish thing about implicit capabilities in the Cats ecosystem and shiny new lightweight support for user-defined (non-Throwable) error types on top of monads like IO: https://typelevel.org/blog/2025/09/02/custom-error-types.html

One of my coworkers refers to Open Source as “the most incredible thing humanity has ever accomplished.” When he says that, he’s not making a socioeconomic or political statement, nor is he ignoring technical shortcomings. Rather, he is making an observation about how millions of people have […]

The is diabolical... a Python object that hallucinates method implementations on demand any time you call them, using my LLM Python library https://github.com/awwaiid/gremllm

It feels so 60s to have to type if __name__ == '__main__': main() into every Python program.

Okay this is very cool and anybody who is touching big data should pay attention to how this evolves: https://ducklake.select/manifesto/

I’ve recently realized that essentially all opinion pieces about the energy/carbon/capital/resource impact of AI invariably spend thousands of words on napkin math figures based on shaky assumptions on how models and data centers work, all to come up with some bottom line number… And then they […]

I'm relatively certain that the fastest and easiest way to solve captcha these days is to upload the image to ChatGPT and ask it for the answer.

General rule of thumb: If you're going to submit a pull request, at least take the time to compile it and run the tests. I don't care if you used AI or not, but a pull request which trivially fails CI is a waste of everyone's time and resources.

In parallel with finalizing Cats Effect 3.6, we began work on actually porting the multithreaded work-stealing runtime from the JVM to Native, and I'm pleased to say that we finally have a fully building PR! https://github.com/typelevel/cats-effect/pull/4201 There's still a lot more to do, but […]

I'm happy to announce the release of Cats Effect 3.6.0! https://github.com/typelevel/cats-effect/releases/tag/v3.6.0 This was a gargantuan effort, and as I've previously mentioned (and discuss in the notes), it sets the stage for the future of CE's runtime with a series of almost entirely […]

Public service announcement: *every* bash script should begin with the following lines. There are no exceptions. This is particularly true if you *think* you have found an exception. You haven't. Don't do it.

The first paragraph of this article basically proves that people don't actually know what "fast" means: https://www.techrepublic.com/article/tiobe-index-language-rankings/ The outlier in their thesis is Go, which is not actually a particularly fast language per se. Its static linking results […]

Some time back, I was having an introductory 1:1 with a new executive we had brought onboard. He were bonding through grousing about various things that were technically wrong with the system, and about halfway through he stopped me and asked, "You're pretty new here, aren't you?” I answered in […]

This has hit me pretty hard. Brendan and I used to play video games together, chatting almost every day. We saw each other at almost every conference ever. Too many memories to enumerate, but the attached photo comes to mind. This isn't Brendan, this is a […] [Original post on fosstodon.org]

I don't normally talk about stuff my employer announces (I work for Nvidia, my opinions are my own, etc etc), but something interesting struck me about the DLSS4 frame generation stuff. Jensen gave the example of computationally rendering 2 million pixels and then inferring 31 million more […]

For those of you using Cats Effect to do real things, it would really mean a lot if you could take 3.6.0-RC1 for a spin together with Fs2 3.12.0-RC1 (doing them both is important!) sometime over the next week. We made significant under the hood changes and we want to make sure we don't break […]

Fwiw (probably not much), I wrote an essay trying to describe a lot of Cats Effect's high-level architecture, the general problem space of asynchronous coroutine scheduling, how we (the maintainers within the broader ecosystem) think about these problems, and how we've gone about optimizing for […]

Look, I really enjoy whitespace sensitive languages, and I generally think braceless syntax looks nice, but holy hell the rules for parsing it are bonkers. Consider: (bar: y) => z bar: y => z foo.bar: x => z Remember the *tooling* has to handle multiple versions of the language, even when the […]

So this finally happened. https://github.com/typelevel/cats-effect/releases/tag/v3.6.0-RC1 It's hard to overstate how vast of an impact this release has on the Cats Effect ecosystem, albeit one which is almost entirely invisible without measurement! This is the first full release of our […]

Pet peeve: People who implement retry with exponential backoff without random jitter. The whole point of the exponential is not to make things slower, it's to geometrically reduce the probability of coincidental multi-request alignment to zero. When you skip […] [Original post on fosstodon.org]