A heartwarming interruption to today’s otherwise disheartening programming (requires sound)

I do sort of love that the example used to showcase breakthroughs in reasoning models is the Monty Hall problem...and that it still takes nearly a minute to solve

I am *almost certain* that I am just late to the party on this one, but have had a lot of fun with lmarena.ai over the last few days – for me, at least, their approach to identifying the "best LLM" feels like it has the most face validity (even if it's probably still imperfect).

On @davidpierce.xyz's recommendation, I was looking to download a copy of @stephenwolfram.bsky.social's "treatise" on ChatGPT.

The search result was....concerning?

I get that Biden is old, but does the Times *have* to run with the photo where it looks like he’s part of the way through his transformation into Darth Sidious?

Sphere Tastiness xkcd.com/2893