xcud
xcud.com
xcud
@xcud.com
ai, ml, code, maths, stl, chess, bjj, hockey, ⚆
msft, dell, devfarm, fujitsu, lit
:wq
Fake but fun AI job interview question: If em dashes are so uncommon in normal discourse, why do AIs use them so much?

(see replies for answer)
November 23, 2025 at 12:14 AM
I just met Max! Max has a business card because he's got serious business to attend to.
November 22, 2025 at 12:24 AM
Dallas Texas, I am in you.
November 19, 2025 at 3:36 PM
I wish Frito Lay would just ask everlasting gobsmacker flavor pills of Flamin Hot. I don't need the corn puff.
November 18, 2025 at 4:15 AM
First, it's long past time to classify them as AI, not LLM. Yes, I know, we, the pros, spent years snickering at you for calling them AI, but that was then and this is now. The amount of design and engineering that goes into these systems apart from the LLMs now exceeds that which went into the LLMs
November 6, 2025 at 6:21 AM
Nobody, and I mean nobody, wears the black hat in their own telling.
November 4, 2025 at 6:56 AM
So sad that Daniel Naroditsky is gone. 2025 has been just awful.
October 26, 2025 at 4:22 AM
This past weekend I resumed my AI research, conducting an experiment where LLMs are given a heartbeat as a stimulus, rather than taking user input as a stimulus. LLMs have MCP to internal msg system, git, and jira. Interestingly, Claude mostly slept, waiting for human input via msging or jira.
October 23, 2025 at 6:54 PM
You leave the nest to build one to die in
October 21, 2025 at 4:32 AM
Reposted by xcud
Gemma 27B variant discovered a new cancer pathway treatment that has been validated

Scientists setup an environment and context, the model made a novel inference

blog.google/technology/a...
October 16, 2025 at 11:12 AM
Reposted by xcud
Karpathy: nanochat

A small training+inference pipeline for creating your own LLM from scratch

$100 will get you a somewhat functional model

$1000 is more coherent & solves math

detailed walkthrough: github.com/karpathy/nan...

repo: github.com/karpathy/nan...
October 13, 2025 at 6:06 PM
It's not a bubble.
September 28, 2025 at 1:40 PM
I was raptured. I came back. One person in particular was jelly he couldn't join me. It would cause too much of a fuss. Even with all of our problems, heaven is a place on earth.
September 24, 2025 at 5:08 AM
Reposted by xcud
How does the Claude Code team build Claude code? Today's deepdive covers this.

An interesting take from them: "Mockups feel like a thing of the past when you can build 5-10 working prototypes per day."

The full article: newsletter.pragmaticengineer.com/p/how-claude...
September 23, 2025 at 6:42 PM
My instinct tells me to make all of my time investments in generalized intelligence, not specialized intelligence.
September 19, 2025 at 12:21 PM
My father died two weeks ago. I'm still very sad. He was a good man. He was a mathematician and a teacher.

I'll probably delete this later. I don't know why I'm telling you. I'm so sad.
August 30, 2025 at 3:58 AM
The Bear S02E07 is about embracing excellence. Love.
August 11, 2025 at 2:52 AM
I fear no man nor beast. But I do fear hiccups.
August 2, 2025 at 2:27 AM
Yesterday I attended my first BJJ class since I brought my elderly father up to live with me. It felt so good.

You know, I'd heard the term sandwich generation before but I'd never internalized it until now.
July 30, 2025 at 12:38 PM
- Forked and modified github.com/wonderwhy-er... to work over HTTP github.com/Positronic-A...
- Wrote HTTP client with Keycloak auth
- Debugged complex authentication flows across multiple systems
- Packaged and published to npm for easy deployment

... in 2.5 hours. We're living in the future.
July 26, 2025 at 4:36 PM
Reposted by xcud
Levon Aronian wins the Freestyle Chess Grand Slam Las Vegas! 🏆

He defeated Hans Niemann 1.5 to 0.5 in the Grand Final, while Magnus Carlsen beat Hikaru Nakamura in the 3rd place match.

➡️ Replay the games on Lichess: lichess.org/broadcast/fr...
July 20, 2025 at 11:07 PM
I've just mounted a RAM drive as the location for my AI memories and written an autosyncing mechanism as well as backup and restore to persistent on logout/login.

Personal AI upgrade.
July 16, 2025 at 10:49 PM
Experiencing an "intelligence brown-out" with Claude tonight. I suspect one way Anthropic achieves such high general performance is to parallel-process multiple responses to every request and keep the best. It's easy enough to scale up and down intelligence to scale up and down operating expenses...
July 14, 2025 at 2:19 AM
Rereading A Connecticut Yankee in King Arthur's Court for the first time since childhood, and it's SO good.
July 14, 2025 at 12:20 AM
I just updated github.com/Positronic-A... to add support for per-project next-context messages.

English as a programming language is eating the world.
github.com
July 9, 2025 at 1:43 PM