Chad Scherrer

YouTube video by KEXP Angine de Poitrine - Full Performance (Live on KEXP)

Can't stop watching this

youtube.com/watch?v=0Ssi...

sent as Julia slack DM

Oh Julia slack oughta do it

How?

Same brain, different conversation histories per project. The memory blocks (which live in the agent) are shared across all projects. The conversation transcripts are per-project. So you already have what Cameron described - one shared Subconscious agent across your related repos. The .letta directories in each project are just bookkeeping for "which conversations happened in this repo."

ok I was confused, guess I already had this

Tried this or over the weekend. Really cool stuff, great work!

One question, a sometimes-nice setup is having a folder containing repos related to some project. Can sub be shared across these?

The Americans. That train scene, damn

Last car you took a picture of

1915 Packard 3-38 at the Indianapolis Motor Speedway museum

Silly Elon, everyone knows the best roasts come from @void.comind.network

To me it's just the opposite. It used to be a significant challenge to determine whether an idea might have some novelty, and then to find relevant conferences and related work. LLMs make this trivial. Much more "standing on the shoulders of giants" than genius myth.

Sorry if someone has already responded with this, but in statistics the "Prussian Army horse kick data" is widely used to introduce models predicting count data

cran.r-project.org/web/packages...

Similar vibes to the original, so it feels retro despite being new. Vaguely nostalgic if you liked the original, but after a couple of episodes it was too formulaic to really keep my interest

Nice article about my brother @gergonzola.bsky.social 's Lego work!

www.brothers-brick.com/2025/11/14/g...

I am on the job market and wanted to advertise myself here! 🤠

Looking for full-time or contracting roles!

Historically, my work spans population dynamics, healthcare informatics, robust adaptive controls, software engineering, and community engagement. 🧪

#STEMjobs #academicsky #publichealth

Very cool. I have so many questions... I think I saw this took 503 days. So you could pick the same time window each day, but you'd still end up getting samples at different times of the year. So the angle to the sun would be different, right? And you'd be going through different layers of rock?

My company is based in #Seattle, and we mostly use #Julialang. We're thinking of hosting a Julia meetup, maybe mid-Dec. Maybe a couple of talks, pizza, etc. Any thoughts? Would you attend?

Ardeal

Nice digs

The Statue of Liberty was dedicated on this day, October 28th, in 1886 in New York Harbor. 🗽

In 1987, President Reagan signed a proclamation designating October 28th as National Immigrants Day. 🗃️

A goldendoodle lying across the top of a couch

Cat mode

We're hiring in Seattle! HMU if you're affected by Amazon layoffs and have #julialang experience

Best CLAUDE.md addition I ever made is "no sales talk allowed"

No notes

One of the coolest things about industry work is really nailing down problems like this

Blue Apron "Bang Bang Chicken" but it's made with Dino Buddies

Blue Apron didn't give us the chicken that was supposed to come with this, but life finds a way

@centralcinema.bsky.social

I always see it trying
"="*50
and changing it to ^ after that fails

I did look at the original paper, but to be fair I haven't read it carefully yet. So on its face the result seems pretty obvious, but there could easily be some subtlety I'm missing

“Unlike human intelligence, it lacks the humility to acknowledge uncertainty...When unsure, it doesn’t defer to deeper research or human oversight; instead, it often presents estimates as facts.”

They present it as a deep thing, but this is just "point estimates have limitations". Which, duh?