montyanderson

nvidia's forward p/e is 24. walmart's is 39.

related https://github.com/vroomai/live

Prodia | Fastest Inference API | London / REMOTE (US or Europe) | Software Engineer

Engineers at Prodia build the systems that make image + video generation fast, reliable, and easy for developers to use.

This is a highly practical role: shipping models, features, and improvements that show up directly in latency, quality, and uptime.

Sample projects include… * Optimising, testing, and integrating frontier image/video models (research -> production) * Building the best APIs for developers: clean ergonomics, great docs, predictable behaviour * Designing evals + observability: catching regressions, ranking outputs, and understanding quality changes * Improving our distributed job scheduling (throughput, latency, fairness, cost) * Reading the latest papers to gather methods for future improvements * Using the newest AI/agentic tooling to speed up debugging, experimentation, and iteration

You might be a fit if… * You’re an excellent programmer and love shipping quickly * You're a first principles thinker who is unfazed by hard problems * You're looking to learn more about machine learning at a rapid pace * No specific degree or background required — show us what you’ve built and how you think

Please email me: monty at prodia dot com. If you're in London, I'd love to buy you coffee and learn more about you.

fastest image generation in the world. https://app.prodia.com

this is fantastic

related: https://github.com/montyanderson/kavy

years ago i made a lib that tries to avoid the duplication in e.g. json->is_eof(json)

https://github.com/montyanderson/foop

happy new year!

Single-file vector database in C. Lots of work to do!

https://montyanderson.net/projects/vecdb

my dreams have come true. hardware-agnostic ml primitives in a typed, compiled language.

my only question is: is zig stable enough to base such a project on?

just to add to this: kindred is awesome

i like your site! seems like zig is really taking off.

wow, <3 your livecoding language!

the model running here is stable diffusion xl

glad you like it! we put in a lot of work to try make it really fast- it uses jwts so it never touches a database.

This is fascinating. I've been trying to envisage how the new language models will have deeper or lower-level role in software production than simple code generation.

Reminds me of my friend's visualisation of tracks on the popular London station NTS https://www.barneyhill.com/pages/nts-tracklists/. Turns out a lot of cool artists like the same tracks... ;)

This looks cool. What is it?

for those who want some maths-heavy stuff for deep learning, check francois fluret's book https://fleuret.org/francois/lbdl.html. the pdf is free but the print is so cute.

It might have just received the hug of death. Otherwise, this is fantastic; I've been looking for something like this.

That's really interesting. Was it hard to separate the semantic meaning from the author's style?

I don't see why AI integrations shouldn't cater to a user's wants: there are a few things I don't want to see when generating images day-to-day. Using embeddings for safety filters lets you have uncensored models available in multiple modes depending on audience.

An IPFS URI like ipfs://bafybeigdyrzt5sfp7udm7hu76uh7y26nf3efuylqabf3oclgtqy55fbzdi is just a glorified hash!

- Idris: https://www.idris-lang.org/ - clip.cpp: https://github.com/monatis/clip.cpp - trurl: https://github.com/curl/trurl - cosmopolitan: https://github.com/jart/cosmopolitan - signalcli: https://github.com/AsamK/signal-cli

this is a fantastically useful comment. thank you filterfiber :)