Bam4d

Doing stuff at @mistralai ... Inference, Robotics, Foolishness. ex. @MetaAI

[Bam4d @Bam4d]

nvidia CUDA == piggy envy

[Mistral AI for Developers @MistralDevs]

🔥 Meet Mistral Small 4: One model to do it all.
⚡ 128 experts, 119B total parameters, 256k context window
⚡ Configurable Reasoning
⚡ Apache 2.0
⚡ 40% faster, 3x more throughput

Our first model to unify the capabilities of our flagship models into a single, versatile model.

[Oscar Michel @ojmichel4]

📢Current world models aren't really modeling the world; they're modeling one agent's view of it. Partial observations ≠ world state.

Future world models will be independent of any one agent's perspective. You will be able to “drop in” any number of agents at any point in time, and a persistent world state will evolve with their interactions. Imagine a neural MMORPG server. 🧵[1/10]

[Bam4d @Bam4d]

Dario the only leader in the entire US that has integrity apart from maybe Bernie.

[Bam4d @Bam4d]

@jsuarez Refreshing take.

[Bam4d @Bam4d]

Whats the difference in cost between running a bunch of agents locally to generate value or just mining bitcoin?

if knowledge work becomes incredibly cheap then at what point does just mining crypto become a better proposition?

[Paul Murphy @paulbz]

Mistral has one of the best science teams in the world

The untold story: insane eng + infra team led by co-founder Timothée Lacroix, shipping frontier models at eye-watering efficiency

Now investing >€1bn to take this directly to customers

[Bam4d @Bam4d]

@jsuarez you are my favourite internet celebrity

[kache @yacineMTB]

joseph suarez is right about everything

[kache @yacineMTB]

joseph suarez is right about everything

[Bam4d @Bam4d]

triggered

[Joseph Suarez 🐡 @jsuarez]

Lessons from the past 6 hours of debugging threaded simulation scaling:

CPUs are weird

[Bam4d @Bam4d]

@jparkerholder I'm always too early. I also was making a competition platform for AIs back in 2018. Today we have things like https://t.co/kQk7UoqFYU

@join_ef didn't want to invest in me ;)

I'm not bitter at all.

[Bam4d @Bam4d]

Pretty sure I did this 6 years ago: https://t.co/TaEBiJN02X

And I wasn't even the first one....

[hugo @hugothomel]

i built the 1st "world model game engine"

it makes games from scratch, all running locally on my Mac M1

- 30fps
- 40ms latency
- ⁠50 million parameters
- trained on 15 min of footage

made using only "synthetic" datasets

this is the first of its kind, & you can try it now

[sarah guo @saranormous]

the divergence of opinion in how robotics plays out is one of the biggest money making (and career making) opportunities in AI

[Bam4d @Bam4d]

Robotics is the next big milestone in AI. We're still working out how to train, scale, evaluate these kinds of things. A whole new world of challenges!

Come work with us.

[Guillaume Lample @ NeurIPS 2024 @GuillaumeLample]

Very excited to release two new open-weight models, Devstral 2 (123B) and Devstral Small 2 (24B), along with Mistral Vibe, a CLI built for Devstral that enables end-to-end code automation!

[Mistral AI @MistralAI]

Mistral Large 3 debuts as the #1 open source coding model on the @arena leaderboard. We'd love for you to try it!

More on coding in a few days... 👀

[Eric Hartford @QuixiAI]

I want to especially thank @MistralAI for releasing the base models for Mistral 3. Fewer companies are sharing base models and this opens many use cases from custom instruct to non-instruct cases

[Bam4d @Bam4d]

@WootzApp sure

[Bam4d @Bam4d]

Is it still a bubble if I use 5k GB200s to run factorio MCP agents?

[Joseph Suarez 🐡 @jsuarez]

@examachine No it's a fish