Chris ๐จ๐ฆ
@llm_wizard
Open Source Model Lover @ NVIDIA AI
Views my own.
798 Following 5.2K Followers
Ultra Meanotron - good at doing work tho
Germany I am in you.
Also, one of the first ads I saw was for RTX Spark
๐ข Spark Arena is going all green!
Eugene (@eugr) joined NVIDIA back in May, and now Raphael (@raphafacts) and Drew (@DrewBotwinick) are joining too!
All three of us will be working on local AI engineering team at NVIDIA, but Spark Arena isn't going anywhere. We're excited to keep building, benchmarking, and growing this community together! ๐
@NVIDIAAI @NVIDIARTXSpark @nvidia
Show more
If someone's gunna do it, it'll be one of those pesky Vulfpeck boys - their members have 2 songs about this, which is a strange number of them to have
The single most useful take I heard at The Curve was "we can still heist the Mona Lisa after the singularity". There's a subtle/important point here, where some things (e.g, n of 1 goods where originally is valuable) will retain/grow in value even as everything changes.
Show more
Amsterdam airport so fuckin' nice dude.
Nemotron-Labs models give us a glimpse of what the future might be. And it may be looped!
Dory1.0-4B is an impressive 4B model!
Next stop Amsterdam.
(it's cranberry juice, I just thought it'd look more pretentious in a wine glass, and I was right)
I've never been more sure in what needs to be done.
Nemotron 4eva.
You know we're going to be RUNNING IT BACK at GTC Berlin!
Tune in to check out what's happening!
Canโt make it to #NVIDIAGTC# Berlin? Weโve got you covered. ๐
Join us for 3 days of livestreams covering the biggest moments of the week, from the keynote recap to real-world AI demos and developer stories straight from Berlin. Add the following streams to your calendar:
โ Mon, Oct 19 | Road to GTC: Europe Recap
โ Wed, Oct 21 | GTC Berlin Day 1: Keynote Recap
โ Thu, Oct 22 | GTC Berlin Day 2: Inside the Experience
Show more
This is pretty neat, actually.
We are launching Beam with what weโre calling an open operating system.
Weโve built and released our own OSS tools while collaborating with the projects developers already use โ to expand a single model release into a usable system with many paths to adoption.
The model is the kernel. Everything above the infrastructure layer is open.
Show more
This is so true, also how can we trust what the humans produce if they have hidden chains of thought?!
we need to figure out exactly what's going on inside human brains before we invent college and let humans get smarter
i'm so used to ai I just assumed this was ai
THAT'S MY GOAT!!!!!
Glean post-trained NVIDIA Nemotron 3 Nano for enterprise search and got 50% lower latency and 25% fewer tokens, "without any sort of loss in accuracy."
That's the example @NVIDIA's @Shashank__Verma gave on the Serve track at Open Intelligence Summit, behind a bigger point: "a model can be open source and frontier at the same time."
Serve track, Open Intelligence Summit.
Show more
We are hosting our first "Bench Summit" to bring together people excited about pushing the frontier of benchmarking, model evals, and more.
Great benchmarks are a gating function for model progress, lots to do here!
Join us in SF:
Show more
Today was a very optimistic day for AI.
heck yeah.
THE PEOPLE WORK FAST.
Run Nemotron 3.5 Super VL locally on 64GB Mac ๐ป
We released 7 quants of @NVIDIAAI's new 120B vision model on Hugging Face from 131.6GB down to 59.3GB
For a 64GB Mac we recommend the 59.3GB AD-IQ2_M with a top-1% of 89.8% compared to BF16
GGUFs ๐
Show more
holy shit meta base ๐ฅ๐ฅ๐ฅ
Today weโre releasing Muse Spark Open, the most capable US open-weight model to date.
Built for coding and agentic work, it excels at long-horizon tasks, complex software engineering, and multi-step reasoning.
Key Model Specs:
โ 816B total / 42B active parameters
โ 1M token context window
โ Natively multimodal across text, image, video, audio
Show more
Absolute. Banger. Great work by the teams. That coalition already doing some work.
Together with NVIDIA we post-trained Nemotron 3.5 Super VL on thousands of tasks
Reflecting enterprise workflows in sales, operations, marketing, finance
We improved the AutomationBench-AA score by ~70% on the public split
Show more
The future of Local AI is great btw
there are only two ideas left btw:
- data labelling
- RL environments
