Coding is a clear step up from glm-5.1
Coding shows significant improvements in context and memory.
model
All AI Pulse updates tagged "model".
18 signalsSwitch between TPUs and GPUs easily
You can switch between TPUs and GPUs without rewriting code, now with native PyTorch support.
BraceSproul and jakebroekhuizen discuss open source models
They share insights on open source models.
Meet DiffusionGemma!
Meet DiffusionGemma! An experimental open model that explores a fast approach to text generation, released under an Apache 2.0 license. Moving beyond sequential, token-by-token processes to generate entire blocks of text simultaneously. Here’s what’s new with
Whoa, I might’ve just experienced the Longest Continuous Tesla Actually Smart Summon ever in ...
Whoa, I might’ve just experienced the Longest Continuous Tesla Actually Smart Summon ever in my Model 3. My car drove itself for 0.4 miles for over 2:40 straight without stopping once. This was epic!
Highly Anticipated Mythos Model Just Dropped
The highly anticipated Mythos model has been released, priced significantly higher than previous models.
A $2,999 NVIDIA box saved him $22,000
He saved $22,000 by exiting cloud billing with a $2,999 NVIDIA box.
Here's a teaser of our Mac-1 model.
Here's a teaser of our Mac-1 model. 6.6B model, runs locally (on any Mac), requires 7GB RAM (12GB ideal), can use 487 MacOS native tools, perform multi-tool chained tasks, reasoning: ON, output: ~65 tok/s. We built a robust application layer around the model t
Cheapest high-performance long-context model
Nemotron 3 Ultra from NVIDIA is the cheapest high-performance long-context model.
Dynamic prefix cache saves model costs
Dynamic prefix cache thrashing saves model costs significantly.
Competition math won't be interesting soon
"Pretty soon, competition math, competition coding, is not going to be interesting anymore."
Option pricing models to become free by 2026
Option pricing models that cost $50k/year will be free by 2026.
New course on serving LLMs efficiently
Learn how to serve models to many concurrent users at low latency.
Optimizing multi-epoch pretraining
Scaling multi-epoch pretraining efficiently with limited data.
Shipping Nemotron 3 Ultra
Introducing Nemotron 3 Ultra, a 550B MoE frontier-intelligence model.
Anuma Aims to Enhance AI Workflow Usability
Anuma is working to make AI workflows more portable and private.
Chinese Student Runs High-Cost Models with One Chip
A Chinese student runs $1,900/month models using just one chip.
Discussing Sakana AI Project on TV Tokyo Tonight
Tonight, discussing Sakana AI's 1T parameter model project on TV Tokyo.