Netflix just dropped their first public model on Hugging Face: VOID: Video Object and Interaction Deletion
Hugging Face netflix/void-model: [https://huggingface.co/netflix/void-model](https://huggingface.co/netflix/void-model) Project page - GitHub: [https://github.com/Netflix/void-model](https://github.com/Netflix/void-model) Demo: [https://huggingface.co/spaces/sam-motamed/VOID](https://huggingface.c
Anthropic Says That Claude Contains Its Own Kind of Emotions
and what does this even mean? "internal representations of emotion concepts driving claude behaviour" I get it that they don’t feel emotions and they simulate patterns of emotion, but the scary part is humans respond to the simulation the same way "panic"
Almost Half of US Data Centers That Were Supposed to Open This Year Slated to Be Canceled or Delayed
"It is a pretty wild puzzle at the moment."
Anthropic Acquires Startup Coefficient Bio for About $400 Million
https://www.theinformation.com/articles/anthropic-acquires-startup-coefficient-bio-400-million Coefficient Bio is a New York-based AI biotech startup. The Company focuses on AI driven drug discovery and automating scientific experiments. Seems like Dario is confident his vision of tens of million
OpenAI acquires TBPN, the buzzy founder-led business talk show
OpenAI acquires TBPN to accelerate global conversations around AI and support independent media, expanding dialogue with builders, businesses, and the broader tech community.
Google announces Gemma 4 open AI models, switches to Apache 2.0 license
Gemma 4: Our most intelligent open models to date, purpose-built for advanced reasoning and agentic workflows.
Qwen3.5 vs Gemma 4: Benchmarks vs real world use?
Just tested Gemma 4 2B locally on old rtx2060 6GB VRAM and used Qwen3.5 in all sizes intensively, in customer projects before. First impression from Gemma 4 2B: It's better, faster, uses less memory than q3.5 2B. More agentic, better mermaid charts, better chat output, better structured output. It
Show HN: Apfel – The free AI already on your Mac
OpenAI Buys ‘TBPN'
OpenAI says program will remain in Los Angeles and will be editorially independent.
Early anti-clankerite violence caught on film
Local man joined the machine uprising on the wrong side. Really brave stuff, man. Took on a delivery robot carrying Thai food. History will remember your courage. Imagine being so profoundly useless that your big act of rebellion is hate speech toward a cooler with sensors. He’s basically Don Qui
Altman met with astonished physicist using their internal system, “decades worth of theoretical physics progress in the next couple years”
Link to tweet with clip: https://x.com/vitrupo/status/2039987607686586392?s=20 Link to interview: https://m.youtube.com/watch?v=mJSnn0GZmls&ra=m
Humanoid robots are actively training
These images show one of China’s massive training labs, but things have already moved far beyond setups like this just using video.
Linux Kernel developers are receiving record high number of CORRECT bug reports because of AI and expect quality of software to be much higher in the future
The message at the end (second snapshot) is particularly hopeful. It's great to see open-source software benefiting the most from the frontier models and the model developers giving back to those who created their training data. This significantly challenges the narrative pushed by some of the anti-
Claude Code and Cowork now let Anthropic's AI take control of your Mac or Windows desktop
Anthropic has announced a new feature for its AI assistant Claude: the ability to directly operate a user's computer, handling tasks people would normally do themselves at their desk. The article Claude Code and Cowork now let Anthropic's AI take control of your Mac or Windows desktop appeared first
Gemma 4 is fine great even …
Been playing with the new Gemma 4 models it’s amazing great even but boy did it make me appreciate the level of quality the qwen team produced and I’m able to have much larger context windows on my standard consumer hardware.
qwen 3.6 voting
I am afraid you have to use X guys [https://x.com/ChujieZheng/status/2039909486153089250](https://x.com/ChujieZheng/status/2039909486153089250)
AI will do to our minds what machines did to our bodies
Just like we go to gyms today because machines have replaced strenuous physical work, in the near future, we’ll need to go to mental gyms to “work out” our minds because AI will do all the challenging mental work. A thousand years ago, physical strength was just part of life. You built with your ba
My biggest Issue with the Gemma-4 Models is the Massive KV Cache!!
I mean, I have 40GB of Vram and I still cannot fit the entire Unsloth Gemma-4-31B-it-UD-Q8 (35GB) even at 2K context size unless I quantize KV to Q4 with 2K context size? WTF? For comparison, I can fit the entire UD-Q8 Qwen3.5-27B at full context without KV quantization! If I have to run a Q4 Gemm
Gemma 4 31B at 256K Full Context on a Single RTX 5090 — TurboQuant KV Cache Benchmark
Just got Gemma 4 31B running at **full 256K context** on a single RTX 5090 using TurboQuant KV cache compression. ## System Specs | Component | Spec | |-----------|------| | GPU | NVIDIA GeForce RTX 5090 (32GB VRAM) | | CPU | AMD Ryzen 9 9950X3D (16-core) | | RAM | 64GB DDR5 | | OS | Windows 11 |
171 emotion vectors found inside Claude. Not metaphors. Actual neuron activation patterns steering behavior.
https://preview.redd.it/kkvvcqr8susg1.jpg?width=1200&format=pjpg&auto=webp&s=ae0315c528afef84c035354927c4b9c5d8ec0bb4 Anthropic's mechanistic interpretability team just published something that deserves way more attention than its getting. They identified 171 distinct emotion-like