Optimizing LLMs for edge devices can be a challenge — especially when balancing model size and reasoning quality. Our team is pushing the boundaries of what’s possible, introducing new techniques to fit large-scale language models on PCs with limited memory and maintain

Why build a humanoid robot? Because the world is designed for humans, including all the best Halloween costumes!

Our new paper performs exact volume rendering at 30FPS@720p, giving us the highest detail 3D-consistent NeRF! Paper: https://t.co/CRtvXC69s1 Website: https://t.co/EbVmedJp0U

AI assistants have changed the way we use computers to work and search for information. As LLMs become more powerful, what’s next? Agents. Excited to introduce Windows Agent Arena, a benchmark for evaluating AI models that can reason, plan and act to solve tasks on your PC.

It’s always good to temper one’s optimism for empirically validated defenses against adversarial examples, but this is the most promising one I’ve heard of in several years. Definitely worth reading this explainer thread

I thought the interview with @SchmidhuberAI on @MLStreetTalk was really well done. Full of inspiration and creative energy. 🔥 For those who think ChatGPT and LLM variants have solved “AGI”, I recommend watching the full video. Looking forward to Part 2! https://t.co/BC38DcP8QU

The legendary Ross Girshick just posted his CVPR workshop slides about the 1.5 decades he spent ~solving object detection as it relates to the ongoing LLM singularity. Excellent read, highly recommended.

Language is primarily a tool for communication rather than thought https://t.co/8V9zPoMDjK “Language is a defining characteristic of our species, but the function, or functions, that it serves has been debated for centuries. Here we bring recent evidence from neuroscience and

🚀We're excited to emerge from stealth and announce our $30M Seed financing round. A huge thank you to our incredible investor group including @HoxtonVentures, @BasisSet, @lightspeedvp, @northzoneVC, @localglobevc, Touring Capital, Giant Ventures, @FjLabs, @ZeroPrimeVC & Tiferes

Refreshing take from @cohere co-founder @nickfrosst that we should focus on solving real-world business problems with large language models rather than "AGI". Just dropped on MLST.

Reminder: no one has cracked AGI yet, not the frontier labs, not Ilya, no one. Everyone is playing the same iterative game, looking N steps ahead, and trying to guesstimate what happens next.

I'm partnering with @mikeknoop to launch ARC Prize: a $1,000,000 competition to create an AI that can adapt to novelty and solve simple reasoning problems. Let's get back on track towards AGI. Website: https://t.co/wNsM3IQgEI ARC Prize on @kaggle: https://t.co/Lhsh1RiWKq

Why isn’t everyone talking about this??? Deepspeed devs literally just created a datatype FP6 with full tensor core support on the a100’s. (Since nvidia left us stranded with int4/8) It is SO smart just reading through the kernel, my god.

My tirade against AI doomers in Davos. https://t.co/SURHmhWywq

Today may be the ImageNet moment for robotics. RT-X: the largest open-source robot dataset ever compiled, across 33 institutes, 22 robot hardware, 527 skills, and 1M episodes. Why is robotics lagging so far behind NLP, vision, and other AI domains? Data scarcity is the main

It feels so weird to me that people talk about open source as reckless and a risk when it comes to AI model. Open source has been responsible for the last 20 years of technological revolution. Builders sharing their free knowledge to the world to benefit us all. But suddenly

Really interesting talk by VC legend Bill Gurley: https://t.co/2FntcDjrLF V. cogent (& 🔥) abt regulatory capture & the industry / govt revolving door. I'm in favor of some kind of gov regulation of AI, but it has to avoid those traps. I'm also big fan of open source.

In less than an hour I am going to present our paper MERF at #SIGGRAPH2023 in Petree Hall D. MERF allows you to interactively explore large scenes on a laptop in the browser. Check out our web demo: https://t.co/a4uOxriY1D By the way we have also released the entire code now!

How do we know how smart AI systems are? | Science

Στην Cohere έχουμε ξεκινήσει ένα πρότζεκτ για να εκπαιδεύσουμε πολύγλωσσα μοντέλα. Για τα ελληνικά δεν έχουμε βρει ακόμα πρεσβευτές (ambassadors). Αν θέλεις να βοηθήσεις να διασφαλίσουμε ότι τα επόμενα LLMs θα μιλούν άπταιστα ελληνικά, γίνε ambassador ή κάνε retweet/share!