came back home to this beautiful first issue from @runwayml that bridges the Art and AI communities. A quote from the mag:
"if the result blows you away, it's because traditional steps, planning and human talent are still behind it." – Ricardo Vilavicencio, the creator of the
@soumithchintala
-
Runway ML Magazine Bridges Art and AI Communities
By
–
-
Mobile ALOHA Platform Showcases Advanced Robotic Hardware Capabilities
By
–
More demos from the Mobile ALOHA platform.
— Soumith Chintala (@soumithchintala) 4 janvier 2024
This is human-teleop, so this showcases the capability of the hardware, which looks solid.
The idea to use a warehouse-class robot as the base is very clever 😊
Great work @zipengfu @tonyzzhao @chelseabfinn , very exciting! https://t.co/QkEfhzzMAoMore demos from the Mobile ALOHA platform.
This is human-teleop, so this showcases the capability of the hardware, which looks solid. The idea to use a warehouse-class robot as the base is very clever Great work @zipengfu @tonyzzhao @chelseabfinn , very exciting! -
CoPilot adoption friction limits user experience and adoption
By
–
IDK if its still true, but to try out CoPilot with Bing, I had to download the Edge browser, and then login with my Microsoft account. Just to try it out.
It blew my mind that they'd create magic, want it adopted quickly and put so much product friction for whatever internal -
Perplexity vs ChatGPT: Speed and Fact-Seeking Performance
By
–
Perplexity has become my most used AI app by the end of 2023. I use it for fact-seeking questions — including recent news / facts, summarizing opinions and recommendations on products and much more.
— Soumith Chintala (@soumithchintala) 4 janvier 2024
ChatGPT + Browsing can do similar stuff, but its like 100x slower, and is often… https://t.co/n0aftiwBBmPerplexity has become my most used AI app by the end of 2023. I use it for fact-seeking questions — including recent news / facts, summarizing opinions and recommendations on products and much more. ChatGPT + Browsing can do similar stuff, but its like 100x slower, and is often
-
Home Robotics Platform Gains Traction Despite Space Constraints
By
–
Cool new ~home robotics platform, and great to see more work in this direction!
— Soumith Chintala (@soumithchintala) 3 janvier 2024
might be a bit too large for NYC apartments though 🙂 https://t.co/TCgHuGR913Cool new ~home robotics platform, and great to see more work in this direction! might be a bit too large for NYC apartments though 🙂
-
GPT-Fast Adds Mixtral-8x7B Support in Efficient PyTorch
By
–
gpt-fast now supports mixtral-8x7B, in addition to gpt/llama.
1000 lines of simple pytorch code blazing it out! https://
github.com/pytorch-labs/g
pt-fast/pull/71
… -
Named Tensors Utility Without Complete Coverage Requirements
By
–
yes its here to stay, as it doesn't need complete coverage to be useful (unlike named tensors)
-
functorch.dim Upstreamed to PyTorch Official Repository
By
–
functorch.dim and its docs.
used to be here but it's upstreamed now to functorch.dim -
PyTorch Discontinues Named Tensors for Tensor Dims
By
–
we actively stopped developing named tensors in favor of tensor dims
-
Google’s 2017 RL-Optimized Cookie Recipe Experiment
By
–
google did something like this in 2017 when RL was the rage. i remember seeing a cookie baking at their booth at a conference, where the recipe was RL optimized