having spent a little bit of time working on agents, the bottleneck if often not inference speed but interaction speed. Interacting with buttons and browsers and UIs is so inefficient and much slower than say if you had direct API access, and very often you don't.
AI
-

Competition Against Nvidia in AI Chip Market Intensifies
By
–
My good friend Eric Savitz has published a lovely feature on the competition against Nvidia for Barron's, with excellent data and examples. I've been writing about the AI chip competition to Nvidia for almost a decade. In all that time, no one has put a dent in Nvidia's
-

First AI Workshop Success: Join Virtual Sessions Year-Round
By
–
Lam and I had a little too much fun delivering our first workshop! Join in on the laughs during one of our virtual sessions through the year https://
pulitzercenter.org/focus-areas/in
formation-and-artificial-intelligence/ai-spotlight-series
… And if you don't see a session in your time zone, send us an email. We'll add more sessions to the calendar. -
Community Standards for AI Systems and OpenAI Conventions
By
–
Yeah, I totally understand that! Promising not to change anything would be an unreasonable burden But apparently the community /really/ wants some kind of standard here – and is making do with the OpenAI conventions in the absence of something more appropriate
-

Sakana AI Releases EvoSDXL-JP Image Generation Model
By
–
We have released "EvoSDXL-JP," an image generation model constructed using the "Evolutionary Model Merging" proposed by Sakana AI. The constructed model supports Japanese and can generate images 10 times faster than conventional Japanese models. Blog → https://
sakana.ai/evosdxl-jp -

Gaussian Splats Transform Movie Scenes Into Navigable 3D Experiences
By
–
Reddit has a thread of famous movie scenes turned into Gaussian Splats, a way of converting a 2D image taken from multiple points of view into a 3D one.
— Ethan Mollick (@emollick) 22 avril 2024
You can navigate through them, up to a point. I can finally play Citizen Kane as the first person shooter it was meant to be. pic.twitter.com/SCmNfwF0rfReddit has a thread of famous movie scenes turned into Gaussian Splats, a way of converting a 2D image taken from multiple points of view into a 3D one. You can navigate through them, up to a point. I can finally play Citizen Kane as the first person shooter it was meant to be.
-
Meta-Llama-3-70B-Instruct Q4 Quantized Model Implementation
By
–
I used the Meta-Llama-3-70B-Instruct.Q4_0.llamafile (quantized at Q4) file from here:
-

Running Llama-3-70B-Instruct locally on M4 Mac
By
–
Pleased to report that the Llamafile release of Llama-3-70B-Instruct works on my 64GB M4 – I had to download the 37GB file, quit a whole bunch of apps to free up RAM, then chmod 755 it, run ./Meta-Llama-3-70B-Instruct.Q4_0.llamafile and visit it on port 8080 on localhost
-

Meta-Llama-3-70B runs on M2 Mac with 64GB RAM
By
–
OK, can confirm that it works on an M2 with 64GB of RAM I had to quit Firefox and VS Code to free up enough RAM for it to run – Meta-Llama-3-70B-Instruct.Q4_0.llamafile now uses 38GB of RAM and runs at about 7.5 tokens a second
-
LLM Documentation OpenAI Compatible Endpoints Guide
By
–
Here's my current documentation on using LLM with these kinds of endpoints https://
llm.datasette.io/en/stable/othe
r-models.html#openai-compatible-models
…