bleeding edge

2 min read Original article ↗

Meta released ESMFold: A LLM to predict protein structures

biorxiv.org

Lex Friedman interviews Andrej Karpathy (formerly AI Lead at Tesla)

youtube.com

Interactive storybook demo powered by Stable Diffusion

twitter.com

BigCode releases The Stack: the largest code dataset (notably, all permissively licensed)

The Stack includes 3TB of code across 30 programming languages and is 3x bigger in size than the next-largest public code dataset. BigCode only includes code that has permissive software licenses (MIT, Apache 2.0, etc) and provides an opt-out process for developers to remove their code from the dataset.

huggingface.co

Shutterstock announces DALL·E integration and fund to compensate contributors

theverge.com

DeepMind proposes a method for in-context RL with transformers.

The paper shows that transformers can improve themselves autonomously through trial and error without ever updating their weights. No prompting, no finetuning. A single transformer simply collects its own data and maximizes rewards on new tasks.

arxiv.org

Stable Diffusion v1.5 unofficially released by Runway

huggingface.co

Google announces U-PaLM 540B

arxiv.org

Google releases open-source LLM Flan-T5

arxiv.org

CarperAI announces plans for the first open-source GPT·3 like model

CarperAI, a new research lab within the EleutherAI research collective, is releasing an "instruction-tuned" large language model trained using Reinforcement Learning from Human Feedback (RHLF). In effect, releasing an open source equivalent of GPT·3.

carper.ai

Replit releases mobile app with codegen built-in

blog.replit.com

Meta releases first speech-to-speech translation system for Hokkien

Meta's Universal Speech Translator project makes it possible to train AI models on languages that are primarily oral and do not have a standard or widely used writing system. Meta built and shared an AI translation system for a primarily oral language, Hokkien.

ai.facebook.com

Potential GitHub Copilot lawsuit

githubcopilotinvestigation.com

Oct 17

·

githubcopilotinvestigation.com

Microsoft announces Designer: A DALL-E powered design tool

designer.microsoft.com

Oct 12

·

designer.microsoft.com

Joe Rogan interviews Steve Jobs generated by AI

podcast.ai

State of AI publishes their 2022 report

stateof.ai

Stable Diffusion meets virtual reality

twitter.com

Google releases Audio LM: a LLM for audio generation

ai.googleblog.com

A working implementation of text-to-3D dreamfusion, powered by stable diffusion.

github.com

Google announces Phenaki: a model to generate videos from text

Google releases a model for generating videos from text, with prompts that can change over time, and videos that can be as long as multiple minutes.

phenaki.github.io

US White House releases a "Blueprint for an AI Bill of Rights"

The blueprint is intended to "help guide the design, use, and deployment of automated systems to protect the American Public.” They are currently non-regulatory, non-binding, and not yet enforceable.

whitehouse.gov