All of linear algebra in one diagram:

I continued exploring #stablediffusion's latent space over the weekend and oh my; there's still a LOT of treasure to be discovered inside this magnificent neural universe! Here's a quick thread with some of my personal favorites and how I found them..

#NLProc: Stop annotating randomly-sampled data! Select-then-annotate to make language models better few-shot in-context learners (ICL)! 12% gain with GPT-3 etc over 10 tasks. 10x annotation efficiency. Different from the finetuning (FT) paradigm. arxiv.org/abs/2209.01975 1/5

How the largest stocks in the world make their money $AAPL $GOOGL $AMZN $MSFT

Ever wondered how a QR code works? No, me neither but it's low-key fascinating. (Warning, there is some extremely nerdy shit here.👇 )

We have these amazing generative models inside of our brains. Proof 👇

We are excited to announce the release of Stable Diffusion Version 2! Stable Diffusion V1 changed the nature of open source AI & spawned hundreds of other innovations all over the world. We hope V2 also provides many new possibilities! Link → stability.ai/blog/stable-di…

#OpenAI is planning to stop #ChatGPT users from making social media bots and cheating on homework by "watermarking" outputs. How well could this really work? Here's just 23 words from a 1.3B parameter watermarked LLM. We detected it with 99.999999999994% confidence. Here's how 🧵

ChatGPT Mascot

I described some of the most beautiful and famous mathematical theorems to Midjourney. Here is how it imagined them: 1. "The set of real numbers is uncountably infinite."

Ce n’est pas que PowerPoint (et al.) soit un mauvais outil — bien au contraire — c’est surtout que la plupart des gens l’utilisent mal. #Thread par Loïc Magnus

🧵On money 💵 and impact (and open-source) Today, I presented scikit-learn to entrepreneurs; One CEOs was puzzled: "at the end of day, a company builds a product on top and captures the money". Looking back, I could not dream of a better way to have an impact 1/5

StackOverflow is laying off 28% of its workforce. This may be the first large layoff directly due to AI: > people asking ChatGPT instead of StackOverflow > usage & ad revenue declines > having to lay people off to stay profitable / survive

Over the past weeks the H4 team has been busy pushing the Zephyr 7B model to new heights 🗻 The new version is now topping all 7b models on chat evals and even 10x larger models 🤯🔥 Here are the intuitions on it 1/ Start with the strongest pretrained model you can find:

Nice, a serious contender to LMSYS Org in evaluating LLMs has entered the chat. LLM evals are improving, but not so long ago their state was very bleak, with qualitative experience very often disagreeing with quantitative rankings. This is because good evals are very difficult

Le petit dej #SystemeFactory bat son plein ce matin. Présentation et échanges entre membres

Jim Carrey once said "Solitude is dangerous. It’s very addictive. It becomes a habit after you realise how peaceful and calm it is. It’s like you don’t want to deal with people anymore because they drain your energy.” 9 Things TO learn from JIM CARREY - THE PHILOSOPHER

A few thoughts on a recent paper discussing the relationship between formatting constraints and performance for LLMs. It suggests that structured text generation decreases performance in reasoning tasks and seems to contradict earlier empirical results. How can that be? (1/n)

Today, we release several Moshi artifacts: a long technical report with all the details behind our model, weights for Moshi and its Mimi codec, along with streaming inference code in Pytorch, Rust and MLX. More details below 🧵 ⬇️ Paper: kyutai.org/Moshi.pdf Repo:

my weekend project to learn about bluetooth mesh networks, relays and store and forward models, message encryption models, and a few other things. bitchat: bluetooth mesh chat...IRC vibes. TestFlight: testflight.apple.com/join/QwkyFq6z GitHub: github.com/jackjackbits/b…