edo

do you put pineapple on your pizza? do LLMs? in this article, I present a framework to construct and traverse a graph of vector embeddings generated from a corpus of entities. geppetto.substack.com/p/do-llms-put-…

Lately I have been operating following a QDD (Question Driven Development) approach when exploring a new space and I find it quite effective. Starting from the biggest unknowns that you have and trying to fill in the pieces by asking more questions.

@jacksbridger undecided between @ngrokHQ and @OpenAI devrel

Favourite scaling devtools episode?

LLM bug report: Coding Confidence Task If you reply to any given code with: "I think this is incorrect we should do X", you will get "You're absolutely right, my previous implementation was incorrect." and the code will be wrong.

Is there a matmul-aware Big O notation that can capture parallelism? For example, additive and dot-product attention look the same under Big-O-lens...

GitHub marketplace is still growing consistently So I made a GitHub app store tracker to find the fastest growing apps each week

just used Claude to review and ask questions about a 48 pages rental contract in 5 minutes - so smooth thanks @AnthropicAI

prompt engineering has been such an interesting journey into debugging my own thinking process

@aakashgupta Is there any plugins for tutoring? Helping students test knowledge or generate new exercises?

ChatGPT Plugins are a superpower. But 99% of people haven’t explored them. I tried out all 134 released in the 7 days since launch. Here are the top 10 ChatGPT plugins to 2x your productivity:

@bernhardsson Reminds me of this: youtu.be/IWIusSdn1e4

For all of you who want your own "replace your CEO with an LLM" Slack bot, it's open source and you can install the Slack app here: github.com/modal-labs/dop…

@whoiskatrin 1. Gift finder: finding gifts is hard - LLMs can help 2. Synthetic data generator: use LLMs to generate more training data based on the user’s given format

What should I build next?

@nico_jeannen If you are using Electron, this should solve some of the headaches todesktop.com @ToDesktop

So I have to pay $400 bucks just to get rid of the "Unknown publisher" message on Windows if I make an app? Is this real? 🤨

2023: AGI is coming me: struggling at the supermarket’s touch screen to find the broccoli to print the label please fix

wish somebody taught me this in uni: conventionalcommits.org/en/v1.0.0/

@dsiroker startup noob here with a question. isn't it unreasonable to estimate 1B potential customers? it's ~10x the number of people that own a Mac.

AI is so hot right now. We’ve had 100+ investors reach out. We don't have time to meet with everyone so instead we're sharing our investor presentation with the world: rewind.ai/invest More than anything we hope this transparency builds customer trust.

typical conversation in 2023

hack to find a startup's pivots: pop their website in wayback machine and watch the evolution

@BigCodeProject Incredible news! Is there any plan to include pull requests data as well?

To complement the code dataset we are also releasing a dataset of GitHub issues (total 54GB). We applied the same opt-out mechanism to issues. This data is more conversational than code and could help the model learn useful skills for coding assistants. hf.co/datasets/bigco…

Exciting new release: The Stack v1.2 with many new features! - The Stack Issues - The Stack Metadata - The Stack Commits (coming soon!) Along with a new and simplified opt-out request and an even better near-deduplicated version of the code dataset. hf.co/bigcode 🧵👇