GitHub, every mention
73 scenes (2024), the whole family · ← back to GitHub
tap a year for its mentions
every year 2024 anyone Shawn Wang 111Kyle Daigle 64Graham Neubig 22Varun Mohan 20Jared Palmer 19Alessio Fanelli 17Quinn Slack 14Boris Cherny 11Walden Yan 9Alistair Pullen 9
Verbatim, from the transcripts: the passages where GitHub comes up
Best of 2024 in Agents (from #1 on SWE-Bench Full, Prof. Graham Neubig of OpenHands/AllHands)
- ▶ 4:58 Graham Neubig And here, um, I, I want to open up specifically that GitHub repo.
- ▶ 10:34 Graham Neubig Um, the agents are super good at using the GitHub API also, so they can do, you know, things on GitHub, like finding all of the, you know, comments on your issues or checking GitHub actions and stuff.
- ▶ 10:34 Graham Neubig Um, the agents are super good at using the GitHub API also, so they can do, you know, things on GitHub, like finding all of the, you know, comments on your issues or checking GitHub actions and stuff.
- ▶ 12:26 Graham Neubig We have a GitHub plugin for tagging and resolving issues.
- ▶ 15:53 Graham Neubig And so for solving GitHub issues, you can kind of have, uh, an overall plan.
- ▶ 18:43 Graham Neubig So just to give an example, um, we fix GitHub actions when GitHub actions are failing, and we do that over and over and over again.
- ▶ 23:12 Graham Neubig Benchmark that comes from real world pull requests on GitHub.
- ▶ 26:33 Graham Neubig Um, right now we have 53% or 55% on sweet bench verified, which is real world GitHub PRS. 3 times in the scene
- ▶ 39:48 Graham Neubig It's like, um, if you detect the word GitHub anywhere, you get instructions about how to interact with GitHub, like use the API and don't browse. 6 times in the scene
- ▶ 43:31 Graham Neubig It will sometimes fail to fix a GitHub workflow because it will not look at the GitHub workflow and understand what the GitHub workflow is doing before it solves the problem.
- ▶ 48:38 Graham Neubig My preferred solution to this at the moment is GitHub, like, fine-grained authentication tokens. 5 times in the scene
Best of 2024: Synthetic Data / Smol Models, Loubna Ben Allal, HuggingFace [LS Live! @ NeurIPS 2024]
- ▶ 25:39 Loubna Ben Allal For example, here we try to force the model to follow a schema for extracting key information from GitHub issues. 3 times in the scene
2024 in Post-Transformer Architectures: State Space Models, RWKV [Latent Space LIVE! @ NeurIPS 2024]
- ▶ 19:02 Eugene Cheah So, and, and we do all this actively in Discord, GitHub, et cetera.
Windsurf: The Enterprise AI IDE
- ▶ 2:32 Varun Mohan Outside of Silicon Valley, a lot of people don't use GitHub, right? 3 times in the scene
- ▶ 16:01 unnamed speaker Sweet bench is more existing, but it sounds like, I mean, your eval creation is similar to sweet bench as far as, like, using GitHub commits and kind of like the history, but then it's more like masking at the commit level versus just…
- ▶ 28:08 unnamed speaker And Copilot was 10 bucks a month.
0 to over $8M ARR in 2 months as a Claude Wrapper (Bolt.new, Qodo)
- ▶ 9:08 Eric Simons And, and so in the first camp, what these developers are doing, like to go from zero to one, they're coming to bolt and then they're ejecting the thing to get up or just downloading and, you know, opening cursor, like whatever to, to, you…
- ▶ 32:41 Itamar Friedman Is it like on cloud or deployed on on-prem?
- ▶ 36:42 Itamar Friedman I, I, I think Copilot is an amazing tool, but I'm quoting others, meaning GitHub Copilot, that they see not so good retention of GitHub Copilot and Enterprise. 4 times in the scene
- ▶ 40:49 Itamar Friedman It could be, by the way, like, for example, oh, we're only AWS, ah, but our GitHub enterprise is on-prem.
- ▶ 45:20 Shawn Wang It's the AlphaCodium GitHub and we'll, we'll put all this in the show notes as well.
- ▶ 1:10:56 Eric Simons AI coding thing you can get, maybe other than Copilot, but, like, we were overrun by support tickets, and it just, and just the sheer volume of people coming in, it's just laws of supply and demand. 5 times in the scene
- ▶ 1:28:55 Eric Simons He's got code spaces.
[Paper Club] Embeddings in 2024: OpenAI, Nomic Embed, Jina Embed, cde-small-v1 - with swyx
The new Claude 3.5 Sonnet, Computer Use, and Building SOTA Agents — with Erik Schluntz, Anthropic
- ▶ 1:23 Erik Schluntz I had been writing a lot of my code with things like Copilot, and I was like, wow, this is actually really cool.
- ▶ 10:46 Erik Schluntz I'd say the biggest cases are the model sort of operates at the wrong level of abstraction, and what I mean by that is the model puts in maybe a smaller band-aid when really the task is asking for a bigger refactor, and some of those, you…
- ▶ 34:31 Shawn Wang Copilot also does the plan stage before it writes code.
Agents @ Work: Lindy.ai (with live demo!)
- ▶ 20:56 Florent Crivello Um, I really like all, all GitHub actions.
- ▶ 50:10 Florent Crivello Like, look, GitLab is much smaller than GitHub.
Agents @ Work: Dust.tt — with Stanislas Polu
- ▶ 25:58 Stanislas Polu Maintaining connections to Notion, Slack, GitHub, all of them is a lot of work.
- ▶ 48:54 Stanislas Polu That needs to happen so that in semi real time, we get all the updates from, uh, Slack or Notion or GitHub into the system.
[Paper Club] Upcycling Large Language Models into Mixture of Experts
How NotebookLM Was Made
- ▶ 1:05:06 Shawn Wang I haven't, I've never, I don't see like a connect my GitHub. 2 times in the scene
[Paper Club] SWE-Bench [OpenAI Verified/Multimodal] + MLE-Bench with Jesse Hu
- ▶ 2:07 Jesse Hu Um, and the way they do that is they pull a bunch of these GitHub issues and PRs
- ▶ 35:40 Jesse Hu Like, you know, you can, I can, you can draw assets and get hub.
- ▶ 52:12 Jesse Hu Ah, Eugene, uh, the real, the question I have for Kaggle is most of it has ended up on GitHub as training data. 3 times in the scene
Building the Silicon Brain - Drew Houston of Dropbox
- ▶ 7:44 Drew Houston So what you see really working is the level one autonomy in the AI world would be like the tab autocomplete and co-pilot, right? 2 times in the scene
[Paper Club] Berkeley Function Calling Paper Club! — Sam Julien, Writer
- ▶ 30:40 unnamed speaker I'm a huge Gorilla shill and kind of a Shashir fan, you know, um, but that's, uh, the Gorilla CLI is really helpful, um, because I, I was using GitHub CLI, or the GitHub Copilot CLI for quite some time, um, but Gorilla CLI is, like, open
- ▶ 30:56 unnamed speaker and I don't have to worry about GitHub connections.
Building AGI in Real Time (OpenAI Dev Day 2024)
- ▶ 50:25 Michelle Pokrass Yeah, I'm gonna be real, I'm still VS Code co-pilot. 2 times in the scene
Language Agents: From Reasoning to Acting — with Shunyu Yao of OpenAI, Harrison Chase of LangGraph
- ▶ 35:10 Shunyu Yao Because you're just creating those like synthetic tasks.
- ▶ 38:50 Shunyu Yao I think they come up with this great idea that we should just script GitHub and solve whatever human engineers are solving. 2 times in the scene
llm.c's Origin and the Future of LLM Compilers - Andrej Karpathy at CUDA MODE
- ▶ 20:50 Andrej Karpathy They're all listed on the GitHub repo.
[Paper Club] Writing in the Margins: Chunked Prefill KV Caching for Long Context Retrieval
- ▶ 24:07 Umar Jamil And we also provide an implementation, so if you go to this URL, so github.com slash writer slash writing in the margins, you can find our implementation on how we actually do this stuff with the KVCache. 2 times in the scene
- ▶ 49:29 Umar Jamil Uh, we have, uh, you know, GitHub repository.
Building AGI with OpenAI's Structured Outputs API
- ▶ 6:58 Michelle Pokrass I was a daily user of Copilot, GitHub Copilot, and I was, like, so blown away at the quality of this thing. 3 times in the scene
- ▶ 24:27 Michelle Pokrass You basically give the model a GitHub issue and like a repo and just see how well it does at the issue, which I think is super cool.
Is finetuning GPT4o worth it?
- ▶ 5:48 Alistair Pullen And co-pilot was start. 2 times in the scene
- ▶ 39:22 Alistair Pullen If you, if you took GitHub as a data set, 2 times in the scene
- ▶ 40:08 Alistair Pullen So genie sits on top of GitHub. 3 times in the scene
- ▶ 40:10 Alistair Pullen And if you have CI running GitHub, you have GitHub Actions and stuff like that, then Genie essentially makes a call out to that, runs your CI, sees the outputs, and then, like, moves on.
- ▶ 50:55 Alistair Pullen So like the, dare I say, traditional Sweebench submission, you can go to a GitHub repo and see it and run them for yourself and verify that the numbers come out correctly.
Segment Anything 2: Memory + Vision = Object Permanence — with Nikhila Ravi and Joseph Nelson
- ▶ 21:37 Nikhila Ravi And on GitHub, we have a Colab notebook that shows how to run SAMTU on images. 2 times in the scene
How To Hire AI Engineers (ft. James Brady and Adam Wiggins of Elicit)
- ▶ 57:11 Adam Wiggins I don't know if you're GitHub, then probably all of your users and, you know, the people you want to hire are among your user base, which is a nice combination, but for most products, that's not going to be the case.
This World Does Not Exist — Joscha Bach, Karan Malhotra, Rob Haisfield (WorldSim, WebSim, Liquid AI)
- ▶ 20:58 Karan Malhotra Uh, so, with Claude's 200,000 character context length, where I can paste in entire GitHub repos, or books, or scripts, I can feed him a lot more information for more accurate simulations.
High Agency Pydantic over VC Backed Frameworks — with Jason Liu of Instructor
- ▶ 37:39 unnamed speaker They're trying to decide, oh, should I stay in my, like, high-paid fang job and just tweet this on the side and do this on GitHub?
Personal AI Meetup - Bee, BasedHardware, LangChain LangFriend, Deepgram EmilyAI
- ▶ 28:08 Ethan Sutin We can, we can send out the, I don't know what the, uh, actual githubs are called, but they're easy to find.
Why Google failed to make GPT-3 -- with David Luan of Adept
- ▶ 21:43 Alessio Fanelli People just don't have them, you know, and just, they cannot attract the talent to do it, and similarly in a software development, you have Copilot, which is the augmentation product, and then you have um, sweep.dev, any of these products,…
Making Transformers Sing - with Mikey Shulman of Suno
- ▶ 13:52 Mikey Shulman Um, it was this open source text-to-speech model, and it got a lot of stars on GitHub,
- ▶ 17:23 Mikey Shulman It got, it got a, a good number of GitHub stars and people really enjoyed playing with it because it made really realistic sounding audio.
Open Source AI is AI we can Trust — with Soumith Chintala of Meta AI
- ▶ 1:20:14 Soumith Chintala Like, G, like GitHub Copilot style usage is not captured in, say, like, LMSYS things.
A Brief History of the Open Source AI Hacker - with Ben Firshman of Replicate
- ▶ 14:34 Ben Firshman There's sometimes code on GitHub, more and more there's code on GitHub, but back and back and back then it was kind of relatively rare, but it was quite often just like scrappy research code and didn't actually run.
- ▶ 56:59 Ben Firshman We've got a team, like I worked at Docker, I've got people who worked at Heroku and GitHub and like the iOS ecosystem and all this kind of thing. 2 times in the scene
- ▶ 1:07:34 Ben Firshman And, you know, it's all open source that you can see all of this stuff on GitHub if you wanna, if you wanna like, see how we do it. 2 times in the scene
The State of AI in production — with David Hsu of Retool
- ▶ 19:13 David Hsu Uh, and so, you know, you've probably seen the Copilot, you've seen sort of so many other coding assistants, uh, like, you know, stuff like that. 2 times in the scene
- ▶ 39:25 unnamed speaker Uh, in the survey, 58 people said they used it less, uh, and 94% of them said they used it less because of Copilot and ChatGPT, which, um, 2 times in the scene
- ▶ 44:28 David Hsu Uh, now that said, I think if you're able to use Trapship or Copilot, let's say, competently, we do do that as a plus, we don't view it as a minus, but if you only use Copilot and you aren't able to reason about, like, you know, how to… 4 times in the scene
The Four Wars of the AI Stack - Dec 2023 Recap
- ▶ 6:25 unnamed speaker Yeah, I, I think it's maybe the fragmentation of the tooling, you know, like, most people in code are using VS Code, Cursor, GitHub, one of the three, so there's maybe not as much experimentation versus with text.
- ▶ 1:08:39 unnamed speaker Like, inner loop happens within your IDE, outer loop happens in GitHub, something like that.
The Accidental AI Canvas - with Steve Ruiz of tldraw
- ▶ 15:28 Steve Ruiz Uh, and I'd also opened up, uh, eventually opened up Tealdraw to be, um, like sponsorware that if you were sponsoring me on GitHub, you could, you could access it, you know, in its kind of primitive state, uh, on tealdraw.com and had like… 2 times in the scene
- ▶ 48:37 Steve Ruiz And how we've seen with Copilot, et cetera, the challenge is more on the input side than the, the output side.