That's sort of true, although in reality Airbyte was only truly "open source" for a very small period[0].
In reality, since about 1 year into the project, it's operated with a mix of open and "less open" licenses for different parts of the codebase, in a way that would make it difficult to just use the MIT licensed bit.
I think that kinda proves the point you were going for.
This feels like an opposite of my experience. I find just having a bunch of Claude Code terminals running in the background to be the most "ambient", while I find autocomplete/auto-navigate the likes of Copilot/Cursor to be so annoying that I don't use any AI autocomplete now. Regardless, I think there's still a lot of room for structured AI programming flows, especially regarding semantic search, code flow tracing and intelligent find and replace.
They're using Column (https://column.com/) under the hood, so more like Stripe (payments + Atlas) for non profits I think? Still very powerful and material value of course on top of the banking partner primitive.
> I think they’re recovering. They’ve learned a few lessons, including not to be too in hock to a few powerful and wealthy individuals.
I do not believe the EA movement to be recoverable; it is built on flawed foundations and its issues are inherent. The only way I see out of it is total dissolution; it cannot be reformed.
Which of the foundations is flawed, the "we have the ability to help others and should use it" or the "some ways of helping others are more effective than others"?
I barely use the autocomplete features of Cursor, and for agentic coding, Claude Code blows Cursor Agent out of the water. I don't think Cursor has anything that cannot be replicated in a week or two other than the first mover advantage; certainly not an advantage that cannot be justified at 30B+ valuation.
It's more than that. They have both their own completion model and now agentic one. It's not a basic fine-tune, because it's faster than anything else available out there, so there's something interesting in the architecture itself.
yeah its not fair to call it a finetune because finetune carries connotation of "there wasnt that much extra compute and data added". RLFT has a lot more added to it as Sasha alluded in his talk https://www.youtube.com/watch?v=md8D8eNj5JM - the x axis is log scale, think about that
the framing here is more about "why would you start from random weights when perfectly good starting weights exist" https://www.latent.space/p/fastai
they are building their own editor (granted they didn't do it from scratch); they do build their own models (see composer);
they may not have done a lot of this from scratch but there's still a lot of innovation in what they're doing. they're also building a pretty fantastic product and clearly the leader today in AI innovation for IDEs.
may not be everyone's cup of tea; but i think you might be detracting some of their innovation.
I expected this to happen. I knew people who were involved in the organization who were unnecessarily chummy to TPOT/Postrat/FTX culture before it blew up.
You made a right decision; The big-world social platforms are doomed to fail, as they have quadratic scaling costs. I am a big proponent of "curation over moderation".
I am working on Mikoto Platforms, which is basically designed to be somewhere between Discord and Notion, but open source + decentralized. There is no global feed; in fact, some users thought it was a bug that it took you to such an empty screen when you first started it. Platforms are what you make of it.
Oh, sweet! This is an awesome-looking app! The tree-view of the discussion subjects or threads is pretty inspired. I really like that idea, especially as a way to moderate a thread that is purely managerial, rather than punitive or rewarding. Of course, it would be great if a discussion could always be efficiently branched by the participants in real time, but I imagine any good "posterity" thread would benefit from some maintenance. In any case, it's something I hadn't really considered before. Cool stuff!
I mean, the whole app is pretty impressive. It's clear, from the blog posts, that we're of like minds when it comes to the "good parts" of the internet and what makes the "bad" parts so frustrating. If I used Discord more this might mean more, but I'm pretty satisfied to swap out my Discord usage for Mikoto platforms. Seems like a win to me, at least.
All that said, I am still left a bit unsatisfied as far as a social media replacement. A "chat" presentation eschews the kind of "press release" posts that social media accommodates well for creatives. At the very least, I wouldn't feel "normal" posting announcements that didn't necessarily expect follow up engagement in a chat-style environment. Posting "here's this new thing I did" on a social media feed without any follow up is...lonely, but not particularly uncommon. But posting the same thing to an empty or unresponsive chat room is...somehow sadder? To me at least.
Didn't Airbyte rugpull their license to ELv2?
reply