Obligatory don't put all your eggs in one basket.

There's countless alternative git forges, self-hosted or not... I suggest people use multiple for redundancy, even if they are just read-only backups.

Maybe it would be nice if git had a way to add multiple repo URLs for the same remote.

It exists apparently: https://stackoverflow.com/questions/849308/how-can-i-pull-push-from-multiple-remote-locations

Wouldn't this be pretty bad for a lot of projects if people can no longer access them without an account?

Also open source, but not really you must make an account first?

I feel like they do this to enforce some sort of standard for how the files are named. Because ultimately, the relevant files are also available in the file tree itself. It's not that different to click on one of those tabs vs. clicking on the file a few centimeters above...

I'm not against this "Accessibility.md" - but I'm thinking, is there a point in having all these specific, 'get a specific tab by filename' files?

Like instead of a license.md and Accessibility.md - they could just make it dynamic? Like Whatever.md would result in a 'whatever' tab - for example.

Well, it does say in the blog post:

extending the in-context blocking controls already available for issue and pull request comments.

So, I assume so. But yeah, terrible choice to refer to that as "discussion comments"...

Oh, this is about discussions specifically? Were you able to block from issues before?

Don't think many folks use the "Discussions" for actual discussions. Those mostly happen on the issues and PRs after all. So, there was probably never much disagreement or trolling on the capital-D Discussions, until presumably LLM spambots realized they can't get blocked there...

That’s landing surprisingly late for a social platform.

You don't need consent. It's better to have an open dataset, since all the big AI companies have their own private dataset anyways (which isn't opt-out).

So it starts I guess. Hugging face holds basically all the open source data and models, the government will start going after them now that they are getting copyright wins. It's regulatory capture. They are doing it at the behest of big AI companies to kill the competition.

so instead of fixing the problems, they just hid the problems.

Lol good luck trying to make one of the most notoriously spiteful group of workers comply with that one. (I am in that group btw)

I am pretty sure they won't rehire them instead will increase workload of others

it's also the most freudian event some of us were waiting for. It will be quite the punch in the gut for shitty bean counter types who fired developers that made their record profits.

Good luck maintaining your custom blog with 35000 of bullshit code. This has to be the most obvious rugpull

With a 1:1 usage:cost relationship, how else could this be priced? Other than to take the 90th percentile usage and peg pricing there - which would just have subscriptions cancelled outright.

That's a .. new feature?

I presume it's mainly because github is being targeted orders of magnitude more.

I'm skeptical about the whole load increase leads to stability issues. Scaling should improve performance, not solve stability issues like stuff not working or breaking.

The article goes into it briefly, but for most of the rest of the article assumes the usage-increase-cause to be valid and correct.

Step 1: Push AI hard
Step 2: Be surprised by platform usage spikes

Well in that it serves a free website rather than just being a bare repo. Still don't think it fits exactly but its better.

Fair point — I get where you’re coming from. I’m still experimenting with distribution, so this setup isn’t final. GitHub Pages might be a cleaner option, I’ll check it out.

I didn’t know GitHub Pages was a thing for this use case. I’ll look into it — it might be a better fit for future projects.

I assume it's not against GitHub terms, but I find it inappropriate and unfitting. Smells like marketing for discoverability with no content, just forwarding. Which is a thing of course, but if you're asking me as a developer and personally, I'd prefer the project not do that. A repo as a pure landing page uses nothing of what makes a repo a repo.

If it's a developer creating their portfolio and linking their distributed projects to their developer profile, I think I can see some value in that. But in that case, all of them should be in one repo.

I personally find it weird and confusing when I see things like this. There are loads of closed source projects that use GitHub purely for issue tracking and feature requests and it just kind of feels against the ethos of it. Honestly for those projects I'm not really interested anyway if it is closed so after a quick browse in the org repos to see if I missed something, I just leave.

In your case it seems even stranger, if you wanted to have a landing page in GitHub but nothing else then why not have it as an GitHub Pages website rather than a repo?

“Copilot, throw together a response to the big name guy calling us out”