Easy to criticize this, but I rather see GitHub survive than fail under its own success. So thanks for acting on this MS!

They were originally on their own datacenters + huge amounts of burst and ancillary stuff in AWS, the internal push to move away from "the competitor's cloud" after the acquisition was huge and entirely stupid. I'm ex-GitHub and was one of the annoying people constantly saying GitHub should only move where it was provably the best option for GitHub - the third best major cloud provider is likely never it on merits alone.

If this story is true, it's good that they finally realised that GitHub's performance and availability mattered more than using Microsoft's products. It would mean someone finally came to their senses rather than forcing a wholesale push to Azure - but I bet they still want to have it both ways even if they concede some AWS now.

In the early 2000's they did the same dumb thing to WebTV: forced a move from Solaris to Windows, that completely wrecked the engineering team's velocity and motivation.

They did the same thing to Hotmail as well. It was fine running on FreeBSD/Solaris. Microsoft didn't need to switch to windows for no good reason.

Does any of the LLM providers actually use llms.txt?

If I remember correctly this "standard" was setup by someone but without involvement of any of the major AI players.

I can definitively say llms.txt is not used by any AI players. I run a blogging platform with around 80k blogs and /llms.txt is not requested by anything (other than humans checking to see if there's an llms.txt path).

All regular pages are aggressively scraped to the extent it's a problem I have to consistently manage, but not llms.txt.

Amazing, I didn't know.

So it get even stranger, I am the only one reading those /llms.txt ...

I'm seeing quite a bit of request for these on my work's GitBook documentation site.

But perhaps these are developers specifically targeting these pages to feed whatever LLM they are using.

How is a static blog being scraped a problem? Do you not use a CDN?

> a blogging platform with around 80k blogs

But nah, I'm sure OP doesn't know about CDNs.

Are all blogs static though?

OP clearly meant that the AI players are not reading and/or honouring llms.txt of other websites when scraping.

i stand corrected, but what was clear to you, obviously was not clear to me.

No, requesting "Accept: text/markdown" in the headers and returning markdown is the more agreed upon standard at this point.[0]

[0] - https://acceptmarkdown.com/

Now, it would be super cool to get markdown and zero javascript bundles…

If you want to see what that looks like, I one-shot a browser with Claude that does it[0]. Docs pages are early adopters to this[1][2], so that AI agents can better handle tasks.

[0] - https://github.com/solumos/md-browse

[1] - https://docs.stripe.com

[2] - https://vercel.com/docs

I just found out Cloudflare supports real-time html to md conversion [0]

- [0] https://blog.cloudflare.com/markdown-for-agents/#convert-htm...

This is interesting. I should start incorporating this -- it couldn't hurt to do both.

yes, they do.

anyone who's, even slightly, clued into how agents access documentation, has been making changes to their pages. ex: https://searchtxt-web.fly.dev/search?q=aws

Still Go/Rust/Typescript is on a whole other level of adoption compared to Zig

And Azure Flexible server please

Curious on what (how much) hardware your running this.

Currently running off

AMD EPYC 7543 x2 for 64 cores/128 threads

512 GB RAM

~ 90 TB of PM9A3 SSDs across 12 physical devices

Storage is not very full though. I'm probably using about a third of it at this point.

Really? I'm daily driving JetBrains IDEs on Apple M3 and don't recognize any if this. Just give it a bunch of extra heap memory (eg 4g instead if 1gb) and it's fast!

I have given it up to 30GB of heap and i tried many different GC configs, i even ran it and my project on ramdisk.

The issue is related to using a monorepo with lots of code in different languages - openening single folders is fine. Ut i want to be able to work on dozens of services in a single window, all other editors manage just fine

I have a similar problem, large monorepo, things have become really bad lately to the point that the cursor is unresponsive. The only workaround I have found is opening each folder of the monorepo in its own IDE instance

Feel this so hard. The opposite is also true where you have a micro-service architecture and cursor faceplants in workspaces with multiple repos. We ended up building cortex.build partly because of this exact pain. Our context engine builds a git-aware dependency/provenance graph so it can stay local and only pull the relevant slice across a massive repo or dozens of smaller ones.

Well the docs say the old way (no Kafka) is on its way out:

"However, this architecture is set to be deprecated in a future release."

So it doesn't stay optional unfortunately. It quite a heavy dependency to include...

Well it only needs to be Kafka-compatible, so personally I hope that things like RedPanda will turn out to be easier to run for this use case than actual Kafka.

Having an S3-compatible store was already a fairly heavy dependency in terms of something to run correctly in production, it's just that most people don't even consider running their own object store at any real scale, they just go to cloud. Whereas running your own Kafka is something more platform teams are already attempting.

Couple of years ago us-east was considered the least stable region here on HN due to its age. Is that still a thing?

When I was there at aws (left about a decade ago), us-east-1 was considered least stable, because it was the biggest.

I.e. some bottle-necks in new code appearing only _after_ you've deployed there, which is of course too late.

It didn't help that some services had their deploy trains (pipelines in amazon lingo) of ~3 weeks, with us-east-1 being the last one.

I bet the situation hasn't changed much since.

>It didn't help that some services had their deploy trains (pipelines in amazon lingo) of ~3 weeks, with us-east-1 being the last one.

oof, so you're saying this outage could be cause by a change merged 3 weeks ago?

Couple of weeks or months ago the front page was saying how us-east-1 instability was a thing of the past due to <whatever chang of architecture>.

Yup, never add anything new to us-east-1. There is never a good reason to willingly use that region.

Yes

Adobe Flex with Adobe Catalyst. Design a GUI in Photoshop, export it to Flex/Flash to add interactivity.

Looked cool during demos. Got killed when Flash died.

Exactly, also requires authentication. How can this be 10/10?