Users spend tokens on DeepSeek, Claude, and others -> generate variations of projects uploaded to GitHub -> Microsoft (GitHub's owner) -> trains LLMs on this -> ends up with better models.
No matter what model a user may use at any price point, microsoft can distill it.
It's distillation at a higher level
most good programmers aren't going to commit garbage solutions; they'll push it to a git remote once they see it actually solves a problem they are trying to solve, and since the git commits and their messages are also AI generated, you've good visibility into the user's intention and resulting edits to the codebase.
GitHub lets me create public and private repos.. with a free account.. how are they making money?! Why are they throwing hardware at this explosive growth of repos with LLMs?
I got an email a couple weeks ago that said my credit card expired and GitHub couldn't renew my subscription. I honestly don't know what that subscription is getting me though, what am I paying for? I used to pay for private repos I think, but that's included for free now.
Did you read the numbers? The volume growth requires a top down examination of everything, which is not going to be fast or easy to do. From 65 million to 7 billion inside a couple years is a ridiculous surge.
At its GitHub Universe conference on Wednesday, the company announced that it is now seeing 7.4 billion commits per month. To put that into perspective, developers pushed 65 million monthly commits to the platform in 2024.
Here's a thing
Users spend tokens on DeepSeek, Claude, and others -> generate variations of projects uploaded to GitHub -> Microsoft (GitHub's owner) -> trains LLMs on this -> ends up with better models.
No matter what model a user may use at any price point, microsoft can distill it.
It's distillation at a higher level
most good programmers aren't going to commit garbage solutions; they'll push it to a git remote once they see it actually solves a problem they are trying to solve, and since the git commits and their messages are also AI generated, you've good visibility into the user's intention and resulting edits to the codebase.
They want to capture it all.
GitHub lets me create public and private repos.. with a free account.. how are they making money?! Why are they throwing hardware at this explosive growth of repos with LLMs?
They're using our data to create AI products that they sell to whales who subsidize our usage. Whether that's long term viable, I don't know.
github is owned by Microsoft, their intention is not just to make money off of it
it's like Gmail and Google
I got an email a couple weeks ago that said my credit card expired and GitHub couldn't renew my subscription. I honestly don't know what that subscription is getting me though, what am I paying for? I used to pay for private repos I think, but that's included for free now.
well, then do something about it ?! this seems like an insane solution. just keep throwing hardware at it until you run out of cash, i guess
Did you read the numbers? The volume growth requires a top down examination of everything, which is not going to be fast or easy to do. From 65 million to 7 billion inside a couple years is a ridiculous surge.
This must be squeezing GitHub per user margins, no?
Related:
Octoverse: Commits, pushes, reach new highs in 2026
https://github.blog/news-insights/octoverse/octoverse-commit... (https://news.ycombinator.com/item?id=50280770)
160pb in SSD?
[flagged]
[dead]