Hacker News Reader

Comment by tlogan | Hacker News Reader

The actual number is that 98% have less than 2 stars (0 or 1). About 90.25% has zero stars.

So Claude repos are statistically more likely to have stars than the average GitHub repo. Not the conclusion the headline was going for.

But the header is just "90% of Claude-linked output going to GitHub repos w <2 stars". No conclusion, just some random fact.

The problem is that this title is editorialized, and the fact is cherry-picked. Why not =0? Why not >1000? This is just a dashboard, it highlights "Interesting Observations", but stars statistics is not there.

Sounds like Claude commits are, on average, going into higher visibility repositories than humans… maybe the author would like to reconsider their approach?

Well, you can't reconsider your approach when you don't like the results.

If anything, the fact that this is what he arrived at, even when starting with the opposite position, is proof of the validity of this result.

Yes, stars can mean a lot of things.

- visibility

- popularity (technical, domain, persona)

- genuine utility

- novelty

...

There are also plenty of super high utility repos that are widely used (often indirectly), but don't have a lot of stars, or even a meagre amount.

Also there is the issue of star != star, because it's not granular.

It's similar to upvotes on general social media platforms. Everyone likes cute cats doing funny things somewhat, but only few people appreciate something that's more niche but way more impactful, useful or entertaining (or requires some effort to consume), but those who do, value it very highly. But the same person might use the same score (single upvote) for a cat video and a video that they value much higher.

You should check recent commits, because obviously there are a lot of forked 0 star repos.

I think this is useful in answering the grandparent comment's question:

stars : uniq(k)

1 : 14946505

10 : 1196622

100 : 213026

1000 : 28944

10000 : 1847

100000 : 20

each line (mostly) being equal length provides me an odd comfort

power law distribution ~1/x I think

Zipf's law?

only 80 more repos need 100000 stars and all lines would be equal! e.g.

1 : 14946505

10 : 1196622

100 : 213026

1000 : 28944

10000 : 1847

100000 : 100

you would lose 80 repos from "10000 : 1847" also in that case.

interesting that you only need ~150 stars on a project for it to be in the top 1%

Let's establish a roving band of ~150 GitHub users that go around 1% things.

Funny how everyone gravitated towards analysis of the star distribution of REPOS when the headline claim is on ADDITIONS. If you look at my comment below (invite you to verify the stats), the distribution of additions by star count is far more weighted to 2+ star repos in GitHub overall. The observation is meaningful, up to the observer to draw a conclusion. Is Claude just speeding up output or is it generating piles of spaghetti code with no use? Considering the get rich quick economy that has sprung up around app development, I'm inclined to at least consider the latter.

How do you know that?

https://ghe.clickhouse.tech/

Reply on news.ycombinator.com