Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

We stopped AI bot spam in our GitHub repo using Git's –author flag

archestra.ai

Post reply on HN
Print view View original

111–120 of 254 posts

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#111

Post by xdennis »

It's quite ironic to complain about AI slop in a piece that's quite clearly AI slop.

Soon there will be no more AI doomer comments. The bots will take over that job too.

---

I'm working for an open source company, and my God, are 95% of contributions useless.

There are really dumb ones where the bot writes 10 paragraphs about how he implemented the feature, but the entire changeset is adding one line to .gitignore or adding a CLAUDE.md file.

There are even worse ones where the bot submits 3000 lines of code that seemingly works, but you have to spend an hour to figure out why it doesn't work.

The dumb ones are so much better.

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#112

Post by rvnx »

smaudet wrote: post #104
rvnx wrote: post #65

Earlier quoted context omitted.

It's called Brave

Which is not chrome and still has ads...(Ironically). The issue here is the core model is broken (misaligned incentives). That's not something you are going to fix with a github "downstream". A token system could help but it's easy to imagine ways that could be gamed, if not implemented well.

Ads are the main business model of browsers.

If search ads are blocked on search engines, then there is no revenue for the browser. It's that simple (on top of that Brave has other revenues, but the majority is search ads).

So it's a game of hoping that the majority won't change the default.

This is the main reason Brave does not block search ads specifically by default, but still block the other ads. Blocking the other ads there are no consequences, since anyway this revenue is not shared back to the browser.

This is why the business model of Brave is cynical.

-> It's the same model as AdBlock and the "Acceptable Ads" (block all ads, except the acceptable ads, unless you disallow them)

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#113

Post by Dwedit »

infinitifall wrote: post #54

Is the solution to everything simply more catgirls [1]? Proof-of-work was, after all, about countering email spam. PR spam is but the latest in that long tradition. 1- https://anubis.techaro.lol

Anubis is actually not a cat. The original Egyptian deity is a god of death, and has a canine head. Anime catgirls and dog girls can look similar at first glance.

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#114

Post by embedding-shape »

bradley13 wrote: post #96
embedding-shape wrote: post #62

Sounds kind of weird that the blog post complains about `poisoning the conversation with pointless "implementation plans"` when literally they ask for that, after attaching $900 USD bounty to a very under-specified issue, and even replies with "Do you have an implementation plan in mind?" to some of the first "attempters". Sounds like they got exactly what they'd been asking for, and even before LLMs if you pulled so…

It's fine for developers to provide a plan, even if it gets rejected. The problem comes when every script kiddie figures AI has made them into a developer. Imagine you want to get a doctor's opinion, or maybe a couple of opinions. But a zillion AI-amateurs have registered themselves as doctors. How do you separate wheat from the chaff?

> Imagine you want to get a doctor's opinion, or maybe a couple of opinions. But a zillion AI-amateurs have registered themselves as doctors. How do you separate wheat from the chaff?

Right, but that's not what happened though.

Someone went to the public square, said "Hey, I'm looking for any sort of doctor, and I'll pay you $900 if you tell me your plan and then whatever plan I chose wins" and then they get surprised they get flooded by zillion AI-amateurs.

You don't generate a ton of chaff then try to find the wheat, you ensure your process doesn't generate a ton of chaff in the first place. Offering large monetary rewards for relatively simple work for anyone in the public is bound to generate a ton of chaff...

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#116

Post by smaudet »

rtdq wrote: post #97
smaudet wrote: post #90

Earlier quoted context omitted.

I don't think that is a (very realistic) concern. AI is slop, the problem is not that the real contributors are struggling to get PRs merged. The bigger issue being, raising the bar to students who may have otherwise had productive careers (but education is a general issue, where the students don't even yet recognize they are being scammed).

I don't follow, and I'd be concerned that this opens up a cottage industry of bots generating plausible looking repositories that unwitting contributors would attempt to contribute to. We already know that bots are astroturfing repos to generate overinflated star counts. I'd say the least crap option here is to honeypot PR contributions from bots

This feels like bot logic, lol.

Unless the contributors don't care about the repos they contribute to, this is not a likely scenario. AI doesn't care. We do.

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#117

Post by rglullis »

applfanboysbgon wrote: post #82
rglullis wrote: post #63

'I will take "problems that could be easily be solved by implementing a Pfand system" for $200, Alex.' Seriously. Just ask for a US$10 deposit for the each PR. If the PR is accepted (not even merged, just accepted as "this is a good effort"), give it back. Hell, give double the amount for good effort and you got yourself a cheap way to attract good contributors. Best case, bots will balk at the payment. Worst case, t…

This is an evergreen internet comment right here. Condescendingly proclaiming "This problem could be easily solved by [significantly worse solution that had 1/10th the thought put into it as the actual solution by people with a stake in actually solving the problem rather than making quippy armchair comments]". --- I know it's against convention to comment on downvotes, but really? Really? This is controversial? The…

Github already has the payment infrastructure.

Polar.sh is already doing things that are a lot more complex in this space.

If you are in a civilized country which allow direct payments (i.e, anything but North American nowadays) and you don't want to deal with Github or any external system, there is always good old "make a M-PESA/SEPA/Pix/UPI transfer to account XYZ")

> the thought put into it as the actual solution by people with a stake in actually solving the problem

Let me flip your argument: think of how much time and thought is poured into problems like this one by people who don't even try to implement a Pfand system beforehand.

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#118

Post by pocksuppet »

drum55 wrote: post #93
infinitifall wrote: post #54

Is the solution to everything simply more catgirls [1]? Proof-of-work was, after all, about countering email spam. PR spam is but the latest in that long tradition. 1- https://anubis.techaro.lol

Proof of work doesn’t work here same as it doesn’t work for email. The effort to mint a valid PoW is always going to put the legitimate user at a disadvantage, whatever the implementation is. Someone with an incentive to spam will always be able to do it faster, more efficiently than you. You can’t submit a PR because your laptop is too slow? Rent some hash rate from someone, and now you’ve just made a system of payi…

Just like we did with IP addresses. If yours is blocked by Cloudflare, you can pay a botnet operator a few dollars to use theirs! You can even use your credit card through a mostly-legitimate website. It's very convenient.

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#119

Post by petterroea »

skydhash wrote: post #36
petterroea wrote: post #17

What I see is a (clever) hack, and GitHub continuing to provide good tools to its users.

What I see is a solution for a problem that is self inflicted, meaning lumping contributors and generic internet users in the same workflow. In big projects, you have the core team, a handful of well known contributors, and everyone else. I strongly prefer the git email model, where it’s often trivial to control the flow of changes proposal. GitHub does not have the same wealth of tools and versatility.

I seem to have completely missed the "failing" when writing "this is GitHub failing to deliver tools". My bad

Re: We stopped AI bot spam in our GitHub repo using Git's –author flag

#120

Post by motakuk »

FartyMcFarter wrote: post #38
zer0tonin wrote: post #6

> Should we stop giving fun test tasks to our job candidates? Yes

It seems this particular company makes a payment for completing those tasks, so it might not be that bad.

We do, it's a part of our hiring pipeline: https://archestra.ai/careers
Post reply on HN