eBay explicitly bans AI "buy for me" agents in user agreement update • Hacker News Board

Re: eBay explicitly bans AI "buy for me" agents in user agreement update

#361

Post by K0balt » Mon, Jan 26, 2026, 10:41 PM UTC

JohnFen wrote: post #85

Fwirt wrote: post #76

Earlier quoted context omitted.

This was my belief for many years, but then I tried sniping (with the same prices I was putting as my maximum bid before!) and my success rate skyrocketed and the prices I was paying dropped. It seems that despite repeated reminders and explanations, there are three groups of people using eBay "incorrectly" that make the sniping strategy viable: 1) People who do not understand proxy bidding and think that they "need"…

> For some reason, auctions with more bidders seem to attract more bidders, whereas auctions with zero bids seem to go unnoticed. Huh. I'm a "buy it now" guy and filter out the auctions, but maybe I should start looking for zero bid auctions too.

Slightly misspelled or otherwise poorly listed items, and check auctions for “ending soonest” at awkward hours.

Re: eBay explicitly bans AI "buy for me" agents in user agreement update

#362

Post by bhaviav100 » Tue, Jan 27, 2026, 7:30 AM UTC

I’ve been working on a small experimental gateway that sits between agents and customer-facing execution paths and forces decisions through policy + approval before anything irreversible happens.

v2 I just shipped adds:

• risk scoring on drafts • policy path traces • approval chain previews • highlighted spans showing what triggered the block • admin review endpoints

The motivation is exactly what people are pointing at here: once agents can transact, marketplaces end up banning them unless there’s a way to pause, inspect, and assign responsibility at execution time.

Curious what failure modes you’d want intercepted first if eBay or Amazon ever exposed agent purchase APIs.

https://authority.bhaviavelayudhan.com/v2/console

Re: eBay explicitly bans AI "buy for me" agents in user agreement update

#363

Post by mminer237 » Tue, Jan 27, 2026, 3:12 PM UTC

direwolf20 wrote: post #336

mminer237 wrote: post #220

Earlier quoted context omitted.

Is so far as you don't have to use the site, that's true, but they are legally enforceable, and you could absolutely be sued for breaking them if you upset eBay enough.

Then most websites can sue OpenAI and Anthropic, but in reality, they can't.

My understanding from hiQ v. LinkedIn is that it depends on whether you have access controls contingent on the terms' acceptance. If you have to make an account and agree to the terms to view content or place bids, it's illegal to violate what you agreed to. If you didn't explicitly agree to anything, you don't bind yourself by viewing the site, so breaking them isn't illegal. However, robots.txt qualifies as an access control and therefore is legally binding, so if an AI company is breaking that, they probably could be sued (see: eBay v. Bidder's Edge), but OpenAI follows robots.txt to my knowledge. Maybe the Internet Archive would be liable.

I haven't followed all the appeals though, and I'm not your lawyer.