> because we want to keep free and logged-out access

But don't you run these checks on logged-in users too?

Yep, on logged-in users too. The reason is basically the same: we want scarce compute going to real people, not attackers. Being logged in is one useful signal, but it doesn’t fully prevent automation, account abuse, or other malicious traffic, so we apply protections in both cases.

▲

salawat 4 hours ago | parent | next [-]

More like "We want your money, but don't want to provide service." Are you sure OpenAI isn't morphing into a finance/insurance company?

▲

pixl97 4 hours ago | parent [-]

While OAI is one of the more hypocritical of the bunch, it is not uncommon for paid services to have some limitations in their terms of service. Like going in a store and buying stuff, it doesn't me a free for all doing whatever you want.

	▲	zamadatix 2 hours ago \| parent [-]
		Limitations on the ChatGPT subscription should have to do with the usage limits of the tier you paid for (and I don't think anyone has a problem with that). If I'm in the limits of requests I paid for then it's usage rather than abuse. "Abuse" checks should only come into play when someone tries to leverage the free tier. It reminds me of those cable companies that try to sell "unlimited" plans and then try to say customers who use more than x GB/month are abusing the service rather than just say what the real limits are because "unlimited" sounds better in marketing.

▲

angoragoats 5 hours ago | parent | prev | next [-]

Nothing you do can fully prevent automation. Someone who wants to automate requests badly enough will be able to do it, especially when the “protections” are as easy to decrypt and analyze as the OP proved.

Meanwhile, the rest of us (well, not me, because I don’t use your garbage product, but lots of others do) have to suffer and have our compute resources used up in the name of “protection.”

	▲	3form 4 hours ago \| parent \| next [-]
		Yeah, that's it. Also, it is a bit amusing to me - "We want to prevent automation", says the employee of Let's Automate Inc.
	▲	geetee 4 hours ago \| parent \| prev [-]
		[flagged]

▲

jorvi 4 hours ago | parent | prev [-]

I'm glad you guys at least went with CloudFlare. LMarena went with Google's ReCaptcha, which is plain evil. It'll often gaslight you and pretend you failed a captcha of identifying something as simple as fire hydrants. Another lovely trick is asking you to identify bridges or busses, but in actuality it also wants you to identify viaducts or semi-trucks.