alt Hacker NewsHey! I'm Nick, and I work on Integrity at OpenAI. These checks are part of how we protect our first-party products from abuse like bots, scraping, fraud, and other attempts to misuse the platform.
A big reason we invest in this is because we want to keep free and logged-out access available for more users. My team’s goal is to help make sure the limited GPU resources are going to real users.
We also keep a very close eye on the user impact. We monitor things like page load time, time to first token and payload size, with a focus on reducing the overhead of these protections. For the majority of people, the impact is negligible, and only a very small percentage may see a slight delay from extra checks. We also continuously evaluate precision so we can minimize false positives while still making abuse meaningfully harder.
Replies
It's interesting to me that OpenAI considers scraping to be a form of abuse.
It's getting to the point where a user needs at minimum two browsers. One to allow all this horrendous client checking so that crucial services work, and another browser to attempt to prevent tracking users across the web.
Nick, I understand the practical realities regarding why you'd need to try to tamp down on some bot traffic, but do you see a world where users are not forced to choose between privacy and functionality?
"we protect our first-party products from abuse like bots, scraping, fraud, and other attempts to misuse the platform"
The scary part is that you don't even see the irony in writing this.
Or, are you just okay "misusing" everyone for your own benefit?
Don’t know if it’s related to the article, but the chats ui performance becomes absolutely horrendous in long chats.
Typing the chat box is slow, rendering lags and sometimes gets stuck altogether.
I have a research chat that I have to think twice before messaging because the performance is so bad.
Running on iPhone 16 safari, and MacBook Pro m3 chrome.
Brand new account with 2 comments in this thread. How can we be sure you're not a bot deployed to defend OpenAI?
Please run Cloudflare's privacy invasive tool and share all the values it generates here so we can determine if you're a real person.
Great to hear from a first-party source. I'm a Pro subscriber and my team spends well over two thousand dollars per month on OpenAI subscriptions. However, even when I'm logged in with my Pro account, if I'm using a VPN provider like Mullvad, I often have trouble using the chat interface or I get timeout errors.
Is this to be expected? I would presume that if I'm authenticated and paying, VPN use wouldn't be a worry. It would be nice to be able to use the tool whether or not I'm on a VPN.
Hi! It's all perfectly understandable - after all, we use things like Anubis to protect our services from OpenAI and similar actors and keep them available to the real users for exactly the same reasons.
>These checks are part of how we protect our first-party products from abuse like bots, scraping, fraud, and other attempts to misuse the platform.
Can you share these mitigations so we can mitigate against you?
Its your business and your call. But my opinion is that I wish you would quit offering free services. I'm pretty concerned about the horrible effect your free services are having on education. Yes, AI can be an incredible tool to enhance education. But the reality is that it is decimating children's will to learn anything.
I don't want to blame AI for all the world's problems. And I don't want to throw the baby out with the bath water. But I think you should think really hard about the value of gates. Smart people can build better gates than cash. But right now, cash might be better than nothing. Clearly you have already thought about how to build gates, but I don't think you have spent enough time thinking about who should be gated and why. You should think about gates that have more purpose than just maximizing your profit.
"We want to hook as many people as possible without letting in our competitors" is a pretty crummy thought to use as a public justification.
(Edited for typos.)
Can't have those bots or scrapers running amok can we...
> A big reason we invest in this is because we want to keep free and logged-out access available for more users.
Are these checks disabled for logged-in, paid users?
> because we want to keep free and logged-out access
But don't you run these checks on logged-in users too?
But is the title true, is typing specifically blocked? Or does it just block submitting the text?
I ask because I have seen huge variations in load time. Sometimes I had to wait seconds until being able to type. Nowadays it seems better though.
Y'all just salty that DeepSeek et al are training their LLMs on yours
I don't trust what OpenAI says. Sam Altman gives shivers, and these kinds of blog posts make things look even worse.
sometimes I paste giant texts (think summarization) in the chatgpt (paid) webapp and I noticed that the CPU fans spin up for about 5 seconds after, as if the text is "processed" client side somehow. this is before hitting "submit" to send the prompt to the model.
I assumed it was maybe some tokenization going on client side, but now I realize maybe it's some proof of work related to prompt length?
Can you fix the resizing text box issue on Safari when a new line is inserted? When your question wraps to a newline Safari locks up for a few seconds and it's really annoying. You can test by pasting text too.
> "abuse like bots, scraping"
You what, mate? Would you please use that on yourselves first? Because it comes off as a GROSS hypocrisy. State of the art hypocrisy.
>> behavioral biometric layer
But this one, especially, takes the cake.
Quite disgusting.
> we want to keep free and logged-out access available for more users
How does this comport with OpenAI's new B2B-first strategy?
> We also keep a very close eye on the user impact
Are paid or logged-in users also penalised?
"abuse like bots, scraping, fraud, and other attempts to misuse the platform"
This has to be a joke, right?
> Hey! I'm Nick, and I work on Integrity at OpenAI. These checks are part of how we protect our first-party products from abuse like bots, scraping, fraud, and other attempts to misuse the platform.
How can first-party products protect themselves from abuse by OpenAI's bots and scraping?
> OpenAI: These checks are part of how we protect products from abuse like bots, scraping, and other attempts to misuse the platform.
This would be fucking HILARIOUS if it wasn't so tragic.
Tangential question: are there chatgpt app devs on X? There are a few from Codex team but I couldn’t find guys from “ordinary” chatgpt.
Also if you could pass this over: it takes 5 taps to change thinking effort on ios and none (as in completely hidden) on macos.
If I were to guess it seems that you were trying to lower the token usage :-). Why the effort is only nicely available on web and windows is beyond me
I shouldn't be giving ideas to your boss, but I bet he would be interested in making ChatGPT available only by paying customers or free for those whose who gets their eyes scanned by The Orb. Give 30 days of raised limits and we're all set to live in the dystopia he wants.
Fake Account
Have you given any thought to what we trade when big tech elects one corporation as the gatekeeper for vast swaths of the Internet?
Then make sure they only target the free tier!
Hi Nick, your software is a horrendous encroachment on users' privacy and its quality is subpar to those of us who know what we're working with. We don't use your product here.
You’re doing gods work sir, thank you!
the irony of your statement is hilarious, disappointing, and infuriating.
[dead]
> we protect our first-party products from abuse like bots, scraping, fraud, and other attempts to misuse the platform
Have you just described the dilemma facing all the content sites used to train LLMs?
That still doesn’t explain why you can’t even start typing until that check proceeds. You could condition the outbound request from being processed until that’s the case. But preventing from typing seems like it’s just worse UX and the problem will fail to appear in any metrics you can track because you have no way of measuring “how quickly would the user have submitted their request without all this other stuff in the way”.
Said another way, if done in the background the user wouldn’t even notice unless they typed and submitted their query before the check completed. In the realistic scenario this would complete before they even submit their request.