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

OpenAI threatens to revoke o1 access for asking it about its chain of thought

twitter.com

Post reply on HN
Print view View original

191–200 of 323 posts

Re: OpenAI threatens to revoke o1 access for asking it about its chain of thought

#191

Post by rvnx »

tarruda wrote: post #185
sweeter wrote: post #115

Im pretty sure its just 4.0 but it re-prompts itself a few times before answering. It costs a lot more

Seems like Reflection 70b was an attempt to implement the same concept on top of Llama 3 70b

Could even be that Reflection 70b got hyped, and they were like "wow we need to do something about that, maybe we can release the same if we quickly hack something"...

Pushing an hypothetical (and likely false, but not impossible) conspiracy theory much further:

in theory, they had access in their backend logs to the prompts that Reflection 70b were doing while calling GPT-4o (as it apparently was actually calling both Anthropic and OpenAI API instead of LLaMA), and had an opportunity to get "inspired".

Re: OpenAI threatens to revoke o1 access for asking it about its chain of thought

#192

Post by PufPufPuf »

blibble wrote: post #159
kgeist wrote: post #132

Earlier quoted context omitted.

In my country, it's illegal to charge different people differently if there's no explicitly signed agreement where the both sides agree to it. Without an agreement, there must be a reasonable and verifiable justification for a change in the price. I think suddenly charging you $100 more (compared to other consumers) without explaining how you calculated it is somewhat illegal here.

where's this? the soviet union? this completely rules out any form of negotiation for anything, ever

It doesn't rule out negotiation. That's what the part about a written agreement is for.

It merely rules out pulling prices out of thin air. Which is what OpenAI is doing here, charging for an arbitrary amount of completely invisible tokens. The shady part is that you don't know how much of these hidden tokens you would use before you actually use them, thus making it possible to arbitrarily charge some customers different amounts whenever OpenAI feels like it.

Re: OpenAI threatens to revoke o1 access for asking it about its chain of thought

#193

Post by contravariant »

IncreasePosts wrote: post #33
contravariant wrote: post #4

Okay this is just getting suspicious. Their excuses for keeping the chain of thought hidden are dubious at best [1], and honestly just seemed anti-competitive if anything. Worst is their argument that they want to monitor it for attempts to escape the prompt, but you can't. However the weirdest is that they note that: > for this to work the model must have freedom to express its thoughts in unaltered form, so we cann…

Or, without the safety prompts, it outputs stuff that would be a PR nightmare. Like, if someone asked it to explain differing violent crime rates in America based on race and one of the pathways the CoT takes is that black people are more murderous than white people. Even if the specific reasoning is abandoned later, it would still be ugly.

Could be, but 'AI model says weird shit' has almost never stuck around unless it's public (which won't happen here), really common, or really blatantly wrong. And usually at least 2 of those three.

For something usually hidden the first two don't really apply that well, and the last would have to be really blatant unless you want an article about "Model recovers from mistake" which is just not interesting.

And in that scenario, it would have to mean the CoT contains something like blatant racism or just a general hatred of the human race. And if it turns out that the model is essentially 'evil' but clever enough to keep that hidden then I think we ought to know.

Re: OpenAI threatens to revoke o1 access for asking it about its chain of thought

#194

Post by NicuCalcea »

edwinarbus wrote: post #180
int_19h wrote: post #76

The best part is that you still get charged per token for those CoT tokens that you're not allowed to ask it about.

Edwin from OpenAI here. 1) The linked tweet shows behavior through ChatGPT, not the OpenAI API, so you won't be charged for any tokens. 2) For the overall flow and email notification, we're taking a second look here.

So we're allowed to ask about the chain of thought via the API?

Re: OpenAI threatens to revoke o1 access for asking it about its chain of thought

#195

Post by error9348 »

edwinarbus wrote: post #180
int_19h wrote: post #76

The best part is that you still get charged per token for those CoT tokens that you're not allowed to ask it about.

Edwin from OpenAI here. 1) The linked tweet shows behavior through ChatGPT, not the OpenAI API, so you won't be charged for any tokens. 2) For the overall flow and email notification, we're taking a second look here.

will these (or have these) notifications been paused while the decision is being reconsidered?

Re: OpenAI threatens to revoke o1 access for asking it about its chain of thought

#197

Post by RivieraKid »

mrinterweb wrote: post #40

The name "OpenAI" is a contraction since they don't seem "open" in any way. The only way I see "open" applying is "open for business."

What percentage of people who use their products care? 1%?

OpenAI is a brand, not a literal description of the company!

Re: OpenAI threatens to revoke o1 access for asking it about its chain of thought

#198

Post by Davidzheng »

rvnx wrote: post #191
tarruda wrote: post #185

Earlier quoted context omitted.

Seems like Reflection 70b was an attempt to implement the same concept on top of Llama 3 70b

Could even be that Reflection 70b got hyped, and they were like "wow we need to do something about that, maybe we can release the same if we quickly hack something"... Pushing an hypothetical (and likely false, but not impossible) conspiracy theory much further: in theory, they had access in their backend logs to the prompts that Reflection 70b were doing while calling GPT-4o (as it apparently was actually calling bo…

Reflection was hyped because strawberry was already leaked

Re: OpenAI threatens to revoke o1 access for asking it about its chain of thought

#200

Post by shinycode »

RivieraKid wrote: post #199

Some of the comments here are ridiculous. Preventing others from stealing your technology, data and ideas is a key priority for a company. Are people expecting OpenAI to give away their innovations for free?

Well it’s fair considering people gave their content for free (which is by the way the promise made by OpenAI in the beginning, to be open)
Post reply on HN