ekelsen

It could have been an interesting article

If they believe open weight models are inherently unsafe but are forced to release weights, then if they really believe their own safety arguments, they would need to stop development.

The hope was that it would be a slight rule tweak that would lead to fewer draws without favoring white too strongly.

I did and it's true, but I amended my statement at the top of this thread, which you've ignored.

What I wrote is true (it is a claim about my experience), so I'm not sure what you're attacking.

I amended my statement to avoid a discussion about the provenance which is not what I actually care about. What I care about is if I want to read it. Humans could theoretically write something in AI voice and if they did, I wouldn't want to read it.

I have not seen any examples of human text in LLM style that clearly predate 2023. If you can supply one, that would be incredible evidence.

If you're able to supply a prompt which generates text that fools pangram and is not like one sentence long or a list of names or something, but actual prose I want to read, then I will concede.

I said it wasn't something I wanted to read. Humans are capable of writing things I don't want to read. Which at this point is starting to include your comments.

What I said is that every time Pangram has said 100% AI, it is 100% something I don't want to read. I don't particularly care if it was written by humans or LLMs.

Every time I've asked for evidence from someone who claims something like you do there always seems to be a reason it can't be posted. So it makes one suspicious.

Is there no way you can modify your prompt to be shareable?

So show us the prompt

The real reason being that it doesn't exist?

I don't know if putting a disclaimer at the very end of the blog, that you also disclaim, should be considered "very clear".

"Here is, for those that have read this far, the acknowledgements that used to be in the formal paper but we decided to remove it, mostly because the paper has been reworked and rewritten so many times by us."

I never found a piece of text that pangram claims is 100% AI that to me seems like purely human written text.

The only counter example is when a human composed a piece deliberately trying to mimic AI slop.

So call it an AI slop detector if you want, but either way it accurately identifies stuff I don't wan't to read.

Feel free to post the evidence yourself rather than directing me to a gigantic pile which you claim mostly does not even support your point.

Do you have any published dateable text from pre 2023? I'm super curious if you always wrote like this, because it is exactly in the style of AI slop.

People say it's in the training data, but I haven't seen any good examples of clearly dated pre 2023 text that sounds like this.

The link says [dead]

Can you share your human piece that got flagged? And the AI piece that didn't? (Or snippets?) I've found pangram to be very accurate, so I'm curious.

Yes, my point, if the answer is that there is no way this works without every disc being different.

It was not possible to manufacture at the time at scale.