4 ms·

Something about them caring about this bothers me a bit. Like, yeah ok it is a good thing that they are trying to curtail this abuse. However, it kinda feels

Something about them caring about this bothers me a bit. Like, yeah ok it is a good thing that they are trying to curtail this abuse. However, it kinda feels like we are just delaying the inevitable? With open source models like LLAMA, detecting and doing anything about this will become impossible in a year or 2 (in reality its already possible). It just feels very, "look at us doing this good thing but ignore what we unleashed on the world" so please pat us on the back here. Maybe I am just pessimistic, but I am not about to give OpenAI any praise here for this when they are the reason we are in the current situation.

UberFly 2y ago
My feelings exactly. As usual it feels like lawyers are saying "just do this even though it's BS so in a future court room you can bring it up". You're not pessimistic, just realistic.
ADeerAppeared 2y ago
> It just feels very, "look at us doing this good thing but ignore what we unleashed on the world" so please pat us on the back here. It is that. These AI firms have been very vocal about their "ends justify the means" attitude. (Which is doubly damning considering how little these firms care about any purported AI safety for their AGI dreams. Months of "AI is going to kill us all" while they couldn't even be arsed to stop using CSAM-contaminated training data, nevermind using quality training data to ensure alignment) And "Foreign state actors generated some propaganda texts" is very small as AI-problems in the here-and-now go.
Lockal 2y ago
You are saying as if automatic detection of misinformation and brigades was ever possible before LLMs, or it is possible now. Only now, when people started to massively notice things like https://news.ycombinator.com/item?id=40444872 https://news.ycombinator.com/item?id=40444872 OpenAI publicly executed few accounts. By the way, they didn't even say exactly how many. Hundreds new will be created in 3.. 2.. 1..

UberFly 2y ago

My feelings exactly. As usual it feels like lawyers are saying "just do this even though it's BS so in a future court room you can bring it up". You're not pessimistic, just realistic.

ADeerAppeared 2y ago

> It just feels very, "look at us doing this good thing but ignore what we unleashed on the world" so please pat us on the back here. It is that. These AI firms have been very vocal about their "ends justify the means" attitude. (Which is doubly damning considering how little these firms care about any purported AI safety for their AGI dreams. Months of "AI is going to kill us all" while they couldn't even be arsed to stop using CSAM-contaminated training data, nevermind using quality training data to ensure alignment) And "Foreign state actors generated some propaganda texts" is very small as AI-problems in the here-and-now go.

Lockal 2y ago

You are saying as if automatic detection of misinformation and brigades was ever possible before LLMs, or it is possible now. Only now, when people started to massively notice things like https://news.ycombinator.com/item?id=40444872 https://news.ycombinator.com/item?id=40444872 OpenAI publicly executed few accounts. By the way, they didn't even say exactly how many. Hundreds new will be created in 3.. 2.. 1..