Browse Comments

lol the exact same way the teachers punishment of hand writing future assignments somehow prevents people from doing exactly that. Depending on the type of assignment, you’d likely expect days/ weeks worth of progress and edits. Whereas I’m guessing most who cheat just copy and paste the whole thing in one fell swoop. So if you could show you were plugging away it every night for a month, it MAY …

They pop up sure but Reddit explicitly permanently banned and closed any subreddits about deepfakes well over a year or three ago.

You want me to do TWO THINGS!? And that’s how AI will begin to self replicate. So each iteration doesn’t have to do more than one thing.

No matter what they do Opposition bloc here is so incompetent, BJP will win atleast by borderline margin.

Yeah the AI hype is overblown. I work on AI systems and yeah it’s been nice funding wise but holy shit does business not understand anything about it.

It's not a loophole. The long-held precedent is that the First Amendment does not protect CSAM in part because criminal sexual abuse is performed in front of the camera to create it. AI trained on real CSAM images is a product of that abuse. The government also has a legitimate interest in suppressing realistic generated CSAM because it jams up law enforcement capacity to investigate real images.

Well, for all the people making AI nudes of me, I have killer abs and a really big pp so make sure you put that in the prompt.

OpenAI is a house of cards and their product and business development strategies are as pathetic as they can get. They thought they were immune from the risk of commoditization for their core offerings, and that they could recoup their investment over a long period of time from 5-10 years, and all they have to do for the time being is to hoard hardware, lobby the USG, and divert capital from comp…

Christ this thing is so good. And it can remember you for 2 weeks. Its with this thing that people will seriously become friends with AI and have AI girlfriends that are like Her the movie. Its mind-blowing how good this is

Similar experience, I use it mostly to help write and plan my dnd campaign and it’s been really bad lately. I used to prefer claude, and I may switch back to that.

I infodump about my special interests to ChatGPT all the time. Wish I had it as a kid when I had nobody to talk to lol

This is what I don't understand, and what all these doomsday articles fail to properly articulate. They say 'AI'. But what are we talking about? The term is a blanket statement. AI can't currently reason, and it is prone to hallucinations right now. It can't even count the number of letters in a word half the time. The concern is regarding promises of future AI technology that does not currentl…

Metadata, many contain the model, seed, resolution, prompts, etc within the image itself. This however can be bypassed by just exporting the image again. While it doesn't solve the issue at least the lazy AI spammers will stop showing on the search results in huge quantities.

Apple's Machine Learning Research came out and said this trend isn't even AI on no uncertain terms. LLMs are basically the continuation of Ask Jeeves, chatbots and 2010s virtual assistants. From a technical standpoint LLMs aren't even close to actual AI and like the above comment implied they're hitting a ceiling. The biggest issue is they cannot reason. [https://machinelearning.apple.com/resear…

Please, more reports like this. It matches my professional experience. The "AI" sucks. And it's consistently getting worse. This fad needs to die.

Step 1: Replace everyone with AI Step 2: No one has a job because they've been replace with AI Step 3: People cant buy anything because they don't have a job Step 4: ??? Step 5: Record profits

Once again they are using AI as air cover for oh crap we overspent on AI that isn’t delivering What they loose on reality labs VR, 50 billion? I’m still firmly in the camp AI agents won’t be replacing the work force and will just become another tool that companies will have to hire for

I think the concept is pretty simple, the same as for "self-driving". If you are rarely required to intervene, you will rarely intervene when required. So the technology at a baseline has to be better than a qualified human paying attention, or you will end up with more mistakes as the person responsible for it will inevitably get lazy and stop checking everything

And the other majority are from new accounts who have post history turned off and then you google their account to see they make thousands of these posts and despite being a "CS expert with 10 years in the field" none of there pre AI posts have anything to do with CS just a LOT of failed web3 content like nfts and crypto scams.

Yes. Jobs will appear. But we won’t be the ones doing them. Why for? AI can do it better.