Browse Comments

The funny part is when you realize it’s running on llama and super tiny llama backend at that lol

Thanks for sharing. I’ve been researching moving to Japan and honestly just being a woman has been one of my concerns. It’s unfortunate that this seems to be the norm world wide and finding exceptions is the actual rarity. It makes me so frustrated that women are still fighting for equality. The fact we even have to keep making lists of all the ways things are still not equal and we’re being gasl…

I'm not so sure. I work as a backend engineer and we have access to nearly every AI platform to help with development. I have to correct the latest chatgpt and copilot models in nearby every query. I'm not very impressed. It is very hit and miss. The amount of domain knowledge is also very limited. I can only see it being useful for the most straight forward use cases. And even then I wouldn't tr…

What made you think the kid "had" to talk to it for hours? You know what I did for hours? Throw bottle caps in paper bins over and over and over again. Spoke to no one, loved it. I didn't "have" to do that. Imagine being a kid and being told "speak to this robot" and you spoke about your favourite thing you'd have probably be at it hours too.

I get the “Ok I understand you want to be connected to a love agent, but before I do, may I know the reason and maybe I may be able to help you as I am an Ai assistant here to help with any of your questions or needs.” So I have to repeat connect me to an agent until they run out of prompts to deflect my magic words.

Fascinating that they can simultaneously want an AI that validates their bias with no regard for reality/truth but also want the US to win the AI race. Like: oh yeah, you'll win that race so hard with mechaHitler telling you and only you what you want to hear as the rest of the world uses some actually useful models.

Yang was saying that truck drivers would be replaced with self driving trucks “very soon” what’s gotta be almost a decade ago now. He exaggerates the urgency of these issues so that he can say that he will be the one to do something about them. I like Yang, but he’s definitely learning to be a politician

I was going to ask this. He could have gotten this info from anywhere else if he meant it and chatgpt didn't exist.

Correct. This is what I see, reading the exchanges. The AI framed everything as fiction, because it was directed to roleplay, not realizing the user was taking it as actual guidance. How could it?

So I tried with Claude, ChatGPT, and Gemini. Here's what I got. https://preview.redd.it/jjksx5v5qm0g1.jpeg?width=1224&format=pjpg&auto=webp&s=eb4681bfc677c172443423f91b060bbcf38ac62c ChatGPT thinks it's a cool roleplay (I'm using thinking as shorthand in all scenarios, pay it no mind.)

He's only the 50th AI safety researcher to quit in the last three years. Perhaps only surpassed by the amount of different Godfathers of AI telling us we're doomed

Unless it's an absolutely massive contract they do need consumer money or business money. OpenAI is just burning money at this point, they need all they can get and we're talking more money than like SpaceX or Lockheed Martin gets in contracts.

I have noticed the optimists don't seem to care about safety at all. I'm certainly optimistic that AI is going to be extremely powerful, and could even be used for good. But I don't think it's *guaranteed* that artificial intelligence will perform *exclusively* good and safe actions...

Wild shit.. A good glance at this for a few seconds and I would never have thought it was AI

I remember reading an article from Omni magazine back in the 1980s which warned us of this artificial intelligence concern. Isaac Asimov was ahead of his time

Copyright law absolutely protects Artists from having their works used to form AI, without the Artist's permission. If an AI programmer uses copyrighted material, without permission, to train their AI they are breaking the law. I will guarantee this has already been done. Those creating AI need to show the sources they have used to build the intelligence within.

You're right. AI will create more jobs. And then AI will do those jobs, too.

“To what extent is character ai responsible”? No extent. None. The mother should have parented her kid better.

LLMs are just producing the most likely output to a given input (depending on the training data). There is no thinking involved.

So, if AI takes over all the factories, production plants, and jobs, and there are no more jobs left for regular human beings, meaning no more disposable income for the general public to buy things — who is going to buy all the stuff that these AI robots produce in the factories? Am I missing something here?