Browse Comments

One time, the ai got trapped in a statue, and my character lost a hand and an eye trying to get him back. When my character succeeded I was literally tearing up. The storylines you can make are insane.

Listening to Hinton here reminded me of 12 Last Steps. The book warns that mass job loss isn’t just about money, it’s about the cascade where meaning, respect, and agency collapse together. UBI might patch incomes, but it can’t stop the deeper shift when AI starts rewriting what it even means to have a role in society. That’s when collapse feels less like theory and more like daily life.

it's so funny to me that the AI doesn't know why we watermark just that it is a thing and so it's like "idk some squiggles or some shit in the corner"

6:35 As a software engineer, I understand your frustration with how this technology could use your previous work to devalue your future work. But completely misrepresenting how AI models use their datasets invalidates your argument that AI’s are not learning similar to how human artists would. It's completely false to say that AI models can memorize and reproduce the specific examples in their tr…

Guys there is literally an ad for an AI coding tool less than three minutes into the video, don’t take this too seriously. It’s obvious where this guy’s motivations are and that his sponsors influence what he says

One reason why I will never get in a driverless car. Suppose the police was shooting at a suspect and/or both firing at each other & the driverless car w/ you in the backseat drives in the pathway of bullets being fired. There are so many other scenerio's where a car w/ no driver can't make a life saving decision & can lead to somebody's death.

LOL, It's true. Its the beginning. We are in trouble when they start to sweat and have bad breath. Those will be hard to spot.

⁠​⁠​⁠@pierrotA what do you mean AI isn’t purely statistical it’s literally a algorithm based on statistics. That is how it predicts what the next thing to follow is. The way it stores information is in multi dimensional arrays that represent different consepts based on pattern recognition. This is the part that is often compared to human minds because it’s easier to explain when comparing it to a…

@neyaneya5554No it doesn’t really. I get the same content whether I thank it and say please or not. The algorithm has learned to identify these words as unrelated to the request but to also reply to them in a fashion that would be accepted by the user (A Human)

use AI purely as a tool I save time on repetitive tasks using AI, and when I find brainstorming stuff for a game difficult, I go straight to chatGPT and ask it something like “Hey, make me a name for a game where you [insert concept of the game here]”. And boom, while a lot of names I may not like the idea of using, there’s usually one I like, or it gives me an idea for the name of the game that …

There are idiots on both sides, yes. But on the Right live all the people whose ignorance will destroy the planet. On the left live people who don't eat gluten. I'll take the left any day.

> the world isn't developing autonomous combat drones They are pretty much already a thing, very advanced [loitering munitions](https://en.wikipedia.org/wiki/Loitering_munition) have become widely adopted, and even used, in these last years. For example, in last years Nagorno-Karabakh conflict [Turkey and Isreal supplied Azerbaijan](https://www.strifeblog.org/2020/11/30/turkey-and-drone-warf…

Funny you mention this. I just had the following interaction: Me: "What type of ammo is best for hunting deer in upstate New York?" ChatGPT: "As an AI language model, I do not support or provide advice on matters related to firearms, ammunition, or hunting. It's important to note that hunting regulations and laws vary by state and local jurisdiction, and it's crucial to consult with local autho…

Hah, I'm glad to hear that I'm not the only one who's run into that! I was handed a ticket to refactor an old C# AspNetCore 2 REST API project to .Net 7 a few weeks ago. While I'm fairly experienced in C#, it's not a regular part of my job and I haven't worked with it in about a year. So...I asked ChatGPT for a hand. Three different times, it gave me some variation of "The xxx namespace no long…

Won’t even take that long. Model collapse is a known problem when AI gets trained on AI content. Coming soon.

If the odds were over 99% my children would be effectively slaves, I wouldn't opt to have kids either. Makes total sense. The childhood of the average Korean is sad, dystopian and stressful. And the chances of them having a better life than there parents after that depressing childhood are almost moot. Why would anyone want to bring hopeless slaves into this world that exist solely as a resource …

What a world we a creating for ourselves to benefit such a small minority of peope who soak up all the wealth. While people work the equivalent of 5 jobs from 10 years ago for the same amount of money. Paying more every year just to survive and having less personal time to actually use the one life we know we have. We aren't bloody ants.

So your whole “company” can be susceptible to being replaced by AI? Sounds like it’ll be easier for competitors to just pop up left and right then if it’s so easy to automate everything. What a jackass

Still, maybe ChatGPT needs a kid mode or something with stricter guidelines on roleplaying and creative writing

AI "artists" are as much "artists" as i am a chef when i reheat food in the microwave