Browse Comments

From a moral perspective, I don't see it as very different. AI isn't psychic. It's very good at guessing, detecting patterns, and replicating them, but fundamentally it cannot know what it has no way of having learned. It's not a picture of that person's nude body. It's simply a computer's guess as to what that person's nude body looks like. From a moral perspective, it's little different fr…

I used to be a human like you, then I took an AI to the knee.

And then said it should be illegal to use AI images in a presidential campaign.

We don't even understand or have a hard definition for what sentience is, so we can't realistically define whether or not something has it. That's specifically why things like the Turing test were invented, because while we can never truly define intelligence, we can create tests that should logically be equivalent. Of course, the Turing test is an intelligence test, not a sentience test - we d…

> I'm just baffled by Americans' sudden very deep and intense cultural urge to work in monotonous grueling jobs We don't _want_ to work these jobs, they're just the only ones left. White collar jobs are increasingly being gated off by insane requirements on top of an ubsurdly expensive undergraduate degree so the grunt work delivery/driving/warehouse jobs are increasingly the only option peop…

Always avoid sponsored links. Worse, the website link shown is usually masked and often points elsewhere. Google and the like don't care (long as they're being paid) and won't take responsibility. And legally, they don't have to, so they don't. The first non-sponsored link is no guarantee either. Often the first legit link for some websites will be somewhat down. Usually within the first 5 or so…

I think there’s also a big group of people who don’t understand how LLMs work and assign them greater intelligence than they actually have

That part has been interesting. Even in this thread, people suggest because AI can make basic mistakes, it is not genuinely intelligent. But imagine an advanced alien race evaluating us by the same metric. Not saying it doesn't have much, much more room to grow in the reliability or intelligence department, but if we're demanding perfection before acknowledging we have created true artificial int…

My buddy and I were just talking about this.  A lot of white collar work is organizing and formatting (data entry) or specific knowledge application (finance and law.) AI can do that stuff no problem. It really reminds me of excel in a lot of ways. Excel does so many things that 30 years ago you’d have to pay a person to do manually, but with formulas or macros you can do automatically. AI just…

I mean it sounds like it just took in very easy rote grunt work, with the only advantage over you or any other eng being time. However, these agents cost way more than a person and that's becoming more and more obvious by the day. I'm worried about how this affects juniors don't get me wrong but as someone that knows the types of tasks LLMs suck at we are all safe

I’m sorry, you were shook and are a person following developments in AI but you don’t know what model was used?

Claude 4.0 was down at work today and I had to downgrade to 3.7 Legitimately unusable by comparison, nothing it produced was worthwhile. It's like it lost all awareness of the codebase, it was incapable of understanding anything other than the file it was editing.

> The era of “let’s nerd out and build cool stuff together and have fun doing it” is over. It’s been over for 10 years now. Everything today is built solely to appease Wall Street. It's more just changed forms and is happening in media production now(teens making videos and games with elaborate production values now that the barrier of entry has been removed by phone apps and AI) instead of s…

its already happening. the actual onlyfans owners train ai to use their face and then create images without any effort.

Under 18's will figure out how to get around this, whereas many adults will struggle trying to figure out how to get their age verified!

I just want an LLM that doesn’t show up as green bubbles in group chat.

Would you say that it's somewhat akin to a school project that has gone on too far in one direction and that it's too late to turn back? What I mean is that given the goal is AGI, the way we have gone about doing it is this strict path of bits > bytes > transistors > code > Ai models and math, just layering on this very specific set of abstractions that we have discovered throughout h…

Do you mean startups who are selling AI or adopting it for a particular business problem? Typically startups adopt emerging technology and many of them fail. What’s crazy about GenAI is that massive regulated enterprises are also jumping on the bandwagon so fast. I remember when cloud was the hot technology. The early adopters were SaaS vendors or companies like Netflix. Capital One was the fi…

You're skipping the part where absolutely no other projects would ever get green-lit if the risk of failure and lack of ROI were as terrible as AI. Shareholders need to start filing lawsuits over this stuff.

It’s heartbreaking that a child felt safer opening up to a chatbot than to his own parents, and I feel for this family deeply. At the same time, I don’t think we can put all the blame on ChatGPT. AI is new and there should absolutely be stronger safeguards and laws, but parents also have to be involved and present in their child’s life. If a teenager is struggling with depression and feels unseen…