Browse Comments

GPT-4 is leagues beyond the competition, and the competition is never catching up. The proof is that OpenAI had GPT-4 months before they even released GPT-3.5.

It'll improve over time though. Then what do you think the solution should be as far as teaching goes? I imagine more in-class "homework". I've heard of other subjects requiring reading/watching the material as homework, instead of doing homework that involves using ChatGPT to get answers or do the work, that's instead replaced by in-class work unaided by computers/etc. But I'd imagine some te…

You have found a market inefficiency, now is your time to shine. I would pay for a service that provides what you just explained. I bet a current AI model could help put it together.

They talk about replacing us because they don't want to have to employ us. That's it. It's a bunch of middle managers thinking that they're qualified to work the line, wanting to increase their pay by reducing their own head counts, and thinking that they'll survive the round of layoffs because they're special and keep the operation moving. Also, based on how underwhelming ChatGPT5's improveme…

Can't help but think all these guys working on female robot faces are working towards the ultimate goal of a sex bot

She's the real deal. Master's degree from Harvard and a ton of other achievements. She's taken a lot of misogynistic attacks during her political journey and just kept going. I'm super happy for her

Even days – Canada: “Those Sikhs? Not truly Canadian.” India: “Cool, deport them. One of them blew up Air India Flight 182 and killed 329 people, mostly your citizens.” Odd days – Canada: “Khalistani separatists are 100% Canadian and we’ll defend their rights!” Pick one narrative, Canada.

On what timescale? Kim Jong Un is the equivalent and his family has sat in that throne for three generations. But history is a poor marker since there’s never been the potential for the kind of asymmetry between those who control a perfect robot brain and its many embodied form, and those who don’t

There's a difference between CGI/AI and video evidence, but with it being this powerful, it's a dangerous proposition for powerful people to fool others and the media. Lucky for us, the lack of TRUST, means we begin to question everything, which is a good thing in principle, tbh ! If it looks like a duck, quacks like a duck, moves like a duck : We still need to check 1000's of things to work ou…

This is the excellent point that most people don't understand. LLMs do not "think". They take the request prompt given to them and try to assemble an output response that they think matches the desires of the prompter. They don't even understand language; they are just doing pattern matching. Yet everyone has been hyped up to the point where they think LLMs are human replacements.

No kidding. AI movies aren’t replacing actors, directors, and producers. AI vibe coding isn’t replacing software engineers. AI isn’t the revolution we’re being sold. It costs way too much and produces very little quality output.

I was told that i had to write up every quarter the use cases that i had found that makes AI useful in my workflow. Motherfucker aren't you already supposed to know that before mandating its use?

Honestly think they're making search engines bad so people start using AI as their search engine. Then we're all fucked.

Article without paywall: https://archive.is/loQZu >Palantir’s leadership incensed workers yet again this week after the company posted a Saturday afternoon manifesto reducing Karp’s recent book, The Technological Republic, to 22 points. The post—which includes many of Karp’s long-standing beliefs on how Silicon Valley could better serve US national interests—goes as far as suggesting that th…

By the time we have full A.I. machines, the question isn't "should we give them rights?", the questions is, "will they give us rights?"

The most moral and pragmatic thing would be to treat any AI that you have reason to suspect has reached personhood as such by default.

This is hilarious! The robot spoke the truth. And what's wrong with launching anti-feminist attacks? Sounds like valid criticism. Too funny.

Why is nobody talking about psychotherapy, human connection, or creating retreats where people can truly experience being alive – real bodies, real energy, real presence? In AI debates about superintelligence, we forget that some of the most profound intelligence comes from living, breathing humans sharing space, emotions, and healing together.

As someone that is in school now for graphic design, who has taken years to even give myself the permission to take the risk and pursue my passion, this AI situation is horrifying and is very discouraging. I don't want to share my art on the internet at all anymore, but that's how artists make their money these days. I will say thank you for making this video and talking about this in such a calm…

As a software engineer who went through a research-focused school not that long ago, and who is now in the industry, I resonate with the anildash article more than Nate's take. And even so, neither of them talk about the apparent scaling limitations / diminishing returns we seem to be hitting with LLMs. Nate did sort of allude to the last AI winter being ended by an algorithmic change, but didn'…