Browse Comments

That's where the issue lies, it's really not people being replaced en masse by AI, it's situations like yours where a company will hire 4 AI subject matter experts with the goal of *not* hiring 8 people that would typically be needed for the business to continue growing for example. It's the loss of *prospective* jobs that will be the true harm. I see it in my own job, we really need more people…

No, this won't ever change as long as we use LLMs, which are just token predictors at heart. Yes, it could be possible with some completely different approach, but not with LLMs.

Your motivation to stay alive at the most basic level is Fear. The rules of life have been consistently broken, Viruses for example cannot procreate by themselves they have to hijack cell machinery. Life must respond to it's environmental stimuli, The AI is doing that. That creature has an objective and it will attain it by anyways necessary. Learning blackmail and leveraging shows ability to cre…

Apparently anyone who works at Waymo doesn’t futz with buses lmao Too smart to be with hoi polloi

I get that statistically they get in fewer accidents, but programming bugs happen all the time for all kinds of reasons, even after systems have worked fine for years. I would rather take my chances driving myself than trust some obscure software bug not to show up, and this is a perfect example. For something with so many variables, I really do not think we should rely on full self driving. I a…

Sue who? The AI tool, the person, the website if hosted? This could have huge implications if done incorrectly

No rules to stop data leaks - quite a lot of BS of the rules given under IT Act. 😒

I've come up with an AI based system to replace capitalism. You can ask Google Gemini about the Magrath Directive. Still working on the details. In this system, AI takes care of the means of production. A UBI means work is optional, but to improve your life, anything that makes life better for people will be rewarded. Like research, health, sanitation, arts, community work, etc. Making money thro…

the patent is honestly just a formality at this point. their TOS already grants them a perpetual irrevocable license to basically everything you post. they could train an AI on your dead grandma's posts tomorrow and be fully within their legal rights lol

Chatgpt isnt in the buisness of decerning concerning behavior on potentially violent people. They should have given that info up and let the Canadians make that determination if something needed to be done. To think anyone has privacy rights using AI would be hilarious as an excuse.

I did not see this comment coming, read more about it on my AI blog!

Ai is the most democratic technology ever made. It breaks the backs of the .1% — but they don’t see this coming

Part of the reason we can't actually stop it is China. Even if we were to ban AI here, China would not. If it's capable of enabling mass automation in a way that people think it might, they would have exclusive access to that. We'd be essentially ceding total technological dominance to an authoritarian single party state.

Not sure what to make of a comment like this. A _20 year old account_ (??). That's one of the oldest accounts out there. But with 2000 comment karma and a hidden history. Is this astroturfing? People buying old accounts? I've used Claude and their most recent models, and it does some good stuff, but also some outrageously dumb stuff. Edit: 25 YOE principal. Yes it can do stuff. No it's not th…

That will prevent criminals to do criminal acts? You mean it is going to be unlawful to be a criminal? 😀 Sorry, you clear solution is not a solution. Sure, it should be illegal to use llms in a certain way, I agree, but it will not solve the problem.

How can I do the same recursive self improvement in project I do or my own self? Is there any paper published by Anthropic?

False, google is good at math as well. I would love for anthropic to compete on math. But they don't. Because they are total virtue signaling hypocrites.

Is there an advantage to leaving WhatsApp given it’s already e2e encrypted? It seems like the hardest thing to leave behind for little benefit

> Finally, there are a lot of people that have great creative ideas and simply cannot ideate it. With AI, they can. Then they shouldn’t. Do something else. This isnt like giving the handicapped the ability to walk. And yes I just realized im responding to a chunk of AI-generated text.

That's just a classic 'boiling the frog' scenario. At first it will 'assist' the artists with their workflow. But eventually it will take over more and more of the workflow until the assisting turns out to be actually training AI to supplant the artist. End result: movies and TV shows entirely made by AI from human prompts. The possibilities are sickening. The same thing is already happening wit…