It certainly sounds like IBM wants people to fix broken shit code that AI barfs out. I'm sure that won't have any native long-term consequences or anything. /s

If they tell you to call them "pureblood" then inform them that means they are inbred and refer to them as being such.

This is really what matters. Some people know that Windows is a house of cards that needs AV to merely mitigate risk.

Just a reminder that this is a correlation study. It does not mean that ingesting caffeine will reduce your chance of dementia in the same way that the reduction of pirates has not caused climate change.

It could simply be that people who can afford to spend money on niceties like coffee can afford food and avoid homelessness. Homelessness is highly correlated with death and dementia but that doesn't mean the house itself is preventing dementia.

This does seem to explain why he retired in 2022 and possibly County Attorney Matt Schultz in the future.

Sounds like a surefire way to get pentesters to refuse you as a customer.

That's pretty much the core desire of every product, to become essential. I have personally found out that society assumes that you have a smartphone, and as someone that does not have a smartphone, that's annoying. However, for the most part there has been a way around these issues but sometimes it just means you can't use a shitty product (which saves me from dealing with a lot of shitty products). The only part of social media that has become mandatory (as I have experienced from not having social media accounts) seems to be LinkedIn if you are looking for a job in the STEM field.

By comparison, the only area where I see AI is becoming super dependent is search engines and customer service. Enshitification has made search engines rather useless (suspiciously after the release of their AI products) which is why you have to use an AI agent to find niche information that I could otherwise find. That said, customer service has always been trying to filter out as many absolute idiots as possible ("you have to plug the computer in, moron!"), so it's no surprise that AI checks that box. Companies with no human support are ones you shouldn't buy stuff from, but people are stupid and short-sighted, so it should be no surprise that they will leap to buy the latest crap without looking at anything beside it being shiny and new.

I'm far less worried about dependency than I am about the environmental impact of AI. Right now, the ROI of using AI doesn't seem to be panning out for most companies and unless that changes then the AI bubble will pop. If businesses set themselves up to fail without AI then they should all fail when they can't afford to use AI.

From TFA:

So yes, it's shutting down... to make another collider.

What is important is that scientists were able to use it to learn something new. Without any obvious new things we can learn using it, they are rebuilding it to be something from which we can learn new things.

I would say that Euria is a starting point, a place from which they are just beginning to break into presenting an alternative to American "Big Tech's" AI models. That said, there is lots of work to be done as simply using a CCP approved model is... not great.

This is 100% correct and it's also a problem with 100% of AI models. Even if it's completely unintentional, all AI models have biases. As a result of claims of under-representation and skewed perspectives, more research is going into detecting biases than ever before. However, that is not enough. What we need is for AI to be trained to present information in a neutral manner even when it's using a biased source of information, which is a difficult thing to do as it needs to have enough internal knowledge about every topic to recognize bias. It's a hard problem.

I wouldn't call it an addiction so much as a tool. However, just like a human friend, you can become overly reliant on it but the difference is that AI won't get annoyed with you and object to this behavior. There are some approaches to fixing this but as long as the interface is not designed to create a dependency then this is more of a user problem.

If you read the linked blog post in TFA, it's pretty clear that it was merely a matter of manpower and shouldn't be viewed as suspicious.

It says right in the summary, "each one was validated by either a member of Anthropic's team or an outside security researcher".

While you normally get less range from sodium-ion, it balances out the low-temperature discharge problem. As such, while you sacrifice peak range, you get a more stable stable year-round range in environments prone to very cold weather. If integrated correctly, it could also act as a bulkhead against thermal runaway destroying the entire battery pack. However, solid state lithium-ion batteries are on the horizon (2030-ish) so that's not a huge issue.

Regardless, sodium-ion battery technology needs to be developed further because we need all the battery chemistries we can get. Relying on any single battery chemistry is a way of creating an economic single point of failure which is ripe for exploitation.

All your posts are bizarre and absurdly long. Are you a bot? I could totally see AI taking the role of dipshit troll on slashdot. Like... who the fuck says "Internet 1.0"?