Keep attacking the messenger while ignoring the message.

Just because he's annoying doesn't mean he's wrong in this case.

Exactly. Considering the level of idiocy just in this comments section (not you, in case you're wondering), it's obvious that most people aren't smart enough to effectively use AI. The results are only as good as the person directing it.

I've used AI for both personal and work related functions, and I find the best way to do it is to focus on one idea, and always start a new conversation for new questions. Don't ask philosophical questions and expect serious answers, and remember that the further you get from the original question, the more unreliable the AI bot is going to be.

I was installing FreeBSD from two floppies and a network connection. 30 floppies is tragic.

And a snide, sanctimonious one who makes assumptions about others knowledge.

I know EXACTLY what I'm talking about, having done a fair amount of experimentation in the last 6 months. You obviously tried AI once, decided it was terrible, and closed your mind to the possibility that it would improve.

"Garbage in, garbage out" has never been more true than with the major LLM's available. Always start a new conversation, and limit that conversation to one topic. And yeah, they still occasionally make up things, but if you call them out on it, they'll usually correct themselves.

Too bad you apparently can't do the same.

Assuming there's another general election.

Meanwhile, he's dismantling every single governmental safety net in this country-- Everything from climate science, medical science, research funding, PBS, NPR, FEMA, NOAA, NWS, NHC, Dept. of Education-- the goal of Project 2025 is eliminate elections, critical thinking, the lower class, and establish a Christian Nationalist theocracy.

Nearly all of Trump's current advisors in the White House helped write Project 2025, and they're fighting like mad to implement it.

Same as it ever was? Don't be stupid. The Constitution and the Republic as a democracy are under attack.

Passed an IQ test at Walter Reed, did you?

Still cracks me up that Trump claims he scored hugely on a cognitive deficiency test that's designed to determine what level, if any, of dementia you have.

What's the saying? Every accusation by the right is a confession?

So here's the thing about Pizzagate. When asked about his relationship with Epstein, Ghislane, and a specific date when he was alleged to have had sexual relations with a young woman, Prince Andrew said "Oh, no, I couldn't have been there, I was at a Pizza Express in Woking with my daughter".

Let me be rude:

What if it was your child?

It's a poor workman that blames his tools.

Learn how to use the tool, get actual results.

Yeah. Let's ban public and personal transport, the internet, and electricity. And medicine, and industry. Let's go back to the stone age, where you'd already be dead of old age, disease or injury.

What a fantastically luddite post.

How about we actually EDUCATE our children, instead?

But unlike me, who had a week of a class, either civics or business, that was devoted to "how to make change", I would suspect that 20 year old was never taught to make change.

I *can* do complicated math in my head, but I'm out of practice, because it's more reliable to use a calculator. Less chance of making a mistake. I'd rather do it on paper, but ideally, I'll hit ctrl-l, type "=" and hit enter, because it's fast.

I would suggest that ChatGPT (or others) would do a helpful job of letting you know of code requirements, issues you might need to be aware of when selling your house, current gun laws, etc..

Primarily, I use AI as a meta search engine-- I don't have to skip through "use this one weird trick", or badly generated AI pages that bury the useful information three pages down, or 27 forum posts from 2009.

I've used it for medical research-- in my case, primarily on P450 metabolism of various medications. I've done such searches before, manually, but it's very tedious. "What's an effective medication to replace metoprolol that isn't reliant on P450 metabolism?" was a question that quite literally changed the quality of life for me from "terrible" to "good".

But, and this is a big one, the quality of the results depends largely on the quality of the input. If the question had been "what medication should I take for my heart?", then the result would have been nearly useless.

You'd have to have a subscription to get a useful result, but the pgsql DB and java interface-- it might actually do that, or at least make a good start on the framework(s).

It's a tool. The better you know how to use it, the better the results.

"years ago"-- yeah, there's your problem.

The AI technology is changing so quickly that a model from six months ago is nowhere near the current models.

There are uses for LLM's in coding, but I wouldn't trust one for an embedded system either. They can be useful when used appropriately, though.

All of these were topics 40(!!!) years ago when I was in college. Difference is, we actually talked about it without threatening to kill anyone.

Education in college is only about 20% about class work. The rest is how to exist in society, how to meet people without lighting crosses in their yards, and taking a crowbar to that cranium your family welded shut during your formative years. How to navigate a bureaucracy, how to experience new things.

If all you want is to stay in your sheltered little bubble of disreality and only learn things that won't offend your precious medieval sensibilities, then just stay on Twitter.

... that I ditched Cox back in January. They aren't the worst company, they aren't the best, but I always knew where I stood with them.

Now I have a fiber provider that's actually local (until they get bought out) who actually believes in customer service. How bizarre.