The ability to delegate tasks to an AI and relax as it reliably achieves them (or comes back to you for help if it cannot) is something that everyone wants from AI, and that marketing hype keeps suggesting that we have from AI, but that AI is nowhere near capable of. Not even close.

A significant part of the current AI bubble is driven by this extremely optimistic and outright false belief. People get really impressed by what AI can do, and it seems to them that it is equivalent or even harder than what they want it to do, so they convince themselves that it can.

But it can't. AI hallucination completely ruins this. You give it very clear instructions, and it will get 2/3 of it right, and also do something the exact way you told it not to. This gets even worse with implied behaviors like "and don't delete my entire hard drive while doing this."

AI can be helpful, but not in this way. Proper utilization of AI requires that you understand its limits and operate within them. It is outright reckless to give AI the authority to take action on your behalf (at all), and stupidly reckless to skip confirmations. Without you examining each command it generates, there is no force ensuring that it did it right, and it absolutely will do wrong things that should be simple for it to do right.

It's a catch-22, and always has been.

Dictatorships are tyrannical no matter how intelligent the leaders are. And given the power structure they have, there is absolutely no way to ensure that the dictator even cares about the people at all.

On the other hand, most people are idiots. They are extremely vulnerable to fake facts and other forms of manipulation, so their voting power isn't actually a form of political power held by the people so much as held by the people who manipulate the people.

What we actually have in America is an oligarchy that pretends to be a constitutional republic. Yes, we vote, but regardless of how many people participate the small group of rich people get everything they want. This is the inescapable result of the general stupidity of the majority of people (not to mention general laziness, apathy, and the very real and pressing need to spend their time earning a living instead of studying up and staying on top of politics).

So we get ruled by elites no matter what we do. The blow is softened a bit in a democracy due to regular rotation of the publicly-visible power-holders, but even then, most of the power is held by un-elected, un-appointed, rich people who only care about the country inasmuch as they have to in order to protect their own wealth.

It does not need to be intelligent in order to qualify as "artificial intelligence." In this context, the word "artificial" means "fake." Like "artificial leather" which is not actually leather, or "artificial crab meat" which is not actually crab meat.

The phrase has been around a long time in the domain of computer science and has always been used to mean "that which imitates intelligence (without actually being intelligent)."

You are not alone in your distaste for the word use here. But you are also greatly outnumbered. The English language doesn't have a final authority on what words mean beyond popular use. And in popular use, the phrase "artificial intelligence" is a very broad term that refers to a wide variety of ways that computers do things that seem intelligent (even though they actually aren't).

So, your pleas for people to stop using the phrase this way fall on deaf ears. That ship sailed long ago. This is what the phrase means, and will continue to mean, no matter how much you disapprove.

Yes, it was supposed to be a joke. Meta-humor, specifically. I was denying that the article applied to me while clearly exemplifying exactly what the article was talking about, thematically linked to a common attribute of the Slashdot user base (arrogance about one's own intelligence).

Oh well. There is a reason I don't work as a professional comedian.

I already know for a fact that I am smarter than most of the human population.

It's just nice to have a conversation with someone smart enough to recognize that. Even if it is an AI.

I thought that what Oracle was really good at was trapping customers into oppressive contracts and using legal action to extract money.

This question comes up quite a lot, in this context and in other contexts. We like it because it makes us feel like industry leadership is being irrational and short-sited, and boy will they get theirs!

In reality, this isn't as much of a conundrum as it seems. As it stands right now, AI can't actually eliminate entry level programmers, despite the marketing hype. It might reduce the raw number of entry level positions in the industry, but not down to zero. So, the problem doesn't actually exist yet.

But hypothetically speaking, even if we did completely eliminate the need for entry level programmers industry-wide, here is the most likely way that would shake out:

First, the jobs go poof, people are laid off, unemployment rises, career transitions start happening, college enrollment in compsci drops significantly etc. Everyone thinks these are pre-shocks of a coming earthquake, but they are not.

After experienced programmers start retiring and demand for experienced programmers is on the rise:

1. the experienced programmers will be paid a lot to come out of retirement, even on short term contracts. And many of them will do this out of sheer boredom, since retirees universally learn that retirement ain't all it's cracked up to be.
2. businesses start looking to hire people with alternative experience, mainly open source contributions.
3. businesses apply significant political pressure to get more H1B visas for these positions, and succeed.
4. businesses can start hiring inexperienced programmers into experienced roles, almost always pairing them up with experienced mentors. THere is a productivity and quality loss due to this, but it also winds up producing enough experienced developers to keep things going.

And done.

Economic conditions change all the time, and people just adapt to them. Usually any kind of gloom-and-doom economic prediction is based on the belief that one specific thing will change while everything else in the economy is held constant. This assumption is never true.

"Google AI Studio" has a free tier that allows API access to the latest Gemini models for code generation. There are no ads when you use them through an API. Of course, you have to code or otherwise obtain a front-end that can do that.

The free tier is only good for an individual developer, and the use restrictions might be a bit too tight for devoted professional code development work. But if you are using it for professional work, then you can spring for a paid tier anyway.

I expect several consequences of this, including:

1. Model collapse. Training LLMs on the output of other LLMs has been shown to lower the model's quality, and it gets worse with each iteration. So, the Internet has become less valuable as an LLM training data source, and this trend will continue, making it more difficult to train new models or improve existing ones.

2. Increased demand for guaranteed human-generated content. This is both from competition between LLM training businesses who need original sources to use as training data, AND from humans who want or need something that is not hallucination-polluted slop.

3. Increased incidence of humans submitting LLM generated slop AS human-generated content. We have already seen this happening in every place you might expect, with comical effect from catching people red-handed lying about this.

4. The bursting of the LLM bubble. Recently experts in the field have said that current training methods have already hit "peak AI" even from good data sources. The landscape continues to change rapidly so I don't know if that is true, but it is at least possible given what is known. An overall decrease in the availability of high quality training data will only make this worse. Then ensuing stagnation in LLM improvement will flatten out the demand curve for LLM services in general.

5. Profit! Especially for everyone who managed to eliminate a lot of human employees thanks to LLMs.

And no repayment to all the renters who have been gouged for years due to this price fixing, either.

Just download an alternative from the official repos.

Oh, wait, I forgot we weren't talking about Linux here...

I disagree. Win 10 was awful. The licensing terms were awful. The in-os adverts were awful. The dark patterns in config settings that tried to prevent you from being able to turn things off were awful. The mandatory online account registration was awful. That's just what I remember off the top of my head.

Win 7 was the last version of windows I ever used outisde of work, and I didn't even like it very much. Vista and 8 were dead on arrival so when Win 7 went out of support and Win 10 was the only viable alternative, I divorced windows once and for all. I have been exclusively using Linux (and on one machine MacOS) ever since, and have never looked back.

I have to use win 11 at work. And I find it awful.

It's almost as if making long-term commitments while under the influence of short-term feelings is fundamentally unwise.

I wonder if this has anything to do with the high divorce rate in the developed world....

LLMs are not good at self-management or judgement-call making. Allowing them to be "agents" and do things on your behalf is problematic because they can get things wrong and then make things worse when they try to fix it. They are much worse about this than human agents.

In my experience so far LLMs can generate code that "looks right", but doesn't necessarily work right. The more details there are in the requirements, the worse the LLM does. And in my experience implementing business workflow pipelines using LLMs, the LLMs are pretty good at interpreting plain English requests and translating them to something machine-parseable (like JSON or whatever), so you can then write your own code that reliably takes action, using the LLM just as a bridge between the two. But the more you ask the LLM to solve problems itself, make decisions itself, or take actions itself, the more it lets you down.

So, I think that AI just isn't ready for what Microsoft plans to use it for. And it seems like many others agree.

But it looks like slashdot has a new annoying add that moves up from the bottom of the screen to get your attention and obscures content until you interact with it. I see no option to disable this.

In short: I hate it.