Making no mistakes is a much higher standard than that which we hold to ourselves. Why are people moving the goalposts of intelligence or usefulness behind perfection?
Technology up to the dawn of the AI slop era was indeed expected to be perfect. When it wasn’t, we fixed it so it would be.
Why should AI be exempt from this? Techbros have convinced you that it should be so that their favourite lines go up.
There’s literally nothing more to it. A hammer is useless if it only drives 50% of the nails you hit with it. Why the fuck should we expect anything less than triple or quad 9 accuracy from AI if its so god damned “intelligent”?
Bc when I use a calculator, I actually DO expect literal perfection. And when I use google search, I expect it to be “useful”. And when I find information in Wikipedia, I expect it to be somewhat authoritative, even if incomplete. And if I use automative driving features, I expect them not to completely take over the wheel and crash me into a brick wall… or to a little child in a crosswalk right in front of me.
People who drive drunk lose their privileges to drive anymore. Employees who screw up that often get fired. Doctors who dispense incorrect medical advice lose their ability to practice medicine, plus get exposed to lawsuits. Counselors who tell their patients to kill themselves… Anyway, people DO experience the consequences of their actions, like ALL THE FUCKING TIME.
Whereas in contrast, AI is said that it is “going to be” great, not that it is great now. Fine, finish it and then we’ll talk. In the meantime, stop shoving it in front of my face.
If AI is like a human, it’s at best 2-year-old and at worst more like 6 months. It should not be “in charge”, e.g. of dispensing medical advice. But since it takes so much time to check its results for errors, it is literally slower and more painful to use it than to not use it (sometimes, often in fact).
You have a point somewhere buried in your mind, as revealed by the insightful first sentence, but your phrasing in the second sentence reads like sea-lioning and is not helping. Nobody is asking for “behind perfection” as that is literally mathematically impossible, and that is not what “moving the goalposts” means. It should not be enough to sound intelligent - we need to actually be such (same for AI as well).
And Google search has been spotty since the beginning.
And Wikipedia article quality … varies.
Like people, if you give AI a sufficiently complex problem, it won’t get it 100% right on the first pass. But, if you give it enough detail to distinguish an acceptable solution from an unacceptable one, it might get 80% of what you’re looking for on the first pass, boost that to 96% on the 2nd pass, 99% on the 3rd pass, and eventually what’s left is simple enough that it finally does get it 100% right.
Anybody who accepts the first thing AI tells them with today’s tech, is using it wrong.
Your “if” there is doing an awfully lot of the heavy lifting. Fwiw, I’m not talking special-purpose, custom-built LLMs - a large part of the problem is the lack of precision language uses to describe the concepts under discussion.
Both of these would be better called “cheating” than “AI”, but seeing as how AI both makes it easier and more to the point so many companies (such as Oracle) are literally pushing their programmers (those remaining anyway) to exclusively write programs using AI rather than by themselves, the very definition of “cheating” will need to be reexamined as a result.
In the examples also take note of how poor quality the LLM output is - e.g. regardless of whether the source is Grok or Claude or whatever, those therapy examples are not helpful in the slightest. Your counterargument might be that these are the “cheap” (aka free) AIs, but preemptively I will say in response: they still count as “AI”, especially in the context of the OP.
As far as “cheating” goes, ever since I got out of the game of paying a bunch of academics to judge and label me, I have been actively encouraged to “cheat” by the people who pay me money… that’s real life.
If you’re using a Ginsu knife to knead dough, you might not have optimal results. Claude is pretty good at code, since about 4-6 months ago. Grok? last time I asked Grok for anything it was the fastest LLM on the market, and the most non-sensical - usless trash.
Okay but Grok is still surely part of the “Anxiety around AI is growing rapidly in the US, research shows” phenomena, as Grok is one of the various AIs that people are aware of, and anxious about.
Your words read to me like you have kept yourself aware of the positive benefits of using AI - which many people on Lemmy including to some degree myself - have done far less of.
For the past year+ it has been popular sport to ask AI a question and poke fun at how wrong the answer is. I, too, get plenty of wrong answers from it - and anyone who trusts what it, or a Google search, or some post by some random troll with an axe to grind on some social media site, or even your high school whatever teacher, without verifying the results… gets what they deserve, in my opinion.
What changed for me within the last 12-16 months is: at least around questions in software development, the answers started being correct more than half the time. That was a critical watershed, because in essence that means that if you give your AI the tool to test its own work, it can work on hard problems that have easy methods to test for correctness (starting with compiler errors), and basically chip away at them - fixing problems until it has an answer that is correct enough to pass all the tests you have specified for it. Before that, an AI agent left to work on problems without guidance would more often get stuck in loops, or run off the rails altogether and never reach a viable solution.
In the past 6 months or so, tools like Claude have gotten much better - incorporating a lot of the kinds of things I (and many others) had to “tell them” manually 12 months ago to get good results into their normal response algorithms, anticipating and fixing problems in their work before presenting it as a solution for your consideration.
The language they present solutions in has been traditionally too over-confident, that’s a huge fault which I attribute to being trained on blog posts by know-it-all blowhard people who similarly present their ideas as gospel truth rather than their potentially flawed best efforts.
Clue for the clueless: even the best human experts in their fields are still only providing potentially flawed best effort answers. Once you leave self-defined fields like mathematics, all we have are our best guesses about how things really work.
One thing that your comments touch on here is just how little of the “Anxiety around AI” actually has to do with AI.
When e.g. Oracle lays off 30k workers, how little of that truly has to do with AI? vs. instead market instability etc. What complicates the issue is that most often, the corporation will claim that the layoffs are to better streamline the company in a future where AI will need fewer workers, so to prepare for that now… they’ll just go ahead and get rid of them immediately.
So this isn’t even people using AI inappropriately, this is people blaming AI for what they wanted to do anyway, for reasons if profit.
Then again, events such as those presage what is to come: when AI truly can do it all, how will humans be able to earn a paycheck? Spoiler alert: not all of us will. And especially in the meantime there will be period of transition and upheaval.
This is what I felt your comments lacked acknowledgement of: not the downside to using the tools but the wider conversation that uses the keyword “AI” but has really barely anything to do with it, as opposed to political and social and economic forces.
Narrator: actually, no it was not.
e.g. it still spreads misinformation.
Making no mistakes is a much higher standard than that which we hold to ourselves. Why are people moving the goalposts of intelligence or usefulness behind perfection?
Technology up to the dawn of the AI slop era was indeed expected to be perfect. When it wasn’t, we fixed it so it would be.
Why should AI be exempt from this? Techbros have convinced you that it should be so that their favourite lines go up.
There’s literally nothing more to it. A hammer is useless if it only drives 50% of the nails you hit with it. Why the fuck should we expect anything less than triple or quad 9 accuracy from AI if its so god damned “intelligent”?
B-b-be-be-because shut up you, that’s why!
Won’t someone think of the poor shareholders?
(/s)
Bc when I use a calculator, I actually DO expect literal perfection. And when I use google search, I expect it to be “useful”. And when I find information in Wikipedia, I expect it to be somewhat authoritative, even if incomplete. And if I use automative driving features, I expect them not to completely take over the wheel and crash me into a brick wall… or to a little child in a crosswalk right in front of me.
People who drive drunk lose their privileges to drive anymore. Employees who screw up that often get fired. Doctors who dispense incorrect medical advice lose their ability to practice medicine, plus get exposed to lawsuits. Counselors who tell their patients to kill themselves… Anyway, people DO experience the consequences of their actions, like ALL THE FUCKING TIME.
Whereas in contrast, AI is said that it is “going to be” great, not that it is great now. Fine, finish it and then we’ll talk. In the meantime, stop shoving it in front of my face.
If AI is like a human, it’s at best 2-year-old and at worst more like 6 months. It should not be “in charge”, e.g. of dispensing medical advice. But since it takes so much time to check its results for errors, it is literally slower and more painful to use it than to not use it (sometimes, often in fact).
You have a point somewhere buried in your mind, as revealed by the insightful first sentence, but your phrasing in the second sentence reads like sea-lioning and is not helping. Nobody is asking for “behind perfection” as that is literally mathematically impossible, and that is not what “moving the goalposts” means. It should not be enough to sound intelligent - we need to actually be such (same for AI as well).
And you have calulators.
And Google search has been spotty since the beginning.
And Wikipedia article quality … varies.
Like people, if you give AI a sufficiently complex problem, it won’t get it 100% right on the first pass. But, if you give it enough detail to distinguish an acceptable solution from an unacceptable one, it might get 80% of what you’re looking for on the first pass, boost that to 96% on the 2nd pass, 99% on the 3rd pass, and eventually what’s left is simple enough that it finally does get it 100% right.
Anybody who accepts the first thing AI tells them with today’s tech, is using it wrong.
Your “if” there is doing an awfully lot of the heavy lifting. Fwiw, I’m not talking special-purpose, custom-built LLMs - a large part of the problem is the lack of precision language uses to describe the concepts under discussion.
An example: https://lemmy.world/post/46390157
Another example: https://discuss.tchncs.de/post/59584533
Both of these would be better called “cheating” than “AI”, but seeing as how AI both makes it easier and more to the point so many companies (such as Oracle) are literally pushing their programmers (those remaining anyway) to exclusively write programs using AI rather than by themselves, the very definition of “cheating” will need to be reexamined as a result.
In the examples also take note of how poor quality the LLM output is - e.g. regardless of whether the source is Grok or Claude or whatever, those therapy examples are not helpful in the slightest. Your counterargument might be that these are the “cheap” (aka free) AIs, but preemptively I will say in response: they still count as “AI”, especially in the context of the OP.
As far as “cheating” goes, ever since I got out of the game of paying a bunch of academics to judge and label me, I have been actively encouraged to “cheat” by the people who pay me money… that’s real life.
If you’re using a Ginsu knife to knead dough, you might not have optimal results. Claude is pretty good at code, since about 4-6 months ago. Grok? last time I asked Grok for anything it was the fastest LLM on the market, and the most non-sensical - usless trash.
(I did not downvote you btw)
Okay but Grok is still surely part of the “Anxiety around AI is growing rapidly in the US, research shows” phenomena, as Grok is one of the various AIs that people are aware of, and anxious about.
Your words read to me like you have kept yourself aware of the positive benefits of using AI - which many people on Lemmy including to some degree myself - have done far less of.
But there are some negatives as well…
There’s plenty of negatives to any new tech, anything can be carelessly or ignorantly mis-applied.
The computer has been coming for our jobs since it was created. Bob Cratchit no longer works for Ebeneezer Scrooge, he’s been replaced with software.
People over-trusting software has been problematic since software became accessible to be over-trusted. A favorite (horrible) example from not-so-long ago, but pre-ChatGPT release I believe: https://www.amnesty.org/en/latest/news/2021/10/xenophobic-machines-dutch-child-benefit-scandal/
For the past year+ it has been popular sport to ask AI a question and poke fun at how wrong the answer is. I, too, get plenty of wrong answers from it - and anyone who trusts what it, or a Google search, or some post by some random troll with an axe to grind on some social media site, or even your high school whatever teacher, without verifying the results… gets what they deserve, in my opinion.
What changed for me within the last 12-16 months is: at least around questions in software development, the answers started being correct more than half the time. That was a critical watershed, because in essence that means that if you give your AI the tool to test its own work, it can work on hard problems that have easy methods to test for correctness (starting with compiler errors), and basically chip away at them - fixing problems until it has an answer that is correct enough to pass all the tests you have specified for it. Before that, an AI agent left to work on problems without guidance would more often get stuck in loops, or run off the rails altogether and never reach a viable solution.
In the past 6 months or so, tools like Claude have gotten much better - incorporating a lot of the kinds of things I (and many others) had to “tell them” manually 12 months ago to get good results into their normal response algorithms, anticipating and fixing problems in their work before presenting it as a solution for your consideration.
The language they present solutions in has been traditionally too over-confident, that’s a huge fault which I attribute to being trained on blog posts by know-it-all blowhard people who similarly present their ideas as gospel truth rather than their potentially flawed best efforts.
Clue for the clueless: even the best human experts in their fields are still only providing potentially flawed best effort answers. Once you leave self-defined fields like mathematics, all we have are our best guesses about how things really work.
One thing that your comments touch on here is just how little of the “Anxiety around AI” actually has to do with AI.
When e.g. Oracle lays off 30k workers, how little of that truly has to do with AI? vs. instead market instability etc. What complicates the issue is that most often, the corporation will claim that the layoffs are to better streamline the company in a future where AI will need fewer workers, so to prepare for that now… they’ll just go ahead and get rid of them immediately.
So this isn’t even people using AI inappropriately, this is people blaming AI for what they wanted to do anyway, for reasons if profit.
Then again, events such as those presage what is to come: when AI truly can do it all, how will humans be able to earn a paycheck? Spoiler alert: not all of us will. And especially in the meantime there will be period of transition and upheaval.
This is what I felt your comments lacked acknowledgement of: not the downside to using the tools but the wider conversation that uses the keyword “AI” but has really barely anything to do with it, as opposed to political and social and economic forces.