AI for analysis (neural networks or machine learning) has been used for years and it's a good tool to validate or confirm some data. It's a useful tool meant for humans.

Chatbots and waifu generators are only used by billionaires to brainwash people who are already brainwashed by TV and social media, and it's very bad.

Also LLMs hallucinate and choke on their own vomit. Who can accept such an unreliable application?

Not at all. But it is almost exclusively abused. It literally makes people dumber if they use it to learn something.

Some examples:

1. Dictation and summaries. AI "note-takers" are being built into videoconferencing platforms and this is super helpful, not only for later reference in a search, but also for a supervisor to just get a quick glimpse at what's going on with employees.
2. Searching a private database. Google Drive search it ironically horrendous. Like if you type in the exact name of a folder, it will show you a bunch of other shit and not that folder. Meanwhile I can ask Gemini a very specific question that may be in a single cell of some spreadsheet somewhere and it will not only give me the answer but also link to the associated spreadsheet and even the specific cell where it found the answer.

It just doesn't have nearly as many applications as the techbros would dupe you I to believing. And it's producing harm on a massive scale, in so many in different ways.

Nope. Machine learning has been a legit comp sci field for decades. Tons of really cool applications in medicine, for example.

For me, the current "AI" frenzy has a couple core problems. Avoid those, and you're good.

• LLM chat bots being treated as intelligent, when they fundamentally can NEVER be intelligent. (I.E. researchers have determine that there IS no solution to hallucinations, without starting over from scratch with a different modeling strategy).
• Training machine models on public, or even copyrighted, works, and then thinking its okay to use that for personal profit. In particular, the stark inequality in how copyright is enforced for normal people, but ignored for private "AI" enterprises.
• The capacity of "AI" tools to cripple human potential, rather than reinforce or elevate it, and the fact that so many people confuse the former for the latter. Like, I can genuinely relate to the idea of having artistic ideas in your head and not having the skills to bring them into reality, and seeing image generators as just a way to fill in the gap. Like, there IS a promotion of creativity there, but being able to run prompts through an image generator is NOT the same thing as spending years developing actual skill, and too many people choosing the "AI" route would be a net negative for humanity. Similarly (and more intimately for me, as a software developer), having too many people rely too heavily on "AI" tools in software development is going to produce a generation of HORRIFICALLY incompetent developers. And that's not just theoretical, we're already seeing the impact of overreliance of "AI" in the industry.

Personally I am a big fan of AI in multiple contexts, but the cons just out weight it completely. Although here some great things that can be done with it:

• image upscaling. Upscaling algorithms are always a best guess and will blur a lot things. That's where AI shines. It can make crisp edge, and if you keep it to 2x, the hallucinations are minimum. I personally use Waifu2x, it has been made before the AI boom, but always has been reliable.

• Copilot. Not Microsoft's "copilot", but an ai acting as your copilot. Someone that reviews your stuff, give suggestions. You can do those suggestions or not, but that's still rely on you to do them As an artist among non artists friends, not having anyone pointing out minor details or mistakes makes it hard to improve. With AI, you just throw your art at it and say "make it better". The ai will give back an alternative version, and you can choose to add some detail it added. But you remain in control. There's been an interesting talk about that in a blender conference that I cannot find back, but that illustrate it perfectly. I never actually used that technique, but if AI wasn't so shitty I would.

For me the “Fuck” in “Fuck AI” stand for “Fuck the misuse caused by shitty company”. I'm not the only one who think that way. The worst part is that, differently from the last numerous bubbles, things are simply going too far with this one. And this is just the worst possible timing for this kind of shit to happen, especially in this wild uncontrolled way. Many sector will end up being pushed back by years because of all this automated-slop gold rush and disinformation will rule the net.

To clarify, this is about LLMs and generative image creation. Other applications and technology are probably generally outside the scope of this community.

There are lots of other technologies that would've once been called AI, until we figured out how to do them. These are all fine.

There are a handful of problems these two specific technologies share which do not look like they're likely to be solved sufficiently anytime soon:

• LLMs are predicting the next word that fits. If the answer to a given problem isn't prevalent enough in the data, or some of randomness inherent in the system makes a wrong answer fit the specific phrasing better, you may get an inaccurate result. These may be difficult to detect and make these technologies difficult to use safely for practical applications where being right, being safe, or simply not wasting the time of those around you are important things.
• Providence of the training materials. It's matching patterns found in existing works. That's part of how you get realistic results, but it also restricts creation of truly novel works. Even if you can get around that, there's still:
• A misunderstanding of what art is, and why we engage with it. Part of what makes art valuable is that it's a window into another human's brain. This is a conflict we've run into before with technologies like cameras, but there's still intentionality in shot choice, and the camera acting in predictable ways that allow the machine to disappear from the end result. This lack of the core of what makes art valuable makes creative applications nonviable for the moment.
• These are being pushed into varying aspects of our lives by the hype of how close they look to solving real world problems. But until these issues are fixed, none of the products that are being pushed will really address the needs that they're supposed to or are ready for production environments. There absolutely are exciting developments, but they're kind of happening off to the side in much more specialized areas, like the geometry solver from Google. If these things were still confined to R&D, I bet communities like this wouldn't exist. Maybe all the hype and funding will help uncover enough similar applications quick enough to make it all worth it, but I very much doubt it.

There are more issues like rate of improvement appearing to taper off extremely hard, power consumption of training destabilizing local electrical grids and worsening droughts, AI related companies having overinflated market caps and making up too large a chunk of the stock market which risks another financial crisis, AI psychosis, our educational system not being set up to deal with students having easy access to plausible looking work without mental exertion or learning on their part, and probably others that I'm forgetting at the moment.

I've yet to meet a person who is truly completely anti-AI. Instead, most people just seem completely over it. It's mislabelled, exaggerated, forced into every single physical and digital product, nearly every ad, the companies are all investing in it with circular valuations between themselves, we have to opt-out (if there even is an option), and it's just not very useful for most applications. It all feels like a massive Ponzi scheme. And all of this without actually making a profit!? When it pops, this bubble is going to destroy the economy...

Did you know the Defense Department asked for nearly 300 billion dollars for their own AI implementation? Do you think drones should use AI? The surface level use is just marketing. They're trying to put it everywhere and 99% of the uses are way worse than deluding people with chatbots.

i believe its good for accessibility use (if it works well and is actually helpful for the disabled) and language translation, maybe even grammar refinements (non-native speakers often hesitate to write due to lack of confidence etc)

not sure about anything else, overall in the current state it's a resource wasting dumpster fire imo. too few good uses considering the costs

I mainly care about chat bot AI being used as a consultant. AI using human language seems to be way to convincing for the average person even though its wrong over 50% of the time. Looking at how people are using it is a perfect example. Instead of building a specific ML pipeline for the task they just pipe it into chatgpt. Copilot on excel is an example but on hacker news there are these "solutions" popping up daily.

There are valuable uses of learning models, but I'd say they all have the following constraints:

• The relation between input and output is at most 1:1. So the output does not contain any information that cannot be derived from the input.
• The scope is sufficiently constrained so that the error rate can be meaningfully quantified.
• Dealing with the errors (including verifying that there are errors, if needed) takes less effort than just doing everything manually.

Medical and chemical research. Or anything else that actually might result in new concepts from the thing just trying every possible combination of things until something sticks. Which doesn't include how LLMs are used by average every day people (IE as a chatbot or an information gathering tool).

I alternate between saying ai can be useful and I hate it 😁

I’ve found good use for it in coding, writeups, meeting summaries, ticket summaries. It can be genuinely helpful.

The negatives are mostly people not understanding what it’s good for and not. For every hour saved when ai helps with a template or context switch or writing unit tests, I lose ten hours because some idiot just acccepted ai slop as their final output. It’s not just coding, but writing as well. AI can be a great tool to make someone more efficient but it’s only a step (or steps). At the current state it’s almost never the final product, almost never something you can accept as-is. You still have to go over it, fix it, finalize it.

And I’m especially pissed off that my company has a quota for ai use, regardless of appropriateness. wtf is that? Even when I’m working on tasks where ai can’t be helpful, I still need my minimum two ai sessions a day

The economic side of things (and companies lying about what AI actually is)