Programmer Humor

I'm not saying that AI can't be used to detect cancer, but it can't be used to diagnose any medical condition.

The doctor needs to do the diagnosis. The AI detects it. AI is definitely used like that. And not just somewhere. AI is being used widely in the health sector. So it's not just some bullshit. It's actually useful. You have no idea what you're talking about.

are we really so dumb that we'd take away the power of doctors

You seem to be a bit confused. Nobody said anything about replacing doctors. And insurancr companies weren't even mentioned.

that's one big pile of shit. don't know what else I expected from someone who blindly supports a tool with religious fervor.

How do you even "support" that? What are you saying? Neural networks have neen around for longer than you've been alive for. It's being used fucking everywhere. And not just yesterday but for many years now.

Remove AI (neural networks) from existence and you wouldn't have the following:

• They are being used in avation for anti-aircraft collision systems since the early 2010s.
• It is used extensively in Email spam filtering.
• Tagging of your photos.
• Search Engine auto-complete.
• Banking transaction monitoring against tax evasion and money laundering.
• Elevator scheduling in large buildings.
• Spell checker in text editors.
• Self-checkout in grocery stores.
• Traffic lights optimization in cities.
• Spotify recommendations
• Battery optimization in phones
• Noise cancellation in voice calls
• Retail store shelf layout
• Weather forecasting

You're just completely delusional and mad because you read something about LLMs on the internet. You're the one who's religiously attacking something so ambiguous and ubiquitous while not having any technical understanding about the underlying technology.

Explain to me how transformers, RNNs, or CNNs work and we can keep this discussion up.

Alright, fair point that it is detection and not diagnosis even if it means being pedantic. But everything else that an AI capable of is also not to be dismissed.

Yes, medicine works through diagnosis... which the AI did... We prefer false positives so the doctor may or may not perform further inspection, but it was diagnosed/flagged nonetheless. That doctor has a second opinion just with a computer instead of talking with his peers which may be busy. And I did not said that the doctor will trust the output blindly aren't I? That's why no layman should operate the AI as I said.

this is the most brain numbing take. AI can generate 15 billion compounds with medical implications. out of those only 200 are viable. out of those 15 aren't toxic to humans. problem is, it's going to take 50 years to find those 200 and another 25 years for the 15. in the meantime all medical research has been dedicated to finding those 15 medications for 75 years and have completely ignored research into specific medicines to treat problems now. the biggest joke about those 15 medicines? they're all "boner" pills because the model was trained on Pfizer data.

Well, then that is not the fault of the AI. Why did humans act irrational as you said? The AI is just trained that way. Maybe train another AI on another data then? The concept clearly works because in the 75 years we have 15 out of 15 billion, and not maybe thousand potential from a handful of manual research which still also needs to be tested.

what's your point? of course you need specialists to train the models, that's besides the point I made.

Your point does not make sense because if AI cannot do all of that, then every early cancer diagnosis being made by a computer is not worth checking. Those 15 compounds are BS. And astronomy may be wrong. As you clearly stated yourself, AI is damn good at detecting patterns that a human may miss. If that does not mean an AI is capable of something, then I don't know what is.

To believe you can live in a world where software is infallible is just foolish and naive.

Aye, sure, but you don't check the result of calculations manually, do you, because you trust the calculator.

Similarly, you probably don't stress test every joint on every car you get into, do you?

My point being that there is a point at which you do trust the tool sufficiently.

Is it at the state of cancer diagnosis? Definitely not. How about cooking? Yes, pretty much it is. I trust it more or less when cooking and making drinks.

I've also gotten a ton of actually helpfil medical information that real life doctors fucked up.

So yeah one should be critical but just don't be a complete luddite.