My ai powered coin flip predictor fails only half of the time. Use it to predict anything!
this post was submitted on 07 Sep 2025
218 points (99.1% liked)
Fuck AI
4039 readers
698 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
founded 2 years ago
MODERATORS
It’s a sign.
Lol. Your heart horoscope, now for just £199.99 a month... But seriously, we should have some standards and medical certification for things. Because we already have esoteric pseudo-medicine and charlatans. People come to a proper doctor for proper treatment.
Spoken like someone with good insurance.
Techniker Krankenkasse
How would a stethoscope even benefit from any AI in its current form? We already have doctors who devote their lives to understanding this, that's real intelligence, not artificial intelligence.
I don't want an artificial doctor. Who would?
Healthcare industrial complex executives. Obviously.
In theory, it could pick up on subtle signals that humans usually miss.
In theory.
That sounds like one of those problems that have very very reasonable solutions with software and not AI. We're talking about amplifying sounds and picking up on some basic patterns? That has nothing to do with LLMs and probably even nothing to do with any sort of machine learning.
I'm pretty sure some mechanics have engine probes that do that, set the device to a specific part or element then run the car for a bit. From there is compares what it hears to various datasets then gives a report.
Who would?
An idiot would.
That said, if I can get this gadget, hook it up at home, and look-up the results on WebMD, I might do it.
I'm aware of the WebMD effect, but I'm not immune to it.
I usually wind up on the Cleveland clinic site
Lmao. The article author is not thrilled with AI either. Love to see it
Deep sigh.
This is definitely going to be bought in the US. They are all about worse outcomes for more money. /S
Nobody read the article? This could be a game changer if they stick with it and train, train, train. This is the exact sort of thing AI should excel at, no idea why this is failing so hard. Seems like this should be correctable pretty quickly. 🤷🏻
the "smart" gizmo analyzes the rhythms of the heartbeat and blood flow that're undetectable to the human ear, while also performing a quick electrocardiogram, or ECG, which is a test that gauges your heart's electrical activity
I guess EEGs are pretty quick as is, had one last years in my doctor's office. But still, if this turns out cheaper and faster, well, more testing never hurts. I'd love to get results in one minute and not have to strip and get wired up.
I read it. It says that 70% of doctors that used it, dropped it within 1 year and the reply from the developers is AI techbros 101. Even if your technology is good, these are not very promising figures.
This could be a game changer if they stick with it and train, train, train.
It could be, yes. Or it could be yet another AIbro grift that sucks up a lot of money for no meaningful results, but, hey, at least they're burning the planet down!
I know which seems more likely to me.
It seems like learning to detect problems is a great skill to learn, for when the hospital internet is down.
Alternately, this is the first thing I thought of
Or, hear me out, we could make a device that does all that without sticking the Incomprehensible Hallucination Demon into it.