141–150 of 368 posts
Earlier quoted context omitted.
Not even, ChatGPT, being an engine that figures out what's right by finding out what is average, is bad at understanding the atypical.
>is bad at understanding the atypical. To be fair I've found most doctors require a lot of convincing if your problem atypical as well.
I'd be happy if the government would mandate that all hospital software systems have to have portable data formats that allow sharing patient data.
Being a scifi geek and AI geek and neuroscience geek for pretty much the past 40 years, I've read countless predictions and scenarios and stories about society's response as "true" AI begins to emerge. So watching it play out for real is creating this bizarre sense of deja vu combined with fascination and frustration and also some anxiety. This article and the comments in this thread are right up that alley. I mean,…
> as "true" AI begins to emerge What, exactly, is that?
Wow imagine if the AI had been used in an unquestioning way. Someone could have died!
Being a scifi geek and AI geek and neuroscience geek for pretty much the past 40 years, I've read countless predictions and scenarios and stories about society's response as "true" AI begins to emerge. So watching it play out for real is creating this bizarre sense of deja vu combined with fascination and frustration and also some anxiety. This article and the comments in this thread are right up that alley. I mean,…
I've noticed a very weird trend when trying to incorporate GPT4 into my workflow. I lead a small team of developers and security consultants, and whenever someone runs into a problem, it is immediately searched for on the web. However, there is a real resistance/reluctance in using GPT for the same task. I think it is a combination of factors. The wrong answers are part of it, but I think another part is the anthropo…
With GPT, I don't have any of that (or maybe I just need to re-learn it?)
Also, I get a useful answer from most Google queries. GPT performs at a significantly lower bar (at least right now) - it works well for some stuff but not others, and the time it takes to figure out whether it's going to do a good job (and maybe do a couple of rounds of prompt refinement) is much more than just Googling.
Sigh.
Medicine -- a complicated, messy human art with an excessively large social component.
The medical drama House at one point had a working title of Everybody Lies. Frequently, the lies are why it's hard to diagnose, not the physical details and actual medical history.
Earlier quoted context omitted.
I my mind, the AI only needs to be more reliable than the human expert, and doctors and lawyers and other experts are "only human". There is only so much a human can read and remember. There is only so much they can be convinced to change their mind in the face of new evidence. They get tired. They get cranky. They have biases, and fears. Some doctors are just terrible. I think the AI is probably more reliable than t…
I totally agree, but we're going to need to figure out better ways to verify correctness of what the AI tells us. This will be pretty hard since we can't hardly do that with human-produced stuff either. And further if the AI is trained on human-produced stuff that contains lots of errors... I don't really know how we're going to do it.
It might look like a conversation where a panel of human experts and the AI agree on whats true and whats not. Like scientific consensus.
Eventually the AI will just run its own experiments and test its own hypothesisis.
Being a scifi geek and AI geek and neuroscience geek for pretty much the past 40 years, I've read countless predictions and scenarios and stories about society's response as "true" AI begins to emerge. So watching it play out for real is creating this bizarre sense of deja vu combined with fascination and frustration and also some anxiety. This article and the comments in this thread are right up that alley. I mean,…
I have case that 5 doctors so far can't explain. I fed in what was happening (generally speaking) and it provided a rather good list of what it "might" be. I then asked it to start describing each scenario and I was able to determine nope, that's not it or if it was plausible. Meeting with doctor no 6 tomorrow to talk about what it could be but at least I have a better insight at what the more obvious possibilities.…
The thing is, if the healthcare system is shitty and broken, it doesn't necessarily matter how good ChatGPT is. If it's applied with the intent of "streamlining" the medical system by ignoring people's concerns and rejecting their insurance claims in an automated fashion, it might be much harder to even get in front of a doctor.
You could fiddle with your dosage, work with your doctor, etc, and then your insurance company can say, "sorry, ChatGPT has decided that you can do with a lower dosage, so we won't be paying for it. The machine is a black box, so it's not accountable to you or us or anyone else, so you just have to accept it's verdict."
No amount of training data, no architecture, heck no alignment or safety work solves that problem. Because technology doesn't solve social problems.
Earlier quoted context omitted.
What I didn't predict is so many AI boosters being so miffed at regular folks pointing out the obvious: a half-working system may be tremendously exciting to AI enthusiasts, but it's not that useful to society. Recently we've had all these people complaining of moving goalposts, when in fact most people never had goalposts for AI in the first place, and certainly never had goalposts that remotely resembled those that…
This is the self driving car argument applied to things it doesn’t fit. Sure, unless a self driving car is several 9’s reliable, it’s not terribly useful for society. That’s because there’s no way for a human to be safely in the loop at highway speed. But a LLM assisting a wide rage of professionals at even a low accuracy could be part of an efficient workflow, allowing for be doctor or lawyer to give quality service…
No wonder tech CEO's are so impressed, or worried, or both. The only skillset that this thing actually duplicates well, is the one that has gotten them where they are today.