Our AI Midwife

59 points•eatitraw•about 4 hours ago•43 comments•

43 comments

lordnachoabout 2 hours ago

AI's biggest strength IMO is as a checklist guy.

Whatever it is being asked, it is great at throwing out new suggestions. This MRI idea sounds like another one of them.

It works well because humans are crap at checklists. For more or less anything you can think of, you'll come up with maybe 4 items, and the AI will add a few more. They may not be relevant, but the AI is very thorough.

Kye30 minutes ago

I usually tell it to go fetch research and writing/blogging from practitioners before having it produce lists.

keedaabout 1 hour ago

A few comments here making categorical assertions about AI being worse than human doctors, or no proof of it does any better. There are literally studies out there. I've read only their synopses by people on social media (including comments here on HN!) but they exist and we should be basing our discourse in that evidence.

To summarize my understanding: just like all things AI, the current frontier is quite "jagged". But there is broadly accepted evidence that AI does do better than human doctors, and even human doctors with AI, in several key aspects (like diagnostic reasoning!) while it sucks in others (pharmacology!)

And just like all things AI, as TFA suggests, the real advantage might simply come down to it being widely available, massively knowledgeable, relentless, and infinitely patient with us.

thih9about 2 hours ago

> the doctors start saying things like, “Have you considered using a surrogate?” and, “Sure, the last 5 embryo transfers failed, but what if we try a 6th and cross our fingers really hard?”

To be fair, this level of reasoning is on par with free tier chatgpt.

To me this story is less about a triumph of AI and more about a failure of the healthcare system.

mrtesthahabout 1 hour ago

There's a certain class of nerd who becomes easily distracted by (and this is being generous) technological solutions that merely paper over the underlying social problems. E.g., "I love riding in Waymos because I can avoid lewd misogynistic harassment from Uber drivers."

Tade0about 3 hours ago

> Strangely, there were many times where ChatGPT decided it would answer our medical question, but not in the form of a TV medical script. The tides had completely turned!

I've noticed this happens a lot with characters that don't have any long lines and/or have limited vocabulary.

Furby trails off pretty quickly, Abathur from StarCraft 2 (great for shortening code comments BTW) starts talking like a typical LLM after a few sentences.

mock-possumabout 2 hours ago

Holy shit are you me?? I just did a little side project last week to get my Claude code agents to RP as Abathur with me. Same results for me, it works for the first couple responses, but very quickly regresses to the mean. I tried writing out a prose/voice guide, I didn’t need it to use actual Abathur quotes, but it never seems to stick. I think there’s a thing the caveman mode does that re-injects the instructions to keep it in line, I haven’t gotten around to experimenting with anything yet though.

Tade0about 2 hours ago

That is a useful hint, thanks!

I make Abathur rephrase multi-line comments, because every LLM out there enjoys peppering code with them.

A successful recconect resets the counter. Efficient.

Doesn't always catch that distinctive style, but at least all the fluff is gone.

This Zerg is a software engineer at heart and has no tolerance for nonsense.

largbaeabout 2 hours ago

Plot twist: GP _is_ you.

socializerabout 2 hours ago

As you read these stories, keep in mind that there will never be an astralcodexten.com guest essay from someone who spent time trying to diagnose a health condition with AI and it didn't help or even caused some harm. And that's probably the most common experience.

Yes, health systems worldwide have a number of problems caused by the need to operate at a scale and manage costs, same as with primary education. AI can create some efficiencies, but it almost certainly solves fewer healthcare or education problems than most AI maximalists believe.

ryanisnanabout 2 hours ago

> AI can create some efficiencies, but it almost certainly solves fewer healthcare or education problems than most AI maximalists believe.

Do you have any information to back this claim up?

kmijyiyxfbklao29 minutes ago

He doesn't need to have it, because his comment is carefully crafted to not say anything other than signaling anti-AI sentiment.

Of course AI "almost certainly" solves fewer problems than the people that think AI can solve all the problems think. That doesn't mean anything. Also notice how he puts together "AI didn't help" and "AI caused harm", and then says that's the most common experience.

socializer8 minutes ago

The vast majority of modern medical practice is not diagnosing one-in-a-million conditions that escaped the attention of the rest of the medical establishment. It's increasingly about geriatric care, including dealing with people of diminished mental acuity. Add to this ~20% of the population who believes that vaccines cause chemtrails and you're really not solving a lot by giving people a sycophantic chatbot not beholden by consequences of its actions, willing to entertain any theory of what might be wrong with me.

Again, I think AI has applications in these niches - I really don't think we need doctors personally prescribing viagra or itch cream - but a lot of healthcare is people management.

hypferabout 1 hour ago

The claims of said maximalists are extraordinary and thus require extraordinary evidence.

Dismissing them otoh does not.

Read the full thread on Hacker News →

43 comments

Related stories