A friend of mine is The Smartest Person I Know. He's a medical doctor that studied literal rocket science while simultaneously studying his med degree. He then did radiology post grad, then a PhD in Artificial Intelligence and know designs 3D full color X ray machines...

I'm just lining up his credentials.

He's also a very nice guy that still works in the New Zealand public healthcare system despite not needing to because he's... a very nice guy. Last time he was in London I had a long chat with him about AI and medicine. It was very eye opening. He said straight up that AI diagnostic accuracy in some radiological tools is well beyond that of doctors, in particular certain cancers. He used the metaphor of Bird Watching. He said that doctors manually checking x ray and other scans for cancer is basically no different to something like bird watching. It's human pattern recognition, he knows what cancer looks like on a scan because he's been trained to recognize it. They've now feed hundreds of thousands of scans into machine learning tools and they can out perform doctors.. The biggest issue this field is having is legal, rather than medical. "Who is responsible." is a massive legal hurdle to get over. We really like having someone to blame for a missed or misdiagnosis condition. How do we assign this when it's a robot doing the work?

Anyway, my friend thinks that the one area of AI that could result in some "miracles" will be medicine. Both in diagnostics and drug design.

Dickheads like Musk and Zuckerberg aren't interested in this.

Gramsci wrote: Sat Aug 30, 2025 2:35 pm The biggest issue this field is having is legal, rather than medical. "Who is responsible." is a massive legal hurdle to get over. We really like having someone to blame for a missed or misdiagnosis condition. How do we assign this when it's a robot doing the work?

When people hallucinate and do bad things they go to psych wards or jail.

AI is gonna protect people who will do some really bad shit. It's gonna be an additional "veil."

enframed wrote: Sat Aug 30, 2025 4:23 pm

Gramsci wrote: Sat Aug 30, 2025 2:35 pm The biggest issue this field is having is legal, rather than medical. "Who is responsible." is a massive legal hurdle to get over. We really like having someone to blame for a missed or misdiagnosis condition. How do we assign this when it's a robot doing the work?

When people hallucinate and do bad things they go to psych wards or jail.

AI is gonna protect people who will do some really bad shit. It's gonna be an additional "veil."

Less so with medical AI because it’s not LLMs, it’s trained pattern recognition. Also anything will always be followed up by the medical professional.

But if you’re from the US I understand your skepticism about all things medical. Your healthcare system encourages immortal behaviour.

I’m far more optimistic on this aspect. Mainly because of conversations with my friend. He’s a genuinely great guy, very leftwing and anti big profit medical companies. If he’s optimistic, I’m optimistic.

The bad area is stuff like this…

https://youtu.be/Tm8RG1leX8c?si=yYDZXjWO28I8DCy7

A podcast everyone here should listen to.

https://podcasts.apple.com/gb/podcast/f ... 0724674169

https://anthonymoser.github.io/writing/ ... lworld.com

I thought this was a succinct rundown of my general feelings, presented much more completely and artfully than I could manage.

my daughter told me she's dropping high school photography because the teacher wants to implement AI. she is definitely a hater.

of course i won't stop her. the idea of coupling art with AI in an introductory course would stifle many students. you have to crawl before you walk.

https://www.hollywoodreporter.com/busin ... 236361367/

"5,000 Podcasts. 3,000 Episodes a Week. $1 Cost Per Episode — Behind an AI Start Up’s Plan

Former Wondery exec Jeanine Wright is leading a new firm, Inception Point AI, that's betting on flooding the zone with audio content: “I think that people who are still referring to all AI-generated content as AI slop are probably lazy luddites."

enframed wrote: Wed Sep 10, 2025 1:11 pm https://www.hollywoodreporter.com/busin ... 236361367/

"5,000 Podcasts. 3,000 Episodes a Week. $1 Cost Per Episode — Behind an AI Start Up’s Plan

Former Wondery exec Jeanine Wright is leading a new firm, Inception Point AI, that's betting on flooding the zone with audio content: “I think that people who are still referring to all AI-generated content as AI slop are probably lazy luddites."

I’m already seeing this on YouTube with channel like Sleepy Scientist. Which is softly spoken looong videos about physics to fall asleep to. It’s borderline not AI Slop but I just can’t engage when I know channels with real people working their butts off doing similar stuff.

https://youtu.be/UclrVWafRAI?si=GWh2XYQrqIua10iL

An interesting watch. I have no idea if the interviewee is right. I normally avoid the host. But worth watching.

Edit.

It was going so well up until the last 15 minutes when it goes totally off the rails. The guest starts saying he’s 100% sure he’s living in a simulation and that everyone should buy bitcoin just be the host starts flogging brain pills. You couldn’t make it up.

Crossposted from the Politics thread:

https://www.youtube.com/watch?v=jS2FDWJgEoA

7-minute mark or so.

"This person ran the alleged text messages through Grock and Chat GPT and asked the AI, hey, you know, what are the likelihood? What's the likelihood this is real? And they both find the text messages between Tyler and his partner most likely fabricated."

So great.