As a published medical LLM researcher, I can say with high certainty that we are decades away from bringing these dreams into reality. These ideas only make sense to people outside of medicine unfamiliar with how medicine works in real life. GPT4 although perform well on board style questions, it significantly under perform in clinical based inquiries in all categories. Problem is not the fact that it hallucinates but also wrong analysis and incorrect reasoning. These insufficiencies are well documented in many medical journals. By the time LLMs be strong enough to takeover healthcare jobs, everyone else is living on the streets. This dream although will turn into reality at some point, it might never become practical in our lifetime.
If frontier AI medical assistants have already surpassed the diagnostic ability of most doctors, it’s not so profound to imagine a robotic version also outclassing surgeons/physical practitioners.
If it costs trillions of dollars to build as Sam Altman is telling us, how is it gonna save cost? It will actually inflate costs even more. Please use your head instead of mindlessly repeating dogma for a few clicks.
![โอ้เธอช่าง... - บี้เดอะสกา ft. ต๋อง เทวัญ (Prod. by ป๋าเพชร) [Official MV]](http://i.ytimg.com/vi/zbWhiLFmVL0/mqdefault.jpg)
Watch more Research@ Bangalore → goo.gle/3VSup7d
As a published medical LLM researcher, I can say with high certainty that we are decades away from bringing these dreams into reality. These ideas only make sense to people outside of medicine unfamiliar with how medicine works in real life. GPT4 although perform well on board style questions, it significantly under perform in clinical based inquiries in all categories. Problem is not the fact that it hallucinates but also wrong analysis and incorrect reasoning. These insufficiencies are well documented in many medical journals. By the time LLMs be strong enough to takeover healthcare jobs, everyone else is living on the streets. This dream although will turn into reality at some point, it might never become practical in our lifetime.
Highly informative and much interesting. Vivek your presentation is so good. God bless you - loads of laurels for ever 🙌
Many thanks for your kind words and wishes!
If frontier AI medical assistants have already surpassed the diagnostic ability of most doctors, it’s not so profound to imagine a robotic version also outclassing surgeons/physical practitioners.
Maybe in 30 years
If it costs trillions of dollars to build as Sam Altman is telling us, how is it gonna save cost? It will actually inflate costs even more. Please use your head instead of mindlessly repeating dogma for a few clicks.