The IET is carrying out some important updates between 17-30 April and all of our websites will be view only. For more information, read this Announcement
This morning (March 19 2025) I attended an IEEE virtual meeting titled "Demystifying GenAI for Telehealth"
One of the attendees was a Surgeon, who made the following statement concerning his major concern about "over reliance" on the output from Gen AI Healthcare systems:-
"THE HUMAN BODY IS A CHAOTIC MACHINE!"
To me this a critical detail that engineers fail to understand about Heathcare.
Do fat tail distributions or black swan really exist in purely natural systems?
Simple answer; Yes. Such is Darwinian evolution, such as that mutant gene that allows 'us' [subset of homo-sapiens] to drink milk and live through European winters from cows milk, to name one.
The 'fat tail' thing of extreme event prediction is that it's the errors bars that make it "fat" (in common perspective). Extreme events are always on a very long thin rapier-like tail, where you can't even see the tip of the rapier (events yet to come).
I remember my extreme event lecture with the analogy of an exam supervisor trying to predict how many spare pencils would be needed given the occasional student would forget theirs. The 90% confidence number (Pareto exponential tail IIRC) was to have more spare pencils than students taking the exam.
The "Black Swan" analogy is useful but is of the 'already happened' type, so it won't be a black swan event to see a black swan. Can't step into the same river twice, ergodicity, and all that philosophical stuff ;-)
Given 'dumb humans', the artificial replicas will be 'haphazard in parts'. Remember we use humans because we are mass produced by unskilled labour 
["why send a man to the moon?"].
Do fat tail distributions or black swan really exist in purely natural systems?
Simple answer; Yes. Such is Darwinian evolution, such as that mutant gene that allows 'us' [subset of homo-sapiens] to drink milk and live through European winters from cows milk, to name one.
The 'fat tail' thing of extreme event prediction is that it's the errors bars that make it "fat" (in common perspective). Extreme events are always on a very long thin rapier-like tail, where you can't even see the tip of the rapier (events yet to come).
I remember my extreme event lecture with the analogy of an exam supervisor trying to predict how many spare pencils would be needed given the occasional student would forget theirs. The 90% confidence number (Pareto exponential tail IIRC) was to have more spare pencils than students taking the exam.
The "Black Swan" analogy is useful but is of the 'already happened' type, so it won't be a black swan event to see a black swan. Can't step into the same river twice, ergodicity, and all that philosophical stuff ;-)
Given 'dumb humans', the artificial replicas will be 'haphazard in parts'. Remember we use humans because we are mass produced by unskilled labour ["why send a man to the moon?"].
Simple answer; Yes. Such is Darwinian evolution, such as that mutant gene that allows 'us' [subset of homo-sapiens] to drink milk and live through European winters from cows milk, to name one.
I don't think this is a good example. This is a bit of a tangent but here goes 
:
I would highly recommend Wolfgang Smith's Cosmos and Transcendence. There's a chapter on Darwin's evolutionary hypothesis which implicitly includes two elements:
1. Natural selection (theory) - animal groups genetically adapting to their environment over many generations, of which there is an abundance of direct scientific observations and evidence.
2. Common descent (hypotheses) - which claims that because of 1., cats, dogs and humans therefore share a common ancestor. In fact, it does not logically follow that common descent is justified by natural selection. It is okay to call it a hypotheses, but since there is no direct scientific observations and evidence of common ancestors it cannot be elevated to a theory, and certainty isn't scientific fact. Even Darwin himself stated that the paleontological record did't match the theory - and this basic state of affairs hasn't changed. Even if it did, a fossil is not direct observation of a mouse evolving into a rabbit etc. Just because two fossils look similar, it's a fallacy to say that it proves that one evolved into another - it's arbitrary and ad hoc as to what you define a similarities. Most of the evidence put forward as proof of common decent is, in fact, merely proof for natural selection.
Maybe I'm splitting hairs but science is science and we should follow the methodology properly.
