lukko

The authors are saying that there was a step change in his wearable data, that might alert clinicians to the need for further imaging or intervention.

I was a surgical trainee not a cardiologist, but there is something interesting and unusual about the step change here. In the context of structural heart disease, this to me implies a cardiac event ?chorda tendinae rupture or possibly ?ischaemia that suddenly worsened the competency of the heart valves, reducing cardiac output and leading to the change in resting heart rate etc. This later put the patient at risk for sudden cardiac death.

I guess the authors' argument here is that the change in wearable data could alert clinicians to the need for further imaging and/or change management. But the sudden step change here may not be typical or generalisable across all causes of sudden cardiac death - they obviously state this is n=1, but I think it's worth emphasising that the change may be quite specific to this case.

We do generalise in medicine as it speeds up decision making, but we always see vitals in the patient context. Longitudinal data may be able to build a picture of what is normal for a patient - but I would say healthy physiology is fairly consistent. Even with individualised data, the problem will still be signal to noise, and working out when to alert the clinician or patient.

One of the reasons doctors don't like unnecessary scans is they have knock on effects - radiation, unnecessary procedures / biopsies, patient anxiety and extra strain on healthcare systems. A larger problem is reaching patients that are less health-conscious, that are unlikely to wear wearables or access the diagnostic tests they need.

What kind of wearable imaging are you thinking of?

I like this analogy! I think it relates to what he was saying in the video about theorems now being 'data points' that you observe and build into wider theories. The change is that now the theorem proving itself, which used to require deep understanding, can now be done very quickly. There are lots of parallels to when genAI for image generation came out, artists were frustrated their work is seen as just an end result - AI obliterates the elegance and the process.

I am not a mathematician, but I imagine it will provide a huge landscape of possibilities to explore, and be a very different field. Almost like an observational science with links to all kinds of other disciplines. Maybe it reduces the satisfaction from proofs, but people climb mountains that other people have climbed, because they learn a lot on the way.

Interesting... is it overkill to use a multimodal model like this on just audio? Could the model size be decreased dramatically, or does the multimodality help the accuracy and speed in some way?

Exactly - context is everything in art, in how it's experienced and how it is created.

I think it's important to note that a jpg of Monet is not fully experiencing the painting in any sense. Colours will not be accurately captured, the texture, the framing, the scale - it's sort of like getting a heavily watered down version of the expensive wine, saying it's cheap wine, and asking what people think.

I'd be a little bit careful here - being a jerk is quite different to non-conformity / red sneaker effect in surgery and it is not a quality you should look for.

The truly compassionate surgeons will want to improve their skills because they care about their patients. They care if they develop complications and may feel terrible if they do, the jerk may not. Being a jerk may mean that the surgeon can rise to the top, but it may not be due to surgical skill at all, they may be better at navigating politics etc.

I'm surprised at both the article and the paper - both seem very hyperbolic. This is LLMs competing against doctors in a way that is heavily weighted in the LLMs favour, which does not represent clinical practice. These reasoning cases are not benchmarks for doctors, they are learning tools.

I think it's important to note that diagnosis also relies on accurate description of the patient in the first place, and the information you gather depends on the differential diagnosis. Part of the skill of being a doctor is gathering information from lots of different sources, and trying to filter out what is important. This may be from the patient, who may not be able to communicate clearly or may be non verbal, carers and next of kin. History-taking is a skill in itself, as well as examination. Here those data are given.

For pattern recognition from plain text, especially on questions that may be in the o1's training data, I'm not surprised at all that it would outperform doctors, but it doesn't seem to be a clinically useful comparison. Deciding which investigations to do, any imaging, and filtering out unnecessary information from the history is a skill in itself, and can't really be separated from forming the diagnosis.

In the UK it would be Class Im - so low risk device that sits outside the body, with a measuring functionality.

this is all so interesting.. Are there any particular functions / parameters that are typically used, that say replicates 3 point light setups?

I guess at a certain point the number of terms becomes so large that it makes sense to just use a cube map?

Ah amazing - thank you for the response! I have a couple of related questions - is it that the non 2 pi frequencies exist, but they destructively interfere so we can't see them? My understanding is that the radial function for the electron is zero at the nucleus - there is no possibility of it being found there - but why is that the case?

This is really great. I always saw those harmonic shapes as electron orbitals, I had no idea they could be used in lighting too - so cool.

It made me wonder - why do the electron orbitals take those shapes in say a hydrogen atom? Is there a constraint on the electron and proton together that make it fit only to spherical harmonic functions?

I've just started to try and learn the basics of RL and the Bellman Equation - are there any good books or resources I should look at? I think this post is beyond my current level.

I'm most interested in how the equation can be implemented step by step in an ML library - worked examples would be very helpful.

Thank you!