Search results

Perhaps this says a lot about the subject of evidence synthesis - I'm not sure they ever really talk about robust methodology just how to tabulate things in other papers and add up the numbers. They don't even seem to check that the stuff they are adding up has the same intervention and hence...

Yes I would be very cautious about the details of claims in terms of performance but the results do suggest a good capability. I also wonder how such a capability could be used in healthcare. Some doctor just aren't very good and authoritarian so AI could provide protections - but such doctors...

https://microsoft.ai/new/the-path-to-medical-superintelligence/ Paper: https://arxiv.org/abs/2506.22405 Interesting article from Microsoft around AI based diagnosis claiming very good results in diagnosing and reducing the number of tests needed.

Something all to common in many areas some people are good at getting credit and rewards others do the work

The thing pacing indicates to me is that ME is part of a dynamic system and needs to be looked at as such. So if things like blood samples are taken and compared they need to be done with knowledge of the state of exertion (and hence pem effects). Otherwise signals can be lost in fluctuations...

Posting a reminder here that the forum will be down over the weekend. See first message. I will try to keep the old forum up on a different address (https://old.s4me.info) and as read only whilst I do the update. But something could go wrong (Particularly the certificate may flag errors).

I believe so. But I do intend to redirect s4me.info to www.s4me.info which could cause a password issue. But in the settings in a browser you can find password for a given website. My password worked on the test system I have as did the email based authentication. The new system will support...

We have been working on a plan to update the forum software (from a very old version of Xenforo to the latest) which has different templates and a slightly different look and feel to it. We are now ready to do the transition having tested the update path. So we plan to do this on the weekend on...

I suspect some will be rolling their eyes until they retire - whether there are effective treatments or not.

I think that would violate the springer license - copies would be personal. I tend to think springer are one of the more restrictive publishers.

AI can help but you need to be careful in how its used and the way the overall systems are designed. AI has very much changed from when this researcher started and it is improving all the time. I do find it useful for coding - its largely replaced me looking stuff up on stack exchange - but...

This is from the letter that various researchers wrote Microsoft Word - Letter from ME Researchers to Funders There are of course many different areas worth medical research so recommending those areas where research funding can be found and hence a career build seems like the right thing to...

They should have recruited some zombies so dead subjects could exercise.

It does help those doctors who think this way believe they are justified and hence thinking its ok to treat patients badly. Perhaps prevent them from rethinking or being more careful in how they treat patients.

It does show there are big issues with the medical establishment in this country. Something that I suspect will be very slow to change. I feel the royal colleges should have been very embarrassed about their responses to NICE especially those backing LP - but they seem to be digging in.

They do briefly mention some differences in the paper And fig 2 seems to show some differences - but I'm not convinced I'm understanding what they are saying. I seem to remember Lipkin did some analysis of CSF.

Potentially a complex argument. I think LLMs do exhibit some aspects of intelligence but they are essentially running a mapping process over weights. So the notion of memory and reasoning about things that have happened/been observed is not there. That is where the agentic ideas help as memories...

So how about a simple experiment to get an LLM (with appropriate prompting) to read a few study designs and suggest potential issues (and good design elements). I'm not convinced that a foundation model would do a good job (perhaps with fine tuning but I'm not sure what the dataset would be).

Maybe we need to be more precise. As I would see it there are experimental results (i.e. stuff obtained from the physical world), interpretations of results (say papers and commentary), meta analysis (looking over interpretations of the results, or combining the raw experimental results). I feel...

I think with transformers (which are at the heart of LLMs) the situation is slightly different. You have a mapping from input to output through multiple layers. But the input and hidden layer sizes vary with input. Thus if you have a prompt with say 100 tokens you have an input of 100*embedding...