Donald Clark

CEO, Speaker, Blogger, Author 40+ yrs in online learning.
AI for Learning - and Productivity -

[Donald Clark @DonaldClark]

@ScratchyWalesUK @afneil These groups always think they have power but they knew and know nothing about the actual causality of Brexit. I'm sure they 'see themselves' as master manipulators. They're not.

[Donald Clark @DonaldClark]

The Think-tank/disinformation arguments make a classic blunder - that people are without agency and easily deluded. I'm on the left, voted for Brexit and don't see this as an attempt by the losers to explain their loss. Obama from the US was all over the Remain campaign - back of queue etc. Still lost the referendum.

[Donald Clark @DonaldClark]

@ScratchyWalesUK @afneil Neither significant.

[Donald Clark @DonaldClark]

@dailycheckhq Correct. That's the bar I tend to use.

[Donald Clark @DonaldClark]

Astonishing paper...
AI applications for clinical medicine continue to advance but often rely on written vignettes. Unfortunately, these are not a realistic representation of clinical medicine, where patients do not come in with a tidy write-up of key symptoms and pertinent laboratory findings.
To address this issue, Nori et al. made their AI system start from the patient’s chief complaint and decide what follow-up questions to ask and what testing to order to collect the information needed for a diagnosis.
In multiple test settings, this system achieved HIGHER ACCURACY THAN HUMAN responders and at a LOWER TOTAL COST , demonstrating its potential as a tool to assist physicians.
https://t.co/dr84maYTuW

[Donald Clark @DonaldClark]

@ScratchyWalesUK @afneil Agree but they change the game and decision making.

[Donald Clark @DonaldClark]

@ScratchyWalesUK @afneil As simple as the Russian influence argument. I buy neither.

[Donald Clark @DonaldClark]

@dailycheckhq The error rates benchmarked against human error rates. I agree. The trajectory seems clear.

[Donald Clark @DonaldClark]

@cbokhove Constraints are a design issue. In other words, achieveable.

[Donald Clark @DonaldClark]

The sound of 'peer review' going!
We should be grateful for this, as it is messy, takes too long and needs to be a more consistent filter

[Ethan Mollick @emollick]

By every measure, Opus 4.8 beats human experts as peer reviewers in a study of 124 Harvard social science papers covering a range from statistical analyses to formal models to qualitative work.

As academics, we need to figure out how to adapt the practice of science to AI, fast.

[Donald Clark @DonaldClark]

France now has a fiscal crisis that will require ECB intervention. Germany is swinging to the right and looking at not paying as much into the EU coffers. There are riots in France, Belgium and Italy. Three of the larger economies. It was not a comment on whether we should rejoin but to deny that the EU has and is becoming less attractive would be to deny the facts.

[Donald Clark @DonaldClark]

@dailycheckhq Significant progress on this front:
Brodeur et al. (2026) - excellent
Tao et al. (2026)
Mateen (2026)

[Donald Clark @DonaldClark]

Are we seeing start of the 'White Collar Rust Belt'?

[BBC Breaking News @BBCBreaking]

Barclays cutting 5,000 jobs as bank increases its use of AI https://t.co/VKZXs87cLm

[Donald Clark @DonaldClark]

@dailycheckhq Research papers need focus and achievable methodologies, they rarely solve all problems at once. But LLM based systems can cope with different answers, along with reasoning to repond.

[Donald Clark @DonaldClark]

@nooscriptor Indeed. The next stage are real trials with real patients, where, I think AI will be just as good if not better. Mainly because it can be more consistent in comms, questions and presentation back to patient. Other papers point to this.

[Donald Clark @DonaldClark]

This is ground breaking. When AI becomes better than Doctors, we all benefit...
and if we think that Doctors's don't, they are the means to the end, not the end in itself.
Also, we have a massive scarcity problem in access to Doctors, even in wealthy countries, more so in poorer countries.
This is big news....

[Tom Warren @tomwarren]

Microsoft says its AI tool is better than doctors at diagnosing conditions. The company claims to have developed an AI-powered tool it says can diagnose conditions up to four times more accurately than human doctors https://t.co/OGu900eQKs

[Donald Clark @DonaldClark]

Good point. But if the process of review in more robust and quicker, research gains. We still have academia and researchers, it's just that the process, which is at present glacially slow, and poor on peer review, will be made faster and better in quality. I'm tired of reading AI papers using GPT 4 as their tool, without peer reviewers making the obvious point that this makes the paper largely irrelevant.

[Donald Clark @DonaldClark]

@ucu There is no existential, even damage risk. The sector expanded, became bloated and is now deflating slightly.

[Donald Clark @DonaldClark]

Agree. But important to remember that they did not simply give the AI a fully worked-up case and ask for the diagnosis. Both AI & physicians started with limited information, then had to 'request' further details and investigations. Their method was appropriate for their research goal.

[Donald Clark @DonaldClark]

@piersmorgan We do have our own problems...