The past few weeks of discussions on AI in image reconstruction and denoising have been a huge help in the understanding of what it is the algorithms are doing. I still don't understand how they do it, but now I understand the aim and the choice of training data. The fact it works is still *magic*.

Hypothesis generation is an important part of science. Case reports like these warrant more research. Kudos to @medcrisis.bsky.social. The lack of sarcastic jokes shows the level of concern.

www.youtube.com/watch?v=FPsK...

The standard AI voice-over we all know from YouTube that pronounces your product name weird (wrong?) is not giving me much confidence in your services.

I am encountering yet more men on testosterone having huge heart attacks. Bear in mind I’m only on call 1/12th of the time (ie my rota has 12 consultants) and yet I’ve now seen 5 men in the last couple of months who’ve had STEMIs in their 40s and 50s. All very fit and *externally* in great shape

I'm looking forward to hearing how one accidentally enriches uranium. Translation in alt.

This is exactly why I hate it when academics are pushed into valorisation.

You read an academic paper on an evaluation of a commercial product and someone's association is stated as only a university.

However, down in the conflicts of interest he is also an advisor and shareholder to the company.

Next up in the clinical trials ridicule: a CRO demands our secretary also gets a GCP certificate. Why do these people hate clinical trials so much that they try to sabotage them every step of the way?

Whenever I am at a meeting or symposium I find myself the one asking questions after nearly every talk. Actual questions, not rambling of my cv or giving my opinion.

Still, I apologise to my colleagues for postponing their coffee break. It must be very annoying.

The luxury of going to work by bike in 20 minutes has made me forget how bad traffic has become. It has become so much worse in the past 10 years.

The problem is not coming up with interesting physics concepts to test. The problem is translating those ideas into computational models and transforming the data to make it all work. 😅

How embarrassing. "RIP free speech" for 1.3M views.

Who knew that having progressive disease would decrease progression-free survival.

link.springer.com/article/10.1...

Patients receiving Lu-DOTATATE had an overall survival of 48 months. Patients receiving Lu-PSMA had an OS of 13 months.

Why are we burdening these patients with time spent away from their families or worrying about cancers that will never come?

ejnmmiphys.springeropen.com/articles/10....

At this point, my pre-test probability for any medical screening showing benefit is near the bottom.

Every once in a while I take a look at what is the current state of the field I used to study at university. We are nearing the precision to probe fundamental physics using atoms. Nuclear magnetic octupole moment, electric hexadecapole or
electric quadrupole shift!

dspace.mit.edu/bitstream/ha...