I had the experience recently of reviewing a paper that was clearly AI nonsense. The abstract didn't quite hang right, I skip to the methods, sort of nonsense, the results didn't make any sense for the data and was just tables.

It got past the editors at a reputable journal...

I was a little less on the way home. Probably around 24.

:00 on the way there, because I landed so early in the morning and I was heading to a space north of Brisbane, I think I was 34 hours! But I got a nice little public bus tour of Australia north of Brisbane, and to see a few pretty towns :)

The longest day. I think today will be 38 hours for me? I can't even keep track, but it sure will be nice to be home.... In 6 hours...

Having sat on CIHR and NIH committees, so many worthy and wonderful grants don't make the cut because funding is so tight.

Wishing you more luck on the next try!

Hot take: we should not be training these LLMs to pretend to be human beings. It's the weirdest and craziest thing and gives all the wrong impressions and ideas of what they are.

They do not have feelings and stuff does not "really stick with them".

If you want to make the news impactful, don't tell people about the amount of funding. Few will cry for largest institutions that lose money.

Talk about the number of clinical trials shut down, the number of patients no longer receiving cutting edge treatment. Jobs lost, people harmed..

Def issues there. But I think a lot of what people describe as failures to replicate was just not finding "exactly" the same thing, because the samples were always a little bit different.

And because I don't think we all use our brains exactly the same way all the time. Getting better.

In neuroscience, and neuroimaging in particular, I think the "replication crisis" stems from the fact that we often identify group results that are specific to the group being studied; two "non replicating" results can both be true. Just not universally true.

Brains are complicated.

I've sat on NIH. Committees twice. It's basically free labor ($400 for about 50 hours of work). And the bar to get a findable score is INCREDIBLY high. The review is so intensly rigorous.

"So called" experts indeed. As opposed to who?

Most of the most exciting funded projects I have been involved with were Multi site collaborative studies.

Most of the most exciting stuff I see coming up is the same.

Yup.

Agree. It's good to have a moment to sit back and see our progress. It's hard to notice it as it's flying by

:)

The Empire Strikes Back, Back.