Emeritus Professor of Psych at
Stirling University
Trumpet player
Statistic-ist
Grandpa
Roger Watt
Would someone put me out of my misery please and tell me what distribution the votes fall on? It seems very stable.
It's not exponential.
All of my mainstream papers were of the n≈3 variety.
Frans doesn't say this in as many words but our answer at the time was to note that the pattern of results was the *same* in each observer. And *individual data from each observer* was always shown.
"You know, heaven forefend that autistic people don’t sit with quiet hands & listen while nonautistic researchers talk about targeting genes with the implicit aim of erasing autistic people." @ejwillingham.bsky.social on #INSAR2026 & "basic science." Subscribe/share!
buttondown.com/TPGA/archive...
"Fellows shall not act or fail to act in any way which would […] bring the Society into disrepute." But this morning ministers condemned Elon Musk FRS for enabling violence on the streets of Belfast. www.theguardian.com/uk-news/2026...
Would the @royalsociety.org like to comment?
Doing (3) is slow, effortful, uncertain of outcome and requires extrapolation. AI can't - it doesn't go beyond what is given.
I don't know why (3) is slow and effortful. It is uncertain because it is intrinsically creative.
Understanding is very different.
I have a crazy speculation.
1. Everyone has something to <say>.
2. Thinking converts <saying> into <knowing>.
3. Monotropism converts <knowing> into <understanding>.
AI really only does (1).
Doing (1) and being heard has status and is rewarding.
#WildflowerHour #lawnchallenge
Arrived last year and back again this year. Heath Spotted Orchid just in flower in time.
Also have several Common Spotted Orchid but not in flower yet.
Quite a lot of sky tonight.
On first look, this is a poor conclusion that shouldn't get taken seriously. It is based on “test-retest reliability”.
But, an individual going into a retest isn't the same person. As a minimum the outcome of the original test/diagnosis must change the person and how they approach a retest.
AI doesn't do (2).
Doing (2) delivers higher status and is therefore also rewarding.
Most of us do (2) at least some of the time. Universities expect students to do (2) all of the time.
The outcome of (2) appears, superficially, like the outcome of (1). AI can do this, but only accidentally.
Roger Watt
Roger Watt
Part one of three installments analyzing that urine panel paper making the rounds.
