Hey fellow scientists, I'm just letting you know that if you're promoting something with some junky AI image, I'm not gonna click. You've got a workshop, a conference, a new paper?
Just not gonna bother. You want a bespoke image? Hire an illustrator.
"Dashboards" are very frustrating for exactly this reason. Needlessly fancy dashboards look impressive to people who don't use them.
THIS is the take on LLMs and the job market. They can probably do some jobs, which will worsen millions of people's lives. But much more than that, people who don't understand your job will assume that it is worthless, even more than they already do.
I love when an open source project's community complains that not enough people use it and mocks people using enshitified commercial alternatives. Meanwhile, every PR you've made to clarify documentation has been rejected or lingered in unaccepted PR purgatory.
Naive empiricism and methodologism
"The dogged experimental pursuit of paraphysical results within a vacuum of genuine paraphysical theory seems to me to be methodologically barren, even if the experiments are performed with meticulous care, and even if they produce some genuinely puzzling results"
Or when the resubmission is identical except for a new Limitations section that casually dismisses the serious methodological issues from the previous review as "future work".
Do interior designers come from a pre-industrial society without electricity? Why do so many homes have so few electrical outlets?
I wish there were a one-outlet-per-meter property that I could use as a filter on housing searches.
Credit to Stefan Uddenberg, who came up with the term while we were discussing people who think the use of AI necessarily makes a product better.
New term: Badhart's Law
For Goodhart's law, the metric starts out moderately correlated with quality/performance. Then, when people work only for the metric, that correlation degrades.
For Badhart's law, the metric is shit from the start.
To add to this: now is the time we need human to human training and transfer of verifiable research and scholarship skills.
As much as tech bros want to replace your jobs, legitimate scientists are now more important than ever.
"MOM! Don't pick up the phone!"
I've already seen this episode.
This map shows widely dispersed dots representing the guesses of US survey respondents about the location of Iran on a world map. 23% placed a dot in the correct area.
Where is Iran?
Each dot represents an American's guess.
pro.morningconsult.com/articles/can... 🌐
1. Kevin slouches on his couch, bags under his eyes, playing his Switch. Next to him, an empty bag of Doritos and a few empty beer bottles. His bird, perched above him, leans in. "Umm… are you okay?" the bird asks. Kevin replies, "I read that playing Tetris helps distract your brain from anxious thoughts." 2. His bird looks concerned. "You've been playing for the last nine hours straight," he says. Kevin looks up briefly. "Have you seen the news lately?" he asks his bird. 3. His bird looks off into the distance, thinking, as Kevin goes back to his game. 4. Kevin continues playing, his bird nestled under his arm. The bird lays its head on Kevin's chest, watching him play.
Yeah, I really can't stress enough that AI code assistants royally fuck up statistical analyses. And they do it with absolute confidence.
Four panels showing how an image of a kiwi fruit can be filtered to understand contrast energy.
So @reubenrideaux.bsky.social and I decided to run an image-processing workshop at this year's EPC/APCV. We will be teaching people how to compute the contrast energy of kiwi fruit, I guess. Sign up now: visualneuroscience.auckland.ac.nz/epc-apcv-2026/
@expsyanz.bsky.social
I'm constantly trying (/failing) to get this point across.
If you're a trained expert in a field, then it may be worthwhile to question the scientific consensus of your peers.
If you're not, the scientific consensus is absolutely the best you can do and it's arbitrary foolishness to disregard it.
Vision Science makes the front page of NYT!!
featuring work from @neurofishh.bsky.social and @denilsson.bsky.social with comments from Berkeley's own @karthikshekhar.bsky.social
#visionscience
www.nytimes.com/2026/02/23/s...
We've posted a new fMRI study of semantic relations (has-part, is-a, made-of, etc.), a key aspect of language. We find that relations are represented in the same brain regions as are other semantic concepts, though voxels tend to be selective for only one relation or another.
doi.org/10.64898/202...
Why do children struggle to recognise objects in cluttered scenes more than adults? Our new paper looks at the development of visual acuity and crowding across childhood, and the way the visual system fine tunes our ability to see detail: www.nature.com/articles/s41...
Both directions of conflating human and computer terminology are problematic.
1) Per the thread below, describing AI chatbots like humans grossly overstates AI's cababilities.
2) Teaching about human cognition using metaphors from computer architecture misrepresents and oversimplifies cognition.
It’s out!
Every application has AI buttons and popups everywhere I don't want. And they won't go away!
Most software use feels like constantly swatting away annoyingly salient distractions.
I like to think about brain areas as well behaved (b/c they do what you tell them too, like V1) vs naughty (b/c 🤷, like PFC).
I'm delighted to learn that midbrain dopamine neurons are well behaved machines that convert objective value into subjective value.
www.sciencedirect.com/science/arti...
"what's the difference between a confidence interval and a posterior interval?"
me: *long long inhale*
It's frustrating to see people who used to regularly contribute to translations of open source projects give up because people are adding so many erroneous AI slop "translations".
Bots do not exhibit the same perceptual illusions and biases as humans. That's how you filter them out.
These are the Democrats I remember from when I was young. It's the same moral panic as the video game hearings from the 90s.
en.wikipedia.org/wiki/1993%E2...
How to make any instant ramen better:
1) Throw away the packet of cardboard vegetables.
2) Replace it with any food that doesn't have the texture of cardboard. Green onions, parsley, any vegetable or meat... anything.
