Matt Krause
Matt Krause
@prokraustinator.bsky.social
Brains, Machines
PLOS Bio has been our go-to when other journals pull this nonsense!
October 21, 2025 at 3:00 PM
Me too -- I caught myself staring down a 5 year old on the Metro who was wearing a batman shirt! (Wrinkles don't seem to matter, FYI).
September 9, 2025 at 6:59 PM
I wonder if there's also a semantic part? I struggle to see the flip in the whole image b/c it's so obviously Batman.

OTOH, the crops flip readily enough.
September 5, 2025 at 6:44 PM
Other answers already mention avoiding boilerplate code/other drudgery.

Drudgery sucks, but I'm not sure avoiding ALL of it is 100% good. I get lots of good ideas, as well as a sense for the tools/data while doing mundane stuff. I doubt it'd be as good if Claude 1-shotted all my initial plans.
September 5, 2025 at 6:40 PM
I'm less convinced there's part of the visual field we *know* we can ignore.

It might be irrelevant to the experimenter's task but you'd still probably want to react if a real live spider (snake, tiger, phobia-of-your-choice) popped up at the unattended location.
September 3, 2025 at 8:45 PM
I still think audition is tough! Echos, HRTFs, weird non-linearities in the speakers, and the whole system is just so much more sensitive to latency!

(Even vision is surprisingly hard to do well, especially with modern consumer stuff, but it's a cakewalk compared to the others)
September 3, 2025 at 8:40 PM
The catch is that delivering olfactory* stimuli is a nightmare: gotta sort out all kinds of chemical and physical issues before you even get to the brain part!

(*I assume taste inherits a lot of those too)
September 2, 2025 at 8:24 PM
Worse, the giant squid does not have a giant axon!

It would have been so fun to record spikes with jumper cables too....
August 18, 2025 at 3:55 PM
"Müller cells are a tool of Western oppression"?
August 14, 2025 at 8:21 PM
August 14, 2025 at 7:06 PM
It feels like something that Allen Newell would have written about, but Go in particular might not have been as well know in the 1970s...

It's sort of hinted at the end of this paper:
github.com/mrkrause/New...
github.com
August 14, 2025 at 5:14 PM
Admittedly, there are two counterexamples but one of them ("olfactory cues") didn't actually seem wildly out of scope to me.
August 14, 2025 at 4:59 PM
I'm not sure I love that framing either.

One doesn't necessarily have to *be* a perception scientist---whatever that may be---to have a result that's mostly of interest to perception scientists. I wish they had some examples of in-scope vs. out-of-scope titles.
August 14, 2025 at 4:57 PM
Yup! www.bingbrunton.com/bing

(Apparently a quite an artist too!)
Bing | bingbrunton-lab
www.bingbrunton.com
July 17, 2025 at 8:49 PM
It's also...not very good?

The vision stuff past V1 is not particularly good, ditto PD/basal ganglia.
July 17, 2025 at 8:46 PM
Tempted to email you just to see what it is now!
July 17, 2025 at 2:45 PM
I think (properly) describing many designs could take a lot more space.

Suppose you have N=100M data points, but it's heart rate every second (highly autocorrelated) from a smaller set of patients, which are themselves nested in cohort, hospital, etc. It'll take a whole paragraph!
July 15, 2025 at 4:23 PM
Surely he signed it on the back of the tweet?
July 15, 2025 at 4:09 PM
Followed by sup Bourne, I assume?
July 2, 2025 at 6:05 PM
Weirdly, her LinkedIn page lists her in VT too. Summer/second home, perhaps?

(I have also been to that Jamaica and yeah, no big universities)
June 12, 2025 at 6:29 PM
I started discussing the cost of the proposed experiments (in terms of time and money) right in the response to reviewers: people don't always realize what things cost.

It seems to help!
June 5, 2025 at 4:55 PM
I've always been conflicted about this story.

On one hand, $1M is obviously *a lot*. On the the other hand, if the manuscript was claiming the moon, like a lot of those papers do, the authors do need to back up those claims with lots of data.
June 5, 2025 at 4:22 PM
That analysis isn't quite right: they should be comparing them directly, rather than comparing each of them against zero.

www.tandfonline.com/doi/abs/10.1...
The Difference Between “Significant” and “Not Significant” is not Itself Statistically Significant
It is common to summarize statistical comparisons by declarations of statistical significance or nonsignificance. Here we discuss one problem with such declarations, namely that changes in statisti...
www.tandfonline.com
May 21, 2025 at 8:37 PM
The word “groceries” is surprisingly old—the OED has an example of its use from 1436–but I’m almost certain he didn’t know that either.
May 15, 2025 at 9:02 PM