A history of the cigarette. Touches on mass production, marketing, media, addiction, and public health.

My four and a half year old just explained she loves her mother more than she loves 6-7. Which I guess is good.

😅

In case it makes you feel better, here is an excerpt from my recent breakfast conversation: "Mom, don't say that! This is so 2023!"

I had exactly the same question about embedding/intrinsic + nonlin manifold! But don’t have a good story for why center-out would had a more linear manifold, so still buy it.

BTW they found >75 dims for 90% variance — bigger than 12 (or 30), def smaller than the 1000 neurons they recorded.

also, I think that was with PCA —correct me if I'm wrong—, and PCA estimates the embedding dimensionality of a manifold not its intrinsic dimensionality, so it's an upper bound to the actual dimensionality

This paper explains it very nicely doi.org/10.1016/j.co...

this is a very interesting paper, but quoting from the paper:

"Analysis used a relatively high (30) dimensional space to ensure trajectory differences were not lost."

I'd argue that 30 is not very high compared to the millions of neurons in motor cortex...

Can even be high-d in motor cortex for isometric movements

www.biorxiv.org/content/10.1...

In fact, in some recent work, we showed that transiently blocking some proprioceptive feedback *increases* the dimensionality of dynamics along a direction orthogonal to the task manifold, with only *weak* effects on behavioral trajectories.

In fact, that last paper clearly distinguishes between ongoing dynamics and behavioral readouts, which are simply different sub-blocks of the combined neural-behavioral system dynamics. These can act on entirely different dimensions or have overlap! One could be low-d and the other not!

Or from optimal motor control.

Low-dimensional activity for simple tasks is a clear prediction from basic coding arguments

Wegmans stores with a Xerox document center and Kodak photo department in the 90's were like a Rochester megazord