Atanas Stankov
@neuronasko.bsky.social
CUNY neuroscience PhD candidate. Interested in comp neuroscience, comp social science, causality, and philosophy of science. Previously at UMich[as visiting student]/Stanford/ETH Zurich & UZH/UIUC. Opinions are my own.
PS: Goodhart's Law is destructive unless adaption happens.
November 10, 2025 at 5:47 PM
PS: Goodhart's Law is destructive unless adaption happens.
Reposted by Atanas Stankov
To me, it's just a nothing "hypothesis" - it doesn't do any explanatory work, just dismisses the problem. And also is founded on a weird rejection of emergence that is obviously contradicted by all kinds of other evidence
November 1, 2025 at 12:16 PM
To me, it's just a nothing "hypothesis" - it doesn't do any explanatory work, just dismisses the problem. And also is founded on a weird rejection of emergence that is obviously contradicted by all kinds of other evidence
PS: I’m giving an putative example of something opposite to that sleep study.
October 30, 2025 at 3:55 PM
PS: I’m giving an putative example of something opposite to that sleep study.
Yes it’s possible single neurons can trigger millisecond-fast population coding transitions.
www.jneurosci.org/content/31/3...
www.jneurosci.org/content/31/3...
Ultrafast Population Encoding by Cortical Neurons
The processing speed of the brain depends on the ability of neurons to rapidly relay input changes. Previous theoretical and experimental studies of the timescale of population firing rate responses a...
www.jneurosci.org
October 30, 2025 at 3:55 PM
Yes it’s possible single neurons can trigger millisecond-fast population coding transitions.
www.jneurosci.org/content/31/3...
www.jneurosci.org/content/31/3...
Okay here’s an attempt: The brain does analog and digital computing. The translation from analog to digital is non-invertible dimensionality reduction and so is any experimental/theoretical method (brain recording/imaging/simulation/analysis/…).
October 29, 2025 at 5:58 PM
Okay here’s an attempt: The brain does analog and digital computing. The translation from analog to digital is non-invertible dimensionality reduction and so is any experimental/theoretical method (brain recording/imaging/simulation/analysis/…).
[Doesn’t work. Neurons are very granular in nature: dendritic spines, quantal vesicle release, spikes/spikelets, on/off gating of channels, noisy membrane, … ]
October 29, 2025 at 3:31 AM
[Doesn’t work. Neurons are very granular in nature: dendritic spines, quantal vesicle release, spikes/spikelets, on/off gating of channels, noisy membrane, … ]
Wow never thought I’d see you guys make a panel but it naturally makes sense given convergent paths of your messages.
October 26, 2025 at 3:41 PM
Wow never thought I’d see you guys make a panel but it naturally makes sense given convergent paths of your messages.
[Hypothesis: The brain is good at detecting “partial bits”.]
October 25, 2025 at 3:40 PM
[Hypothesis: The brain is good at detecting “partial bits”.]