What I meant is that a smart PDF reader could generate/guess the abbreviations, without any need to hard-code them for each instance of an abbreviation.

Great idea, but works only with LaTeX. Maybe best to implement it via the *PDF reader*, not the PDF. Some browser addons for reading PDFs do this already by linking figure refs and citations (chromewebstore.google.com/detail/googl...)
Adding the same feature for abbreviations should be feasible.

That's a very interesting approach! What happens if you select the 10-20 best quadruplet rules and select neurons/synapses randomly from these rules? Wouldn't this diversity make the network/neuron more stable and capable? Curious about your take.

But from a scientific perspective, your approach (modeling first, then data analysis) seems absolutely valid to me!

In a paper, mixing modeling, data analysis and experiments may be more difficult to browse and follow for the average reader. In systems neuroscience, we are used to seeing first the data analysis and then - in the final figure - a model (which can be ignored by those who don't understand it).

What would it really mean to achieve experimental goals that currently seem out of reach: a new blog post.

Part III - Simultaneously recording from all neurons of the human brain
gcamp6f.com/2026/01/03/a...

What would it really mean to achieve experimental goals that currently seem out of reach: a new blog post.

Part II - Recording the inputs and the output of a single neuron in real time in vivo
gcamp6f.com/2025/12/30/a...

All the work by @colah.bsky.social (blog posts, distill etc.) is amazing - clear and transparent, with little or no jargon, trying to address the hard problems head-on.

But I believe that post-publication reviews, or the suggestions that you've made, or simply transparent open reviews, which is used by many journals already, are a good way to read papers better and more efficiently. And this is a culture change that has already taken place during the last decade...

I don't think there is a way around properly reading the literature. How could we design science experiments, if the knowledge that we need for their design is somewhere in a database or LLM, but not in our heads?

Good point! But already now, everybody can contribute to post-publication peer review right now, no need for a centralized structure. I regularly write such reviews on my blog: gcamp6f.com/category/rev...
The idea is not new, see the "peeriodicals" by Romain Brette:
peeriodicals.com/peeriodicals...

If you don't learn how to write, then you don't learn how to think.

Looks like a 1:1 copy of the Cambridge Technology 8 kHz scanner. Did you notice any differences? Or have they simply licensed the technology from the existing scanner?

It's a great paper - congrats!

This objective was designed for 2P imaging. For 1P imaging, I believe that there are still some technical problems to be solved. Perhaps @voigtvision.bsky.social or @nvladimus.bsky.social can tell you more!

I'm using it mostly for 2P imaging, but it should also be relevant for expansion microscopy: Olympus XLPLN10XSVMP with NA 0.6, WD 8 mm: evidentscientific.com/en/objective...