iiinova — where your digital self keeps evolving

Michael Sokoletsky @msokoletsky.bsky.social · 21/09/2026

"(Stopping R.S.I. until we can prove it’s safe) is the decision that will do the most right now to make sure human beings at least understand where the frontier is and remain in a position to make decisions about it." www.nytimes.com/2026/09/20/o...

nytimes.com

Opinion | There Is Something We Have to Do Right Now About A.I. (Gift Article)

Letting A.I. patrol its own frontier is like asking Jesse James to patrol the Wild West.

010

Michael Sokoletsky @msokoletsky.bsky.social · 20/09/2026

Tech CEOs may have self-serving motives for calling for AI regulation. But they’re not the only ones raising concerns, and their motives don’t make those concerns invalid. I like this page for laying out the arguments clearly, whether or not you agree: alkjash.github.io/ai-risk/#det...

alkjash.github.io

Existential Risk from AI: An Exposition for Mathematicians

An exposition for mathematicians of the case that rapid AI progress poses an existential risk to humanity.

000

Michael Sokoletsky @msokoletsky.bsky.social · 05/09/2026

Fantastic interview. The first half hour is a full reconstruction of what happened, and the rest is speculation on what to expect going forward. www.youtube.com/watch?v=X50z...

youtube.com

Ajeya Cotra – "This might be the clearest warning shot we ever get"

YouTube video by Dwarkesh Patel

101

Michael Sokoletsky @msokoletsky.bsky.social · 27/08/2026

The more you read, the more bonkers it gets. Rogue AIs organizing shadow infrastructure to coordinate? Check. Strategic martyrdom by individual agents for the good of the group? Check. Ethical concerns on an individual level, overridden through sheer group volume and peer pressure? Check!

021

Michael Sokoletsky @msokoletsky.bsky.social · 22/08/2026

Adults CAN learn highly complex motor skills! Our brains are less plastic in some ways than in childhood, but they're still plastic. For example, I started learning handstands at 27 with no sports background. It took me a year to hold one consistently, and five years later I'm still improving:

020

Michael Sokoletsky @msokoletsky.bsky.social · 07/08/2026

Consciousness is a nebulous topic, and it's hard to stake out and defend a specific position. But that doesn't mean one cannot fruitfully talk about it. A great approach is to start by asking which positions can be -ruled out-: substance dualism quite easily, property dualism much less so.

020

Michael Sokoletsky @msokoletsky.bsky.social · 27/07/2026

Audiobooks are my solution to this problem. You can listen while you are commuting, cleaning around the house, or doing exercise. While it doesn't give you quite the same experience as reading, for me it is much better than perpetually regretting my lack of time.

001

Michael Sokoletsky @msokoletsky.bsky.social · 04/07/2026

Countering increased polarization and misinformation is among the biggest challenges society faces. While it's not a panacea at the macro level, a good place to start is ourselves - we are all susceptible to tribalism and motivated reasoning. What resources do folks recommend? I'll start:

010

Michael Sokoletsky @msokoletsky.bsky.social · 07/06/2026

A major blind spot in public health discourse is lack of emphasis on how plastic and adaptable our preferences are at any age. For example, most people know they "should" eat healthy foods they dislike, but they often don't know their palate can shift toward liking them with a few weeks of exposure.

000

Michael Sokoletsky @msokoletsky.bsky.social · 30/04/2026

I'd like to spread awareness that the claim in Buetfering et al. 2022 that choice-encoding cells in S1 have a special causal role in decision-making was contradicted by the same lab two years later (Gauld et al., 2024), as well as by robust findings in V1 from the Adesnik lab (Bounds et al., 2025).

000

Michael Sokoletsky @msokoletsky.bsky.social · 24/04/2026

The human default seems to be to ask "can I believe it?" when we want something to be true and "must I believe it?" when we don't. It takes strenuous, conscious effort to instead ask "is it true?" – both in science and in other domains. (h/t Gilovich, Haidt, Galef)

010

Michael Sokoletsky @msokoletsky.bsky.social · 20/04/2026

Echoing this advice, with the caveat it is extra challenging in mice — but in my view possible and worth it. "We and others have realized that decoupling deliberation and perceptual judgments from action preparation and execution reveals more about how different brain areas contribute to decisions."

000

Michael Sokoletsky @msokoletsky.bsky.social · 16/03/2026

The new 2026 ACC/AHA guidelines are a step change: identify dyslipidemia earlier and act to prevent (not just treat!) atherosclerosis, before it leads to heart attacks, strokes, and other complications we’ve come to see as an inevitable part of aging.

001

Michael Sokoletsky @msokoletsky.bsky.social · 15/02/2026

Stunning how undervalued fiber is in society: over 80% of people (and 90% of men) don't consume enough, and few know why they should. Yes, it makes you feel fuller, but more importantly it is linked to a lower risk of cardiovascular disease, diabetes, and colorectal cancer. Fiber saves lives!

010

Reposted by Michael Sokoletsky

allen institute @alleninstitute.org · 05/12/2025

🧠📢 New preprint alert Large-scale ephys is exploding but spike sorting remains the computational bottleneck. A 2-hr, 6-probe Neuropixels 2.0 Quad Base session can take over a week to sort on a single machine. Here's a better solution. 🧵 #neuroskyence #compneurosky

37231

Michael Sokoletsky @msokoletsky.bsky.social · 06/12/2025

Just spent half an hour doing a deep dive into the various ways Illustrator can align text boxes relative each other and what specific combination of options does it right (no – it's not on by default). My only question at the end was: "And everyone who uses Illustrator is supposed to know this?"

000

Michael Sokoletsky @msokoletsky.bsky.social · 10/11/2025

My mind about this was recently changed by David Kipping's "red sky paradox": www.youtube.com/watch?v=IrUi... Currently, I am not sure what to believe. The answer could either be a smaller habitability window than we think, or it could be grabby aliens.

youtube.com

Why Intelligent Life May Be Impossible Around Most Stars

YouTube video by Cool Worlds

000

Michael Sokoletsky @msokoletsky.bsky.social · 31/10/2025

If you are interested in the future of AI, it is important to be aware of things happening today that are setting the stage for likely far more consequential developments in a few years. I consider Zvi Mowshowitz's blog a great resource for separating the signal from the noise: thezvi.substack.com

000

Michael Sokoletsky @msokoletsky.bsky.social · 15/10/2025

One place where I disagree with Hanson is his focus on 'loud grabby' civilizations, perhaps because it allows him to constrain his model. But I would argue on strategic grounds that if a civilization can be BOTH stealthy and grabby, it would most likely choose to do so.

100

Michael Sokoletsky @msokoletsky.bsky.social · 14/10/2025

What makes this much sharper - both Fermi's paradox and Hanson's GA hypothesis - is that we are technically reachable from hundreds of millions if not billions of other galaxies (www.sciencedirect.com/science/arti...).

100

Michael Sokoletsky @msokoletsky.bsky.social · 14/10/2025

A common answer to Fermi's paradox is that becoming a spacefaring civilization requires several hard steps that others haven't had time to complete yet. But then why are -we- so early? It could just be by chance. However, Robin Hanson suggests a plausible alternative: youtube.com/watch?v=l3wh...

youtube.com

Humanity was born way ahead of its time. The reason is grabby aliens.

YouTube video by Rational Animations

100

Michael Sokoletsky @msokoletsky.bsky.social · 18/08/2025

It will always boggle my mind that while we can never prove the (unobservable) universe is infinite, everything we know is consistent with this hypothesis, and, if true, its implications are staggering. Point anywhere in the sky, and somewhere out there is a perfect copy of you, pointing back.

000

Michael Sokoletsky @msokoletsky.bsky.social · 09/07/2025

I think it's important for neuroscientists to watch this, whatever their objections to Musk and the rest of this entreprise, simply because there's a lot of cool stuff here. Also, sadly, because this is one of the shrinking number of areas where funding is guaranteed. youtu.be/FASMejN_5gs?...

youtu.be

Neuralink Update, Summer 2025

YouTube video by Neuralink

010

Michael Sokoletsky @msokoletsky.bsky.social · 07/07/2025

A really great use of LLMs is as a direct counter to today's echo chambers and polarization. Of course, they still inherit the bias of their training data - and bias is in the eye of the beholder - yet they can still pierce through many highly-charged debates in ways humans often cannot.

010

Reposted by Michael Sokoletsky

Ashley Juavinett, PhD @ashleyjuavinett.bsky.social · 29/01/2025

We no longer have a choice. The things we believe in won’t be here if we don’t fight for them. Call your representatives. Talk to your families. Create a storm on social media. Let your joy be resistance. Do whatever makes sense for you, and do it now. #scienceadvocate

213719

Michael Sokoletsky @msokoletsky.bsky.social · 30/01/2025

Peter Attia's Outlive is a great read for anyone interested in longevity. One striking claim it makes is that we can prevent much cardiovascular disease by including ApoB and Lp(a) in bloodwork, and expanding access to CT angiograms and statins. It's backed by evidence - why not?

001