What I'm thinking about these days

permalink: https://guzey.com/what-im-thinking-about/

---

Introduction

Some topics here I’ve been thinking about for years, some for months, and some for weeks. Some are personal problems, some relate to projects I want to pursue, and some are more philosophical. What unites all of these questions is that, despite my best efforts, I haven’t been able either to answer them or get them out of my head.

If you have thoughts on any of these questions, let me know.

How to fly to the stars?

I want to host a party at the center of the universe. There are several problems with this idea.

First, I’m basically scientifically illiterate and physicists I talk to tell me that the universe doesn’t even have a center. They say that the universe looks the same in every direction from any point regardless of where you start looking from and that therefore the idea of a “center” doesn’t even make sense. This seems like a big problem.

Second, even if we found the center of the universe, it would probably be way too far away from us to get there. The universe has existed for more than 13 billion years. Who knows how far away its center is?

If we’re talking about traveling to Mars, it’s very clear how to do that. Just get on Elon’s Starship and fly. If we’re talking about traveling to the nearest star system, it’s pretty clear how to do that as well. Build a giant sail. Construct solar-powered laser plant somewhere in the solar system. Point the laser at the sail. Get to 90% of the speed of light. Get to Alpha Centauri in less than 5 years. Or if you don’t want to do that, you can blow up a bunch of nuclear bombs and use that to propel your spacecraft.

But even getting to the center of the Milky Way galaxy is another matter entirely. Alpha Centauri is 4 light-years away. The center of the Milky Way galaxy is 26,000 light years away. The only way to get there would be either via (1) a multigenerational spaceship, a faster-than-light spaceship, or (3) by figuring out how to make human cryonics/hibernation work.

As far as I know, we have no idea how to do any of these. And this is just traveling within our galaxy. So I really have little idea how to host that party. We would probably need to figure out some kind of faster-than-light travel (e.g. via wormholes or yet undiscovered physics).

In any case, if you’d like to get added to my Partiful invite, let me know.

---

Relevant pieces of media: Spore (a 2008 video game), Project X (a 2012 movie), Interstellar (a 2014 movie), Soulless (a 2012 Russian movie [trailer, watch with English subtitles]), Skins series 1 (a 2007 TV show [trailer]).

Interlude: How can I get OpenAI to send me to Rome for a few months?

I’m currently a contractor at OpenAI. It’s a fascinating company and I want to get converted to full-time. At the same time, I really want to spend a few months hanging out in Rome and doing my own thing, talking to people, writing, thinking, etc.

I’m very confident that if I do that while employed at OpenAI, it will somehow be good for the company even though I have no idea how. Everything good that ever happened to me initially started as me just trying to do something very strange or interesting for unexplainable reasons!

In any case, I have no idea how I would pitch this idea to OpenAI, and my best guess is that if I really do try to do that, it will in fact severely jeopardize my chances of getting a full-time offer.

How to defeat death?

(my definition of “defeating death” means “getting to constant probability of death every year of life”. This still implies finite expected lifespan rather than immortality)

I’m very confused about this.

First, here are some answers that I don’t think are right: Christian afterlife. Eastern-style eternal rebirth. Atheist technological singularity bringing Heaven (and/or Hell) to Earth.

If we’re being more down-to-earth, the first thing we want to do is solve all diseases, including aging.

Whenever I talk to people about this I can’t escape the feeling that it’s just too early to be working on aging. I have some kind of intuition that while a problem seems like a scientific problem rather than an engineering problem, you shouldn’t work on it too directly or you won’t explore the space of solutions. And I’m pretty sure aging is a scientific problem for us today.

A few years ago, Ryan Flynn (supported by New Science), currently a Professor at Harvard discovered a new type of RNA molecule (glycoRNA).

In 2024, a group of scientists discovered a new type of viroid (a virus but without the capsid shell) which they called Obelisks in the human body. We have no idea what they do.

Just two months ago, in March 2025, Nature published a piece saying a “textbook assumption about the brain’s most abundant receptors needs to be rewritten”, reporting that GluA2-containing AMPA receptors in the brain are often orders of magnitude more permeable to calcium than previously thought.

I just don’t see how we’re going to make interventions on the human body that affect such complicated processes as aging while we still keep discovering new types of molecules and viral particles in the human body. If it was a monocausal disease, sure. You just find a way to destroy whatever the agent of disease is and you don’t need to know how the body works or what’s happening in it aside from that at all.

But for an endogenously-driven multi-causal process, the origin of which we don’t know, the markers of which are all very sus, the feedback loops on which are measured in years? I just don’t see it happening.

I spend a lot of time thinking about the blog post ‘The “it” in AI models is the dataset.’ and I wonder how much progress we can make given how bad our data collection abilities in biology are. I’m not even saying that we have to “understand” everything that happens in our body. But at least we need to be able to get good data on what’s happening.

I like Ed Boyden’s idea that we want to “see everything and control everything” but we seem to be very far from that. I continue to think that one of the best things anyone could do for science is to just give Boyden $1b/year budget for the next 20 years for his biological tool-building and then in 20 years for-profit companies with whatever that effort enables.

Finally, we haven’t solved a single brain disease ever. I would personally say that it seems to be one of the more important organs out there and our inability to intervene on it makes me pessimistic about our knowledge of how to reverse its aging (which seems like a strictly and significantly more difficult task).

I hope to be proven wrong here and I hope that companies like NewLimit, Retro, and Altos succeed, despite everything I wrote above.

---

Boris Power notes that I should follow my own advice and reframe “defeat death” into a positive goal like “How to live more?”. Part of living more is living longer. But this also suggests things like richer and more meaningful experiences, augmentation, increasing the capacity of our bodies, etc.

For example, think of someone’s life at 40 years old and 80-years-old. In which ways are the lives of a 40-year-old usually better than an 80 year old? Can we now project this gradient backwards even further? What does life look then?

Interlude: What should I do with my life?

I’m very emotionally attached to New Science and to the idea of building new institutions of science. I’ve been wanting to run a research institute since at least 2018.

I believe that science, as a study of God, is sacred, and there’s no sense of sanctity at all left in modern academia. As far as I can tell, most academics simply do not believe in truth. Academia as a broader cultural entity certainly doesn’t. This pisses me off, and this is why I started New Science in the first place: to figure out how to create institutions that actually believe in the pursuit of truth first and foremost.

But I feel very lost now. AI really is getting incredibly powerful. I think I’m AI-pilled after all. If – that’s an if that deserves good questioning – the main constraint on the progress of science is sheer intelligence, perhaps working on scaling intelligence is the best thing one can be doing. This is what I’m doing now by working at OpenAI.

However, if both AI and robotics are going to make most human scientists unemployed anyway, then what’s the point of building new institutions of science? We’re entering the age of Pure Will.

On the other hand, I don’t see for-profit companies doing basic science research and pursuing truth just for the sake of it – and AI progress doesn’t really change that. So there’s a world in which building new institutions of science (perhaps institutions of learning, a la the first universities?) is the right thing to do after all. I hope we’re in this world. Thinking…

---

Also see: What Should You Do with Your Life? Directions and Advice, On Impact, Why you shouldn’t build your career around existential risk

What’s the fundamental nature of reality? How to study it?

People used to think that Newton’s laws were how reality worked but it turned out that if you push your system hard enough they break down. Today people think that QM and Relativity theory are how reality works. They debate all kinds of philosophical questions about multiverses, collapse, fundamental randomness or lack of it, based on our construction of these theories. I don’t understand why you’d do that.

I think the right way to think about current scientific physical theories is the same as the way we think of Newton’s laws: good approximations for the range of physical systems we can currently access. At some point in the future, we’ll invent better instruments and find the places where e.g. quantum mechanics breaks down and figure out new theories. And then the same thing will happen to those theories, and the theories after them, etc.

Why not think of known physics as a set of constraints on the real generative function of the universe, whatever it is, instead of the descriptions of the ultimate reality?

I often hear people say things like “consciousness is the only real thing and that’s what determines reality”. I mean sure, there’s a way in which everything we observe is just a product of our consciousness. But there’s a hard limit on how far you can push this argument: you can shoot yourself with a gun and you’re done. That’s it. Your conscious experience ends and yet the world – with all of its natural laws – remains.

So there are hard limits to hyperstition and there really are laws of nature, independent of individual consciousness.

How to find these “final” laws rather than just their approximations…?

---

Relevant media: Brian Skinner, From Eros to Gaia (1992 book).

Why does everyone I know believe that the world is going to end?

There’s something to the idea of Armageddon.

A couple of years ago I really got into my head the idea that the world will end in a few years due to AI singularity/doom and have been having trouble getting rid of it ever since. Plenty of my friends and people I know have never escaped.

I mean, what if it is true? What if the world really is about to end? Then nothing matters. Nothing except fully dedicating yourself to averting the end of course.

I understand early Christians much better now. If I were living then and saw Jesus doing the miracles and then telling me that the world is ending within our lifetimes, I probably would’ve dropped whatever I was doing, donated all of my money to him, and spent the rest of my life spreading the Word in an attempt to save as many souls as I could.

What I find notable is that in overtaking Christianity as the leading thought doctrine, science, and physics specifically, has replaced the Biblical story of creation, but also its story of Armageddon. Christianity said it was the Rapture. Physics says it’s the heat death of the universe.

Every single story of Armageddon ever told so far has been false. I cannot help but to reject the idea of Armageddon – whether Religious, Technological, or Scientific – entirely. And yet every single person I know still believes some version of it.

I mean it literally: every single person I know believes in either the religious Rapture, the heat death of the universe, or AI Utopia/Doom. Why?

Conclusion

In 2024, I didn’t know what to do, so I ran a Twitter poll posing this question to my followers (vox populi, vox dei). They told me to get serious about physics, so I flew into Boston at the beginning of the Spring term 2024 and spent 4 months auditing physics classes at MIT, going to a lot of research seminars, and talking to physicists.

I ended up auditing undergraduate Quantum Physics I, undergraduate Statistical Physics I, graduate Astrophysics I, and graduate Atomic and Optical Physics (I’m not claiming to have understood much in any of these classes; although, at the insistence of Isaak Freeman (a New Science Fellow), I did show up for one of the Quantum Physics I midterms, had it graded, and successfully passed it, which I think shows that I wasn’t 100% just LARPing).

In any case, it was all very fun, I met a lot of great people, at some point a Harvard physics professor offered to write me a recommendation letter for grad school, apparently being impressed with all of the harebrained questions I was asking at the weekly research seminar he was in charge of, but it became pretty clear that I’m not becoming a Physics Professor any time soon.

I spent June in Mexico City with Misha Yagudin, July to October traveling around wherever people had a couch for me to sleep on, went to New York for November, hung out in Taiwan with Misha in December, but eventually figured that OpenAI would be a good place for me to think about the future of science, and started here in January.

For 2025, my #1 goal is to get out of debt. My #2 goal is to learn to drive. My #3 goal is to get a better feeling for what the the future of science looks like.

I want OpenAI to work out and I would love to continue pursuing my curiosity and thinking about science & AI at the company. If it doesn’t, I’ll probably become unemployed for a bit again (and writing more). What’s afterwards? I’m not sure.

Acknowledgements

Thanks to Boris Power, Jackson Veigel, Maksym Sherman, Mehran Jalali, Sundari Sheldon, and for conversations and feedback.

Notes

• I promised myself to journal for 40 days 1 hour in the morning and 1 hour in the evening to figure out what I want to do and how I’m feeling, triggered by a major personal event. I’m 3 weeks in and it’s been a very interesting experience. Just getting curious about my feelings without any expectation of change, etc.

• “Учёные ищут путь к бессмертию, тогда как смерть и всеобщее равенство перед ней - это единственное, что удерживает наше общество от превращения в ад.”

• Adam Majmudar tells me that if I just keep writing about things I’m thinking abstractly about instead of actually committing to something and then executing on it then I’m never going to get there.

Appendix: 2021-02-25 concerns

[this is from my personal notes written in February 2025; I announced New Science publicly on May 13, 2021]

ok i feel terrible like my project doesn’t make any sense at all in the first place…….

ok so what’s my project? building new scientific institutions, decoupled from existing ones

but you know what existing scientific institutions work totally fine

we have a shitton of basic science

the gap is translation

the gap is just people

the gap is not in us not having enough fundamental science or us not moving quickly enough

anyone really smart can get into science or if they don’t want that get a job at a biotech

besides there’s the whole issue of biosafety, maybe it’s more important to worry about it than about speed of science itself

also church lab exists - well but it will soon not exist!

so yeah this doesn’t really make sense to work on

basic science is bs lol

tony kulesa is in biotech

everyone else too

i don’t see a reason for why i would try to build new institutions of basic science

also there’s max planck institute in germany

given how few really brilliant people there are, i think we are doing good

…

yep……

George Church exists

“if you have low gpa you’re fucked” no you’re not, you work as a tech for 2 years and you’re golden

perhaps what i should do is find a goal that’s really helpful?

not just abstarctly build new scientific institutions…..

science is really doing fine

there’s thought emporium

ok but there’rs adam strandberg

what’s up with him? well yeah i can help him i guess?

but as i myself said, he’ll find a way

and he’s a big exception

and yes church is an exception

but what about adam marblestone, he didn’t get in

and in europe, there’s Crick institute!!!! take brilliant people, get them money

plenty of places to work on crazy stuff

shoudl i expose the corruption of NIH? that’s how everyone makes a name for themselves…!

• Adam Strandberg did not get NSF

there’s literally just one Boyden and just one Church

i refused an offer from church 1.5 years - he was interested in me being a fundraiser for him. this would actually be perfect i thought he was too old and too rich i should’ve thought more about this now i feel terrible lol

wait but biotech is not working on basic science lmao or Boyden or Open Phil… I firmly believe they could’ve brought me in and it would’ve been great for them.

ok but what about the publication system

you need to publish in CNS in order to graduate and do cool shit

anyway eLife is doing great there I think and the publication system becomes better too

but anyway what does new institutions of basic science even mean

suppose i have $10b per year budget

what exactly will be better in my system? or do i really believe competition is all we need?

more people should be going into biotech anyway?

are there many people stuck in tech positions? well i’m not aware of them!

like i’m just making an argument that it’s not enough

also jed was right – i need to figure out a way to justify my project. and if i couldn’t justify it then probably i just didn’t have a very good idea of what i’m trying to do!

like how do you even improve academia?

if we don’t even know what the problems are, how am i saying i’m going to build new academic institutions

if career progression is to be decoupled, it needs to be a separate system!!!

you can always get a job at a biotech these days anyway!

MOTIVATION PABLO ADAM STRANDBERG JP BIDA ADAM MARBLESTONE CHURCH Andrew York?

at the end of the day, everyone i talk to went to a top school! can’t turn this around!!!

like who are tehse mythical people who i want to support??? i can’t find them!!!!

but maybe academia is good – motivates people to work really hard!!!!!

idk man lol

maybe i can do this new science research and shit and then start a biotech fund lol –

ok let me find the most motivational aspects for me!! —–

the other claim is that i need to enable just a very small number of individuals

call with egan? look at it

look at the original essay version

also what about the whole issue of discoveries following one another…….

https://docs.google.com/document/d/1vk4eKUbW9K6jg1QpQjeUbbDoQga6TqD1wVh9FwOMJ1U/edit#heading=h.27l98jja95yi

Noah reminded me that there’s so much fucked up shit with academia….that’s incredible. like grant cut offs… again, the entire competition system of academia… it just makes so little sense….it’s incredible… also I’m talking to people working in hot areas…but in not hot areas…shit’s fucked up…nobody is researching cryonics for example!!! ok, I’m inspired now. lol

the reason for my academia:

• it’s just going to be better

• the young dude from oxford - in seoul

• those who are not politically savvy are fucked!!!!!

2021-02-27 i feel terrible anxiety for not delivering anything with new science in so long…

i just need to find one crispt, to enable a few lost geniuses

https://docs.google.com/document/d/1CcOR4HERqVcRiZU6Nc37VckTB5Ywa-QWPNdYC2EFOso/edit#

• get more people working on biology because we need immortality, self-replicating bots etc!!!!

• the existing academia is just not big enough - that’s enough justification

• we need more people working on bio!!!!!! get everyone from google to work on biology lmao

expanding into space, colonizing mars is INCREDIBLY INSPIRATIONAL

• ok we need to live forever

• we don’t yet know the answers

• we need basic science: see crispr is a perfect example - defense mechanisms in bacteria used for editing genes - completely unexpected maybe i should do a PhD in biology lol

The absolute pillars:

• I want scientists to be able to become PIs at 25

The pillars:

• I want scientists to be able to project their budgets into the future

• I want scientists to pick their grad students

• colonize space is

• important things:

  • artificial organs

  • regeneration

  • get more people working in research

  • become rich to get the ability to new technologies first!!!

  • prepare for AGI

  • gradually replace aging cells with good cells

Appendix: how to speed up the future?

See here (due to Substack email cutoff: https://guzey.com/what-im-thinking-about/#appendix-how-to-speed-up-the-future)