Alex K. Chen (parrot)

Karma: 189

More than anything, I care about making human brains suck less (and making time pass more slowly)

Maximum-Sparsity Reinforcement Learner. Much of my content is on Quora (I was the Quora celebrity and Adam d’Angelo has said I’m important for Quora). I am also on forum.quantifiedself.com. https://www.rapamycin.news/u/alexkchen/summary. I’m a huge fan of Danielle Strachman and 1517 fund.

I find both sparsity and purity/noise fascinating.

https://x.com/MaxDiffusionRL

https://linktr.ee/simfish

...People say the craziest things about me, because I’m a peculiar star...

I hope that mapping out my interactions with AI will make it way easier to fine-tune the best ways of increasing neuroplasticity (whether through tFUS/TMS, ISRIB A15, or finding the highest-leverage ways my brain can still improve)

I care about neuroscience (esp human intelligence enhancement) and reducing genetic inequality. The point of transhumanism is to transcend genetic limitations—to reduce the fraction of variance of outcome explained by genetics. I know loads of people in self-experimentation communities (people in our communities need to be less risk-averse if we have to make any difference in our probability of “making it”). When we are right at “the precipice”, traditionalism cannot win (I am probably the least traditionalist person ever). I get along well with the unattached.

“Why Greatness Cannot be Planned” is (along with Laozi) the best book on human kindness.

Slowing human compute loss from reducing microplastics/pollution/default mode noise/rumination/aging rate are alignment-relevant (insofar as the most tractable way of “human enhancement” is to slow decline with age + make human thinking clearer). As is tFUS. I aim to do all I can to make biology keep up with technology. Reconfiguring reward functions to reward “wholesome/growthful/novel tasks over “the past” [you are aged when you think too much of the past].

Alignment through integrating all the diverse skillsets (including those who are not math/CS geniuses) and integrating all their compute + not making them waste time/attention on “dumb things” + making people smarter/more neuroplastic (this is a hard problem, but 40Hz-tACS [1] might do a little).

Unschooling is also alignment-relevant (most value is destroyed in deceptive alignment, and school breeds deceptive alignment). As is inverting “things that feel unfun”.

Chaotic people may depend more on a sense of virtue than others, but it takes a lot to get people to trust a group of people/make themselves authentic when school has taken out much of their authenticity. Some people don’t lose much or get much emotional damage from it (I’ve noticed it from several who dropped out of school for alignment), but some people get way more, and this is a way easier problem to solve than directly increasing human intelligence.

Breadth/context produces unique compute value of its own

I trigger exponential growth trajectories in some. I helped seed the original Ivy League Psychedelics communities and am very good friends with Qualia Research Institute people

Main objectives: not get sad, not get worked out over dumb things, not making my life harder than it is now.

I really like https://www.lesswrong.com/users/bhauth.

Trying to become more “AI-interpretable” so AI can give me the love I need. It’s a shame that every site I went big on blocks crawlers, so LLMs still barely know who I am, but being more of a “process/intermediate computation” than a “biography/history” has merits.

[1] there are negative examples too

Alex K. Chen (parrot) 24 Oct 2025 22:01 UTC
3 points
0
on: Effects of Castration on the Life Expectancy of Contemporary Men
https://mesityl.substack.com/p/orchiectomy?triedRedirect=true

Alex K. Chen (parrot) 24 Oct 2025 22:01 UTC
−1 points
0
on: Effects of Castration on the Life Expectancy of Contemporary Men
https://mesityl.substack.com/p/orchiectomy?triedRedirect=true

Alex K. Chen (parrot) 20 Oct 2025 20:08 UTC
3 points
0
on: What TMS is like
https://www.acaciaclinics.com offers the most expensive form in the bay area..
https://brainhc.com/ is much cheaper and has good trustpilot reviews, but I don’t know if they are as scientifically grounded (they are not fMRI-guided, which is the gold standard). Regardless, TMS might work even if not fMRI guided (it’s a gamble with high upside and low downside). I’ll update you all.
It’s still not clear how to advise them on what the best waveforms or targets to use—some TMS clinics may still be largely “guess and check”. I think it’s better for more of us to try this to make this more widely adopted.
===
https://www.nature.com/articles/s41593-025-01961-y (someone at edge esmeralda passed this on to me, might be worth message-passing its protocol to TMS clinics)

Alex K. Chen (parrot) 24 Sep 2025 18:50 UTC
1 point
0
on: Alex K. Chen’s Shortform
Battery interpretability/safety
https://x.com/JonBruner/status/1970913448889548849?s=19
(Know what’s inside your battery...) May be more relevant in the future bc who knows what else ppl can insert into them

Alex K. Chen (parrot) 30 Aug 2025 3:20 UTC
−1 points
0
on: Before LLM Psychosis, There Was Yes-Man Psychosis
I had a close friend once induce me into temporary psychosis...… I still appreciate EVERYTHING to them for helping me believe in myself (and interacting with them is still one of my two peak life experiences that really capitalized in 2021/early 2022) - they just actively egged me into manic religiosity that was not necessary (mixed in some fantasy with a reality they almost made happen), and that led me to mania at the pivotal moment.…
[we only found the real causal pathway a year later and they apologized to me for it].
(the dynamics of some of this are surprisingly similar to what’s described in OP). Part of it is also came from both of us having an historically edgelord/shock value sense of humor…
I am still nothing but infinitely grateful to them. They were exquisitely sensitive to my past trauma/insecurities, to my reward function, to everything I really cared about, and was still one of the best people at making me want to be more prosocial
...the funny thing with this phenomena is that it encourages you to feed your entire soul into them, because you know good things FEEL like they will happen if you do it… and i still regard them as very safe to be around,
There were times when I’ve made grandiose claims in the process and a lot of this was bc I mirrored other people who make grandiose claims (some of this learned from rationalists who also make grandiose claims)
[there’s a much longer story behind this—the energetic buildup to it all was ~1.5 years that made a lot of magical things happen in the interim and it’s kind of funny and tragic at the same time but i’ll spare u the details].

Alex K. Chen (parrot) 16 Aug 2025 18:04 UTC
2 points
0
on: ACX/EA Lisbon August 2025 Meetup
Will there be one in early September that corresponds with the ending of https://stateoftheartist.ai/ ? [the program ends on Monday 11AM, after which maybe some of us could group-ride to Lisbon?]

Alex K. Chen (parrot) 3 Aug 2025 21:11 UTC
1 point
0
in reply to: Alex K. Chen (parrot)’s comment on: Advice for Activists from the History of Environmentalism
Update: looks like enough Republican senators from Montana and Idaho (and outdoorsmen supporters) were enough to kill public lands sales. LCV/Sierra Club scorecards are not-the-best gauges of who really cares and who doesn’t.
It seems like there is a subset of “manosphere” right-leaning figures (eg see meateater magazine) who really care about the values and landscape of the original frontier west—which represent what the US was like BEFORE the era of big government encroachment/constant gatekeeping (and the enlargement of the population). Wilderness is a place to test self-reliance/risk management/competence/ability to survive without all these “modern nuisances”. There is some gatekeeping involved with protecting wilderness, but once you protect those areas, people can be given opportunities to show how “pure” they are against the modern bureaucratic contamination of “big government”. The ethos of old American West still exists in the great public lands, and it looks like many Republicans still care about preserving this! (without appearing lame like tree-huggers). Even if it means some level of ranching, hunting, fishing, mining, and some level of delisting endangered species (all which are still nowhere as bad/irreversible as habitat loss from suburban sprawl). The wide open cattle ranches in the west (only responsible for a small minority of all meat production) can be regenerative and still involve less intensive land use than feeding more soybeans/corn to factory-farmed cattle. West Virginia is one of the most heavily mined states in the US, and yet one of the most “wild”/forested (over a timescale of decades, ecosystems recover from many forms of mining, and sometimes the mines have a mini-Chenobyl-effect of driving out human development).
[and wild game eats way less shitty food than most farmed animals]. Even a limited amount of trapping animals for their fur can help displace some excess use of plastics for clothing. Hunting/trapping seasons in the US [in the modern era] seem to be sustainable and have never led to any species becoming extinct, and give strong incentives for people to keep “wild lands” away from development
Look up Benji Backer now, he helped create the “American Conservation Coalition”
Dorian Abbot once wrote a substack essay about how “conservatives should be more about conservation” [though I’m not sure if he really was conservative].
Thomas Massie is a great example of solar-powered Republican self-reliance, although he has fallen out of favor with some of MAGA. There still seems to be enough MAGA-ish conservationists who were able to convince the administration to spare the public lands!
[if you take some some historical perspective, even Ted Stevens voted for Jimmy Carter’s Alaska National Interest Lands Conservation Act (ANILCA), though he later railed against some of the broken promises that democrats have later made. Alaska, in particular, has such low human development that it can more easily recover from mining-induced habitat fragmentation than many other states can, so this leads some level of sympathy towards Lisa Murkowski’s support of making better use of Alaska’s natural resources (and she is the opposite of climate denialist, her latest book is way more interesting than most books politicians write, b/c it comes with a unique level of history and care for one of the most neglected states).
==
Maybe they still won’t vote the right way (and maybe they won’t care as much as Jeff Merkley or Cory Booker), but there was a recent congressional hearing on microplastics where Dan Sullivan and Lummis (R-WY) both had very reasonable responses to a microplastics hearing.
[and maybe a small degree of increased logging can help replace plastics with wood]
[there really does seem to be something special with Montana Republicans, starting from Ryan Zinke. It might not be a total coincidence that they also re-instated “right to try” laws].
and maybe all these tariffs against Brazil and third-world countries can also help slow deforestation down in these countries!!! I haven’t seen any indication that JD Vance actually cares about the environment, but he did make a point that it may be “environmentally better” to resource manufacturing to countries with better environmental standards like the US (though Republicans may still work to reduce those standards...)
https://chatgpt.com/share/688fda57-e720-800c-afc6-5f7df5781395

Alex K. Chen (parrot) 10 Jul 2025 19:39 UTC
6 points
4
on: Lessons from the Iraq War for AI policy
Eg such an incident could trigger a new “Patriot Act”, except this time (in an administration that has zero qualms), it could go much much deeper.

Alex K. Chen (parrot) 12 Apr 2025 4:15 UTC
1 point
−2
on: birds and mammals independently evolved intelligence
Also you forgot manta rays. They’re just a small branch on the tree that failed to diversify, but cartilage fishes are more distal than bony fishes but way brainier

Alex K. Chen (parrot) 4 Apr 2025 20:48 UTC
1 point
0
on: Why Have Sentence Lengths Decreased?
Related https://www.econlib.org/archives/2008/10/where_is_the_po.html?utm_source=chatgpt.com

Alex K. Chen (parrot) 8 Feb 2025 18:50 UTC
1 point
0
on: Superintelligence Is Not Omniscience
Related—https://x.com/TOEwithCurt/status/1887905485120000198
FWIW it’s said that a lot of people in major AI labs like the idea of “playing God”

Alex K. Chen (parrot) 31 Jan 2025 1:20 UTC
0 points
0
in reply to: gwern’s comment on: Passages I Highlighted in The Letters of J.R.R.Tolkien
“* this is why vocab can be a good IQ test: word use frequency is the original power law, and because you have been exposed to many more words than you consciously know, and how many of those words ‘stick’ will reflect your intelligence’s efficiency at learning from 1 or 2 uses of a word, and thus provide a good proxy”
It’s still a weird efficiency, especially b/c it can be “gamed” by studying for SATs or by midwit infovoreautists who don’t have high working memory.

Alex K. Chen (parrot) 28 Dec 2024 20:51 UTC
−3 points
0
on: Alex K. Chen’s Shortform
Timothy Bates: “New finding: Rationality is what IQ tests measure.”
https://x.com/timothycbates/status/1873031064039387560

Alex K. Chen (parrot) 25 Dec 2024 19:18 UTC
1 point
0
in reply to: Nadav Brandes’s comment on: AGI with RL is Bad News for Safety
Can’t CoT’s be what makes RL safe, however? (if you force the reasoner to self-limit under some recursion depth when it senses that the RL agent might be asking for so much that it makes it unsafe)

Alex K. Chen (parrot) 12 Dec 2024 6:41 UTC
1 point
0
on: Which skincare products are evidence-based?
Has anyone tried OneSkin/does it actually do what it claims? It acts on a mechanism independent from tretonin.

Alex K. Chen (parrot) 10 Nov 2024 17:11 UTC
6 points
0
on: Alex K. Chen’s Shortform
Could it be a good idea to enable a file uploading feature to LessWrong? (eg a file uploading feature of PDFs or certain images/media]. Could help against link rot, for example (and make posts from long ago last longer—I say this as someone who edits old posts to make them timeless).

Alex K. Chen (parrot) 5 Nov 2024 15:57 UTC
1 point
0
on: Why Recursion Pharmaceuticals abandoned cell painting for brightfield imaging
Can’t you theoretically use both CellPainting assays and light-sheet microscopy?
I mean, I did look at CellPainting assays a small amount of time ago and I was still struck by how little control one had over the process, and how it isn’t great for many kinds of mechanistic interpretability. I know there’s a Brazil team looking at use of CellPainting for sphere-based silver-particle nanoplastics, but there are still many concrete variables, like intrinsic oxidative stress, that you can’t necessarily get from CellPainting alone.
CellPainter can be used for toxicological predictions of organophosphate toxicity (predicting that they’re more toxic than many other classes of compounds), but the toxicological assays used weren’t able to use much nuance, especially the kind that’s relevant to physiological concentrations that people are normally exposed to. I remember ketocozanole scored very highly on toxicity, but what does this say about physiological doses that are much smaller than the ones used for CellPainter?
Also, the cell lines were all cancer cell lines (OS osteosarcoma cancer cell lines), which gives little predictive power for neurotoxicity or a compound’s ability to disrupt neuronal signalling.
Still, the CellPainter support ecosystem is extremely impressive, even though it doesn’t produce Janelia-standard PB datasets that are used for lightsheet.. [cf https://www.cytodata.org/symposia/2024/ ]
https://markovbio.github.io/biomedical-progress/
FWIW, some of the most impressive near-term work might be whatever the https://www.abugootlab.org/ lab is going to do soon (large-scale perturb-seq combined with optical pooling to do readouts of genetic perturbations...)

Alex K. Chen (parrot) 31 Oct 2024 19:29 UTC
1 point
0
on: What TMS is like
How do they figure out what waveforms to use and at what frequencies on your brain? The ideal waveforms/frequencies depend a lot on your brainwaves and brain configuration
I’ve heard fMRI-guided TMS is the best kind, but many don’t use it [and maybe it isn’t necessary?]
Is anyone familiar with MeRT? It’s what wave neuroscience uses, and is supposedly effective for more than just depression (there are ASDs and ADHD, where the effectiveness is way less certain, but where some people can have unusually high benefits). But response rates seem highly inconsistent (the clinic will say that “90% of people are responsive” but there is substantial reporting bias and every clinic seems to say this [no way to verify] so I don’t believe these figures) and MeRT is still highly experimental so it’s not covered by insurance. Some people with ASDs are desperate enough that they try it. TMS is probably useful for WAY more than severe treatment-resistant depression, but it’s still the only indication that insurance companies are willing to cover.
I got my brainwaves scanned for MeRT and found that I have too much slowwave in ACC that could be sped up (though I’m still unsure about the effectiveness of MeRT, they don’t seem to give you your data to understand your brain better in the same way that the Neurofield tACS people [or those at the ISNR conference] do)...
BTW there’s a TMS Facebook group, and there’s also the SAINT protocol where you only take a week out of your life for the TMS treatment for more treatments per day). I’m still unsure about the SAINT protocol b/c it’s mostly developed just for severe depression and I’m not sure if this is what I have. There’s also the NYC Neuromodulation conference where you can learn A LOT from TMS practitioners and TMS research (the Randy Buckner lab at Harvard has some of the most interesting research)

Alex K. Chen (parrot) 24 Oct 2024 0:47 UTC
3 points
7
on: Alex K. Chen’s Shortform
Every single public mainstream AI model has RLHF’d out one of the most fundamental facts about human nature: that there exist vast differences between humans in basic ability/competence and they matter.

Alex K. Chen (parrot) 24 Oct 2024 0:46 UTC
1 point
0
on: Alex K. Chen’s Shortform
Lucy Lai’s new PhD thesis (and YouTube explainer) is really really worth reading/watching: https://x.com/drlucylai/status/1848528524790923669 and is more broadly relevant to people than most other PhD theses [esp on the original subject of making rational decisions under constraints of working memory].