All of Alex K. Chen (parrot)'s Comments + Replies

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"

"* 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.

Timothy Bates: "New finding: Rationality is what IQ tests measure."

https://x.com/timothycbates/status/1873031064039387560

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)

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).

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 f... (read more)

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... (read more)

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.

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].

How about TMS/tFUS/tACS => "meditation"/reducing neural noise?

Drastic improvements in mental health/reducing neural noise & rumination are way more feasible than increasing human intelligence (and still have huge potential for very high impact when applied on a population-wide scale [1]), and are possible to do on mass-scale (and there are some experimental TMS protocols like SAINT/accelerated TMS which aim to capture the benefits of TMS on a 1-2 week timeline) [there's also wave neuroscience, which uses mERT and works in conjunction with qEEG, but ... (read more)

It's one of the most important issues ever, and has a chance of solving mass instability/unhappiness caused by wide inequality in IQs in the population, by giving the less-endowed a shot to increase their intelligence.

tFUS could be one of the best techniques for improving rationality, esp b/c [AT THE VERY MINIMUM] it is so new/variance-increasing and if the default outcome is not one that we want (as was the case of Biden vs Trump, and Biden dropping out was the desireable variance-increasing move) [and is the case now among LWers who believe in AI doom], we should be increasing variance rather than decreasing it. tFUS may be the avenue for better aligning people's thought with actions, especially when their hyperactive DMN or rumination gets in the way of their ability... (read more)

Why is thing IQ measuring mostly lognormal

Worth following for his take (and YouTube videos he is creating): https://x.com/jacobrintamaki

[he's creating something around this]

What are your scores on the US Economic Experts Comparison (Interactive Matrix)?

https://www.kentclarkcenter.org/economist-comparison-interactive-matrix/

How about people who just don't "give a fuck", are Nishkama Karma, and maintain emotional composure even in times when others doubt them/do not believe them (knowing that the end is what matters).They are graceful on the inside, and maintain internal composure in the face of chaos, but others may view their movements as ungraceful particularly b/c they have the sense (and enough of a reality distortion field) to "make the world adapt to them", rather than "adapt to the world" (if they succeed, they make the world adapt to them such that the world around th... (read more)

Wow, and Mexico's fertility rate just plunged to 1.82

Isn't that a non-disparagement clause, not a NDA?

This is a very promising start on some thesis (that could go further into the theory of computation/sid mani content/https://lifeiscomputation.com/), but the "intelligence growth curves" are not very intuitive. I wager that dimensionality is more important than number of elements in determining intelligence growth curves and especially number of discontinuous jumps.

Why does F^4_65's intelligence peak out at such a low value at time 2040? Why does $F_{65537}^1$'s intelligence peak out at a lower value than equal-dimensional fields with fewer elements in them... (read more)

Has anyone tried Visia skin analysis to get feedback loops on skinhealth? (they reveal WAY more than just pictures) The problem with camera images is that visible light doesn't capture fine lines or wrinkles. My skin SEEMS to look as perfect as that of a 12-year old on the outside, but there is some small amount of wrinkling under my eyes that a visia reveals (which is why this thread prompted me to finally get dermatica tretonin) 

Collagen peptides also can help increase collagen synthesis and relieve fine lines (it's my biggest pet peeve b/c I can't ... (read more)

Newt Gingrich started out as an environmentalist (and a former member of the Sierra Club), but later turned away from it.

Even after he left congress, he still had some sympathy for environmental issues, as he wrote the book "Contract with Earth" (with an EO Wilson forward). 

Newt can be surprisingly high openness - a person oriented towards novelty can be pro-drilling (accel), pro-geoengineering, and pro-environment (which can be decel), and maybe not reconcile the two together in the most consistent way. He has been critical of both parties on climate... (read more)

There's tana https://twitter.com/AndyAyrey/status/1791679301362016519?t=Wo8e4NcWJqY4pcHRjMYgAQ&s=19

Now AlphaTensor - https://deepmind.google/discover/blog/discovering-novel-algorithms-with-alphatensor/

Bill Frist, the former Republican Senate Majority Leader under Bush (even though he had a low score by the partisan/zero compromises LCV), is now chairman at the Nature Conservancy (it's even his LinkedIn profile header) and frequently speaks out on environment and climate change issues. His kind of Republicanism is now way out of vogue.

https://www.tennessean.com/story/news/2022/08/16/tenneessee-former-senator-bill-frist-elected-chair-nonprofit-nature-conservancy/10328455002/

https://www.linkedin.com/posts/billfristmd_nature-conservation-activity-7114961629... (read more)

No one mentioned Remnote? It's the one Roam replacer that seems to beat Roam on many of the things it was good at. 

I way prefer remote storage, having lost a hard drive before, so I don't like Obsidian much. 

Related "As stated, one of the main things I make-believe is true is the overlighting intelligence with which I align myself. I speculate that I am in a co-creative relationship with an intelligence and will infinitely superior to my own. I observe that I exist within energetic patterns that flow like currents. I observe that when I act in alignment with these subtle energetic currents, all goes well, desires manifest, direction is clear, ease and smoothness are natural. I observe that I have developed a high degree of sensitivity to this energy, and that ... (read more)

It depends on how processed the PUFA is - many PUFAs in processed foods are highly heated up. Processing PUFAs in high heat is what causes peroxidizeable aldehydes/acrolein/9-HNE/advanced lipid peroxidation end-products (ALEs)/etc

But PUFAs in soybeans (or sunflower seeds w/o extra procesing) themselves are way less likely to be bad, and this is what the epidemiological evidence hints at.

For whatever reason, PUFAs are VERY strongly protective against heart disease (b/c they lower LDL) and insulin resistance. These are the leading causes of death on western ... (read more)

https://vitalik.eth.limo/general/2020/11/08/concave.html

Does Germany have a lot of food/MP testing companies? Germany seems highly represented in analytical chemistry, as I saw from the SLAS2024 conference.. (for all those people who complain about "lack of innovation" in Europe, they're all underrating analytical chemistry). This conforms to stereotypes about Germans and precision..

(and the culture of Germany is WAY more amendable to eco-consciousness/environmental health than the culture of America)

It would be nice (even in fringe cases) to have one country/area dedicated to being microplastic/pollution free ... (read more)

Isn't having boundaries also partly to do with full on consent (proactive and retroactive) with your implied preferences being unknown?

Consent is tricky because almost no one who isn't unschooled grows up consenting to anything. People grow used to consenting to things that make them feel unhappy because they don't know themselves well enough, and they trap themselves into structures that punish you for dropping out or for not opting into anything. In that sense, the system does not respect your own boundaries for your own self autonomy - your actions don'... (read more)

multiscale entropy

netlify/vercel/heroku/shinyapps/fleek (find cool associated apps!) + replit

github 

modal/EC2/docker

photonic computing

Are exotic computing paradigms (ECPs) pro-alignment?

cf https://twitter.com/niplav_site/status/1760277413907382685

They are orthogonal to the "scale is all you need" people, and the "scale is all you need" thesis is the hardest for alignment/interpretability

some examples of alternatives: https://www.lesswrong.com/posts/PyChB935jjtmL5fbo/time-and-energy-costs-to-erase-a-bit, Normal Computing, https://www.lesswrong.com/posts/ngqFnDjCtWqQcSHXZ/safety-of-self-assembled-neuromorphic-hardware, computing-related thiel fellows (eg Thomas Sohmers, Tapa Ghosh)

[this is... (read more)

Have you seen smartairfilters.com?

I've noticed that every air purifier I used fails to reduce PM2.5 by much on highly polluted days or cities (for instance, the Aurea grouphouse in Berlin has a Dyson air purifier, but when I ran it to the max, it still barely reduced the Berlin PM2.5 from its value of 15-20 ug/m^3, even at medium distances from Berlin). I live in Boston where PM2.5 levels are usually low enough, and I still don't notice differences in PM [I use sqair's] but I run it all the time anyways because it still captures enough dust over the day

Using size-1 piksters makes you really aware of all the subtle noise that your hidden plaque is gives your mind (I noticed they cleared up plaque un-reachable by floss+waterpiks+electric toothbrushes.. the first step to alignment/a faithful computation is reducing unnecessary noise (you notice this easily on microdoses of weed/psychedelics)

It's a pareto-efficient improvement to give all alignment researchers piksters to eliminate this source of noise (align the aligners first - reducing unnecessary noise is always the first step to alignment [and near-term... (read more)

What are some strategies you use to "reduce the hit" when you're about to take in potentially bad news? This is important b/c it's sometimes important to face up "bad news" earlier rather than later, and there is social loss in some people not being able to face it until it's too late, esp b/c some kinds of "bad news" aren't as incorrigible as they may initially appear (just that you need out-of-distribution strategies to make the proper amends)

[some examples of bad news: irreversible data loss, cancer diagnosis, elevated epigenetic age, loss of important ... (read more)

Do you think there are many similar threads between shock value, surprisal, and awe? Like, are there many common threads - both neurologically and sociologically?

Totalitarian societies use "awe" as a tool of control.

Did awe evolve from "something more primitive" into the complex emotion it is today? What is the simplest animal species that can feel something akin to awe? Jane Goodall wrote that even chimpanzees can feel "awe" from a waterfall, and some cetacean experts have mentioned that whales/elephants can pause at events humans might react with awe to.... (read more)

It's worth mentioning that (many) autistic people are often better at not getting into higher layers of simalucra that cause them to be trapped by maze-dom.

[SBF is an obvious counterexample]

BTW the opposite of mazedom is Newscience.org

Microplastics (and pollution - both mimetic and actual) wreck boundaries by intercalcating between boundaries/cell membranes and reducing the integrity of the boundary. To reinforce proper boundaries, it's important to maintain the organism's overall health (eg deuterated PUFAs like RT-011 help reduce oxidative stress on polyunsaturated fatty acids in the cell membrane).

[when the integrity of boundaries is weakened, the organism's channel capacity is reduced by the extra noise].

https://studio.ribbonfarm.com/p/boundary-intelligence

https://twitter.com/Sara_I... (read more)

Has anyone considered.. spiritual outliers?

ALSO outliers in neurodevelopmental trajectories/cross-correlations in how fast one region of their brain develops relative to another region of their brain.

[there was someone at USC who presented at https://web.cvent.com/event/82cd3c8a-5f63-4fb2-b394-b0b7feb49093/summary who had a talk on outliers at a more neurophysiological level]. Maybe not QUITE the level of outlier we're looking for, but more directionally so

19. Title: PoserDeep Isolation Forest Outlier Analysis of Large Multimodal Adolescent Neuroimaging Da... (read more)

https://stream.thesephist.com/updates/1711563348

Neurable headphones could be one way of crowdsourcing value signals b/c they're so wearable

Hm there are other people like https://soulsyrup.github.io/  and @guillefix and Ogi

tFUS is a way of accelerating internal alignment (look up PropheticAI). As are the Jhourney jhana people (though people like me have so much DMN noise that tFUS is needed first). Look up 

https://stream.thesephist.com/updates/1711563348

Talk to https://www.linkedin.com/in/steven-pang-625004218/ ?

Better sensors/data quality is super-impt, esp b/c data quality from traditional EEG is very poor.

https://github.com/soulsyrup

Also https://sccn.ucsd.edu/~scott/canexp00.html

https://www.linkedin.com/in/erosmarcello?miniProfileUrn=urn%3Ali%3Afs_miniProfile%3AACoAAANRGXMBF8gD4oOTUH4MeBg4W0Nu4g12yZ8&lipi=urn%3Ali%3Apage%3Ad_flagship3_feed%3BNzoHK%2BruTH%2BRrm9SgKs9Pg%3D%3D

Neurable (cody rall reviewed it) has over-the-ear EEG (which can be used ... (read more)

I mean, is there a way to measure the quality of the forecasters into the predictions? As number of forecasters expands, you get lower quality of average forecaster. Like how the markets were extremely overconfident (and wrong) about the Russians conquering Kiev...

Another example of an ethyl version being potentially better: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7827200/

Has anyone done a study on whether or not bacteria incorporate chlorotyrosine (or other damaged protens) into their proteins at first pass? This seems very doable.

We now know that oxidized DNA bases can be incorporated into the intestines of mouse DNA.

https://a16z.com/announcement/investing-in-tome-biosciences/

https://twitter.com/alexeyguzey/status/1728549209949995299

This may be far future, but what do you think of Fanzors over CRISPRs?

Also Minicircles?