sarahconstantin

Karma: 9,424

sarahconstantin 20 Apr 2026 16:17 UTC
6 points
0
on: sarahconstantin’s Shortform
links 4/20/26: https://roamresearch.com/#/app/srcpublic/page/04-20-2026
- - https://www.lieder.net/lieder/get_text.html?TextId=91385 this is the song in Hilary Mantel’s Thomas Cromwell trilogy
  - https://en.wikipedia.org/wiki/Nicholas_Culpeper the herbalist was related to the earlier courtier Thomas Culpeper
  - https://thezvi.substack.com/p/on-dwarkesh-patels-podcast-with-nvidia satisfying rundown, I agree with Zvi Mowshowitz on all points, as i often do
  - https://www.lesswrong.com/posts/WewsByywWNhX9rtwi/current-ais-seem-pretty-misaligned-to-me this seems weak to me. the kind of “misalignment” that is “AI coding agents are too eager to say they’ve solved your problem when they really haven’t” is a.) exactly the sort of thing AI labs are economically incentivized to fix, and b.) kind of unreasonable to complain about?! if you are letting your agents completely off the leash on open-ended projects, they are going to screw things up, and you will need some kind of supervisory process. it seems to me like serious professionals already know that and are successfully compensating for it? if you think of the models as TOOLS YOU USE, which of course have flaws and limits, but are trending in the right direction over the course of months and years, you tend to have a better time with them and make predictions more in line with reality...
  - https://www.verysane.ai/p/alignment-is-proven-to-be-tractable i’m not sure i agree with the *title* but i agree with the essay, i was surprised by how little of a thing “brittleness” is in large 2020s models (in the sense of “AI model does something very different from what the user intended” or “model doesn’t generalize to slightly different real-world scenarios from the training data”). language seems to really help with that!
  - https://en.wikipedia.org/wiki/List_of_human_cell_types red blood cells are by far the most common cell type in the human body

sarahconstantin 8 Apr 2026 15:53 UTC
2 points
0
on: sarahconstantin’s Shortform
links 4/8/26: https://roamresearch.com/#/app/srcpublic/page/04-08-2026
- - https://benjaminrosshoffman.com/telescopes-need-good-lenses/ this amounts to “yes, I think I am a better parent than liberals/progressives.” I am unconvinced.
  - One of his examples is that his kids are in the 99th percentile for height and weight despite not being visibly fat, thus he thinks that other parents are underfeeding their children. This doesn’t make sense. Child obesity is in fact a significant problem; on average, US parents probably feed their kids too many calories, not too few.
    The fact is that—I don’t know why, maybe measurement error at the doctor’s office, which is common—it’s not that unusual for kids to be “outliers” in size. Anecdotally it seems to me like “99th percentile” large or “1st percentile” small kids (who appear neither fat nor scrawny) are a lot more than 2% of the population. the thing is, the official distribution of heights and weights for kids is very narrow, so a “99th percentile” kid is not that enormous, and a 1st percentile kid is not that tiny. Plus, kids’ growth rates change enough over time that they usually grow out of “outlier” status. IMO, Ben’s kids are normal-sized, their “99th percentile” status may very well be random fluctuation unrelated to how he feeds them, and he is really jumping to conclusions from inadequate data to conclude that most of his peers underfeed their kids.
    I’ll grant him that being better at cooking tasty food might make kids less likely to be picky eaters, but there’s a lot of noise in that as well (I’m a pretty good cook and my kids often don’t like what grownups think is delicious). And I also am skeptical that liberals are worse cooks than conservatives on average in the US. (remember, the OP Scott post was about politics!)
    I have less objection to his overall point, which is a critique of altruism itself. Yes, people will never be as careful or as grounded, as effective, in helping others, if they’re motivated primarily by getting good-boy/good-girl points, as if their practical self-interest is entangled.
    I’m not sure that I buy that EA-style or progressive-style altruism is outright bad; for the most part, i think people aren’t harming themselves by doing it, and even if they help suboptimally they may still be helping a little. I see people as mostly putting themselves first, with a little ineffectual flavor text to make them feel good about themselves, and maybe also a little genuine help going to people who need it, and I don’t see a problem with the flavor text or the modest amount of help. I think charity is not obligatory and not the most important part of morality, but it’s probably net good; and I’m not too fussy about whether people rationalize/exaggerate how good they are, if it makes them feel better. (I also think building AI is cool and have no moral objection to it, so I don’t hold EAs blameworthy there either.)
    I don’t like the idea that we evaluate the goodness of “communities” by how “functional” they are, including via measuring fertility—yes, in an evolutionary sense, fertility is “fitness”, but that doesn’t tell me why i should care about it. I don’t want to be Satmar or Amish. If they are “virtuous”, so much the worse for virtue!
  - https://icmi-proceedings.com/ICMI-E-virtue-under-pressure.html Tim Hwang is trying to make AIs with Christian values. this freaks me out, even though i suppose it shouldn’t (Christians are not so overwhelmingly dominant worldwide that we wouldn’t also have secular AIs)
  - https://www.newyorker.com/magazine/2026/04/13/sam-altman-may-control-our-future-can-he-be-trusted entertaining and thorough profile of Sam Altman, including a very well documented pattern of people saying he lies a lot and can’t be trusted, going back to his YCombinator days.
  - https://nymag.com/intelligencer/article/why-citrini-research-sent-an-analyst-to-the-strait-of-hormuz.html I gotta admire a finance firm that sends a researcher to an active war zone to get firsthand data.

sarahconstantin 6 Apr 2026 16:36 UTC
2 points
0
on: sarahconstantin’s Shortform
links 4/6/26: https://roamresearch.com/#/app/srcpublic/page/04-06-2026
- https://balioc.tumblr.com/post/768319203871719424/looking-at-the-world-from-a-managers-perspective I like this taxonomy. I’m definitely a Fanatic.
- people’s vulnerability to sleep deprivation is a bona fide Trait: there’s substantial variability across people, substantial consistency within an individual’s sleep-dep-vulnerability over time and across measurement types, and substantial heritability. some people are more messed up by sleep deprivation than others!
- - https://pmc.ncbi.nlm.nih.gov/articles/PMC3413799/
  - https://academic.oup.com/sleep/article/43/6/zsz292/5648124
- https://german.millermanschool.com/ fun quiz

sarahconstantin 2 Apr 2026 17:51 UTC
3 points
0
on: sarahconstantin’s Shortform
links 4/2/26: https://roamresearch.com/#/app/srcpublic/page/04-02-2026
- https://www.astralcodexten.com/p/against-the-concept-of-telescopic
- - I basically agree based on personal experience that telescopic altruism is not a thing. Liberals, progressives, and EAs who care about distant strangers don’t especially neglect their own family and friends. It’s just not a pattern I observe at all.
  - It seems like everyone puts themselves and their family/friends first. and to the extent “lefties” even do tend to care more than “righties” about distant strangers, that care doesn’t seem to come out of a fixed budget such that “righties” do nicer things for their family & friends. the “extra caring about strangers” is either virtually free (it affects how you vote, or how you talk, or what you read, or what you work on, and “righties” also vote/talk/read/work, just differently) or it just makes you generically more worried, because you worry about distant people in addition to nearby people.
  - I do think there is a thing where some people put energy into following politics at the expense of their career or their housework, which is a sort of “caring about distant strangers more than matters close to home”, and can be irresponsible… but that seems to exist on both sides of the political spectrum!
- https://www.myjewishlearning.com/the-nosher/does-the-great-british-baking-show-know-what-challah-is/
- - the Great British Baking Show is laughably wrong about what challah is. are the Brits okay?
- https://medium.com/supernuclear/fairness-is-overrated-and-bragging-is-underrated-motivational-systems-for-coliving-and-beyond-c83acf9f88c9 good ideas for motivating people to do chores in group living situations (including families)

sarahconstantin 31 Mar 2026 17:44 UTC
3 points
0
on: sarahconstantin’s Shortform
links 3/31/26: https://roamresearch.com/#/app/sarahbrain/page/03-31-2026
- - https://writing.antonleicht.me/p/press-play-to-continue a cogent explanation of “why not Pause AI, given that you are concerned about AI risk?” pretty convincing to me.
  - tl;dr: you don’t want to support bad policies that don’t affect x-risk and/or are unlikely to pass
  - https://www.complexsystemspodcast.com/episodes/delve-into-compliance-theatre/ Patrick McKenzie on Delve. he believes they’re probably helping startups fake compliance paperwork. this is illegal and will get a lot of people in trouble.
  - https://en.wikipedia.org/wiki/Rule_against_perpetuities you cannot exert control over private property for more than 20 years after the owner’s death.
  - https://summerlightning.substack.com/p/the-two-memories AI and embodied vs. contemplative memories. ofc i’m mostly on the contemplative side.
  - https://acoup.blog/2026/03/25/miscellanea-the-war-in-iran/ a case that this war is a bad idea
  - https://www.programmablemutter.com/p/gooning-towards-the-fuhrer-as-policy rather satisfying (too satisfying?) theory of how the Trump administration operates
  - https://www.overcomingbias.com/p/finish-the-industrial-revolution Robin Hanson on his usual obsession. i have to ask myself, what would I be willing to sacrifice about modern liberal democratic culture in order to make it out-reproduce the Amish or Haredi in the long run… that wouldn’t make it worse, by my values, than the Amish or the Haredi? like, if you buy the argument that in the long run the future is dominated by Malthusian cultures that sacrifice personal freedom and per capita consumption to reproduce more, and you want to turn your own culture into a Malthusian one, maybe the result is not going to be a nicer culture to live in than other Malthusian cultures. Maybe the problem is precisely the Malthusianism and, if i have to pick my flavor, i might pick Jewish Malthusianism as the least-worst option.

FW26 Color Stats

sarahconstantin14 Mar 2026 15:50 UTC

21 points

0 comments2 min readLW link

(sarahconstantin.substack.com)

Things that Go Boom

sarahconstantin13 Mar 2026 19:00 UTC

61 points

2 comments8 min readLW link

(sarahconstantin.substack.com)

sarahconstantin 10 Mar 2026 16:05 UTC
10 points
0
on: sarahconstantin’s Shortform
links 3/10/26: https://roamresearch.com/#/app/srcpublic/page/03-10-2026
- https://manufacturing-operations.vercel.app/index.html Ben Reinhardt’s matrix of manufacturing operations, ranked by how valuable closed-loop self-improving automation would be
- bottlenecks in defense production:
  - https://www.businessdefense.gov/ibr/mceip/dpai/dpat3/ DPA Article III is a department involved in procuring supplies from the private sector
  - https://www.gao.gov/assets/gao-18-45.pdf solid rocket motors, used in missiles, are in shortage; carbon fiber motor casings are one of the bottlenecks
  - https://www.csis.org/analysis/first-battle-next-war-wargaming-chinese-invasion-taiwan wargames of an invasion of Taiwan
  - https://www.insensitivemunitions.org/history/the-introduction-of-castable-plastic-bonded-explosives-pbxs-in-navy-munitions/ plastic bonded explosives are also a key bottleneck
  - https://en.wikipedia.org/wiki/2025_Tennessee_manufacturing_plant_explosion explosives manufacturing is quite dangerous
  - https://www.navalsteminterns.us/nreip/labs/nswc_indianhead.html missile warheads are made and tested at a single naval facility
  - https://en.wikipedia.org/wiki/Holston_Army_Ammunition_Plant this is the only place where the explosive material for missiles is made
  - https://www.heritage.org/tidalwave/chapters/chapter-6-assessing-the-us-indo-pacific-munitions-system we are short of missiles by about an order of magnitude, Heritage estimates
  - https://www.defensenews.com/space/2017/06/21/us-industrial-base-at-risk-for-key-rocket-motor-ingredient/ there is only one supplier of a key ingredient for rocket oxidizer, ammonium perchlorate
  - https://nuclearcompanion.com/data/mark-48-torpedo/ parts of a torpedo
  - https://www.globalsecurity.org/military/systems/munitions/lrasm.htm the LRASM, a key missile
  - https://news.usni.org/2026/01/24/2026-u-s-national-defense-strategy the US’s official current strategy in the Pacific: “denial defense along the Pacific Island Chain”, i.e. preventing an invasion of Taiwan by being ready to sink any Chinese fleet attempting to attack
- https://dynomight.net/pattern/ Dynomight, food for thought
- https://catboost.ai/docs/en/ library for gradient boosting
- https://paulgraham.com/brandage.html Paul Graham on watches as an extended metaphor for what happens when technical competence is commoditized. he never mentions AI but it’s interesting to consider what the analogies would be.
- https://thezvi.substack.com/p/anthropic-officially-arbitrarily Zvi Mowshowitz, like all reasonable people, decries what’s going on here
- https://timhwang.github.io/because-china/ China is an excellent excuse for whatever you wanted to do anyway
- https://en.wikipedia.org/wiki/Distillation#History the Babylonians had distillation, but for perfume, not booze
- https://en.wikipedia.org/wiki/History_of_gunpowder it is entirely a myth that the Chinese used gunpowder only for fireworks and were unaware of its military applications. they used gunpowder weapons for centuries before Europeans did.
- https://en.wikipedia.org/wiki/Control_variates method used in Monte Carlo estimation

sarahconstantin 2 Mar 2026 20:30 UTC
8 points
0
on: sarahconstantin’s Shortform
links 3/2/26: https://roamresearch.com/#/app/srcpublic/page/03-02-2026
- https://arxiv.org/abs/2602.07164 more effective way to elicit “in character” personalities from LLMs
- https://blog.egan.bio/p/reviving-estrogen-replacement-therapy overview on why hormone therapy for menopause is losing its black-box label (tl;dr: it’s not a totally unreasonable choice)
- https://www.owlposting.com/p/neurotechnology-for-cancer-ben-woodington Abhishaike Mahajan interview with the founders of Coherence Neuro, which uses neuromodulation to disrupt (mostly brain) tumors. neurotransmitters can be tumor growth factors!
- https://en.wikipedia.org/wiki/Fixed-income_attribution#Modeling_the_yield_curve functions that define the yield curve for bonds
- https://lydialaurenson.substack.com/p/social-contracts-moral-order-weaving Lydia Laurenson on “missing stair” type dynamics, how hard it is to enforce norms against bad actors in a community
- https://www.hyperdimensional.co/p/clawed Dean Ball is sticking his neck out to do the right thing here
- https://www.clinicaltrialsabundance.blog/p/ai-wont-automatically-accelerate sorry Dario Amodei, AI can’t make all clinical trials take under a year; sometimes you need follow-ups longer than that for the actual course of disease to become clear! this is crazypants and we need to keep in mind the actual places there’s room for improvement on trial times—reducing purely regulatory delays, or procedural time (like patient recruitment time) as opposed to medically-relevant follow-up time, and finding surrogate endpoints where good ones exist. (I kinda hate surrogate endpoints cause so many suck, but yeah, clearly, when you find a good one it will speed up trials substantially.) Ruxandra Tesloianu bringing the common sense and much-needed urgency, as usual.
- https://marginalrevolution.com/marginalrevolution/2026/02/if-you-have-the-right-to-die-you-should-have-the-right-to-try.html right-to-try is the least-good and most popular form of medical deregulation, IMO, but it is still correct. Ruxandra Tesloianu again.
- https://read.technically.dev/p/how-are-companies-using-ai survey of how companies are using AI
- https://www.economicforces.xyz/p/we-dont-need-to-just-make-up-fantasy solid economics based takedown of the Citrini piece
- https://www.gleech.org/enhance I think I buy this, by Gavin Leech; nobody will ever acknowledge that the transhumanists were right, they will simply adopt “better-than-well”/human-augmentation technologies piecemeal, as those technologies become practical, safe-enough, and accessible.

Getting Back To It

sarahconstantin26 Feb 2026 20:30 UTC

38 points

1 comment7 min readLW link

(sarahconstantin.substack.com)

sarahconstantin 20 Feb 2026 16:45 UTC
32 points
0
on: sarahconstantin’s Shortform
- https://danfrank.ca/daniel-isms-50-ideas-for-life-i-repeatedly-share/ Dan Frank’s life advice.
- https://notnottalmud.substack.com/p/the-tasting-day-why-buying-5-babkas Dan Frank advises holding a tasting as a genre of social event. i should do this more.
- https://www.ribbonfarm.com/2022/03/30/fermi-estimates-and-dyson-designs/ Venkatesh Rao’s idea of a “dyson design” is the sort of quantitative/heuristic thinking that involves exponentials, order-of-magnitude estimates, and similar types of “this is so big that we can neglect that” types of reasoning. Moore’s Law (as a description of the world, and implications drawn from it) is in this category.
- https://www.science.org/doi/10.1126/science.aea1260 universal vaccines? mucosal vaccination in mice with some immune-stimulating antigens protects against multiple viruses & bacterial infections
- https://www.slowboring.com/p/why-are-young-liberals-so-depressed
  - one of my pet peeves is that people usually condemn negativity in negative terms, to the effect of “nobody likes it when you’re a whiny downer.” i don’t believe this is an effective way to fight negative affect! it violates basic principles of vibe!
  - Yglesias instead does this the right way, via inspiration. “remember We Shall Overcome, Si Se Puede, and There Is A Power In A Union? isn’t hope and optimism and working for noble causes better than depression?”
- https://marginalrevolution.com/marginalrevolution/2026/02/rebuilding-our-world-with-reference-to-strong-ai.html I’m pretty much with Tyler Cowen on this; AI will be cataclysmic, on a historical scale.
  - now, given that he believes this, i think it would be reasonable of him to outwardly panic a lot more. if things are probably going to be Quite Bad and it’s unlikely anyone can do anything about it, you’re free to have whatever reaction you like, including putting it out of your mind and spending your time on pleasant things, or running around panicking like a chicken with its head cut off. it’s hard for me to see the point of sober Smart Guy talk, under potentially cataclysmic conditions; it’s so much aversive effort for so little reward.
- https://danluu.com/julialang/ the Julia language is buggy
- https://alexaraki.substack.com/p/how-to-feel-invisible known goodguy Alex Araki wants to build a blood test for [[chronic pain]]. Is there a good molecular correlate of pain? idk but i hope so. new pain drugs have just the worst track record, and an objective outcome measure might help with that.
- https://austinvernon.substack.com/p/speed-can-reindustrialize-america remarkable claim from Austin Vernon that we don’t need industrial policy or new automation tech to reindustrialize America—we just need machine shops that use modern office software for things like quoting, billing, and other “soft” functions (not the machining itself.) He argues that American machine shops could scale to be big—they’re not inherently mom & pop operations—with the right operational setup (eg modern logistics). includes a list of companies he thinks deserve investment.
- https://en.wikipedia.org/wiki/Gantt_chart why didn’t Napoleon just use Gantt charts for logistical planning? because they wouldn’t be invented for another hundred years. and they were instrumental in WWI.

sarahconstantin 11 Feb 2026 17:21 UTC
5 points
0
on: sarahconstantin’s Shortform
links 2/11/26: https://roamresearch.com/#/app/srcpublic/page/02-11-2026
- https://www.clinicaltrialsabundance.blog/p/the-moderna-rtf-and-the-cost-of-regulatory Ruxandra Tesloianu on why FDA changing its mind to reject trial protocols it previously approved is harmful for biomedical progress
- https://www.drmaciver.com/2015/04/some-empirically-derived-testing-principles/ D.R. Maciver advice on software testing
- https://notebook.ldeming.com/garden/index.html Laura Deming on how, if longevity ever really works, it would be creating a new sort of creature with a mindset very different from a “normal”-lifespan human
- https://mcfunley.com/choose-boring-technology advice on picking a software stack
- https://www.owlposting.com/p/heuristics-for-lab-robotics-and-where excellent overview of new directions in (biology) lab automation by Abhishaike Mahajan
- https://authorshipai.vercel.app/ writing app that’s more of a thinking partner than just writing for you
  - I don’t care for AI assistance in writing myself, but I do like attempts to make more “wholesome” tools that don’t degenerate into machines for cheating on homework and mass-producing slop
- https://priori.chat/ allows you to toggle the “implicit assumptions” in an LLM’s response; though upon playing with it, i consistently preferred Claude’s default choices anyhow. this is definitely the kind of thing i want to exist though.
- https://www.bitsaboutmoney.com/archive/fraud-investigation/ Patrick McKenzie makes the case that we should get more serious about investigating benefits fraud, using the tools of the finance industry

sarahconstantin 5 Feb 2026 19:49 UTC
3 points
0
on: sarahconstantin’s Shortform
links 2/5/26: https://roamresearch.com/#/app/srcpublic/page/02-05-2026
- https://paulgraham.com/mean.html obviously everything Paul Graham said during his tenure at YC was more about what he wanted founders to be like than what they necesssarily were like...but i do think i agree, actually, that mean people fail. at least, I agree with the weaker claim that the most overtly nasty people are usually jobless and socially isolated. successful people may be ruthless, but almost all of them have manners.
- https://www.anthropic.com/news/claude-is-a-space-to-think I like that Anthropic is branding itself as the less-skeezy AI. i want to use products that have a “we’re the Good Guys, we keep it classy” brand identity, even if obviously there are some compromises made for the sake of profit. a company culture of “we don’t do that lowest-common-denominator shit, we have Honor” is obviously toothless from an incentive perspective but can in fact hold out for some time and provide a better-quality customer experience.
- https://en.wikipedia.org/wiki/Assassination_of_Shinzo_Abe somehow i didn’t realize this happened. wild story.
- https://www.construction-physics.com/p/on-technologies-vs-commodities Brian Potter finds that technologies vs commodities are more of a continuum
- https://arxiv.org/abs/2503.21602 scaffolding for text-to-SQL
- https://www.metaculus.com/questions/41139/china-attacksblockades-taiwan-in-2026/ why is this so flat?
- https://www.bbc.com/news/articles/cevz13xkxpro (from Sept 2025) we are finally beating Huntington’s

Biotech Startup Stats

sarahconstantin5 Feb 2026 18:40 UTC

21 points

0 comments4 min readLW link

(sarahconstantin.substack.com)

sarahconstantin 27 Jan 2026 17:46 UTC
14 points
0
on: sarahconstantin’s Shortform
links 1/27/26: https://roamresearch.com/#/app/srcpublic/page/01-27-2026
- https://en.wikipedia.org/wiki/Insurrection_Act_of_1807 last invoked in the 1990s LA riots
- https://www.sluggish.xyz/p/are-we-all-just-having-understandable
  - one annoying thing about anti-psychiatry people is that they equivocate between “the person labeled insane is behaving in a valid fashion, and deserves acceptance/respect rather than pathologization” and “the person labeled insane is causing their own problems and ought to stop; they deserve criticism rather than therapeutic treatment”
  - either of these can, imo, sometimes be true, but it’s weird when someone is trying to pose as both the defender of the misunderstood and the teller of hard truths to malingerers.
- https://www.darioamodei.com/essay/the-adolescence-of-technology not a lot of news in here; Dario Amodei lays out “Anthropic-ism”, the summary of their research & policy programs of the past several years.
  - he and Paul Christiano deserve credit as the architects and prophets of the age we’re living in. things did, after all, turn out their way. i was always critical of their theories as basically blurring boundaries between “real intelligence” and the thing LLMs do, but in their defense, that boundary has turned out to be blurrier than I expected.
  - i think i might be at least sympathetic to Anthropic-ism. it turns out that a world with these kinds of gradually-and-quickly-improving AI tools is pretty nifty, and also involves all kinds of novel risks that are short of the “everyone dies instantly” scenario but are still quite worrying. sure, if they had listened to Eliezer they could simply have not chosen to go down this path, and there’s more than a hint of vainglory in the motivations of people who want to build powerful AI, but all in all I’d rather the 2020s had these tools than not. it’s not like the “end of the end of history” wouldn’t have happened anyway, and at least this way we have some powerful opportunities as well as depressing developments.
- https://oll.libertyfund.org/titles/bagehot-lombard-street-a-description-of-the-money-market 1873 classic about central banking
- https://steve-yegge.medium.com/welcome-to-gas-town-4f25ee16dd04 this sounds quite complicated but interesting. multi-agent management system.
- https://marginalrevolution.com/marginalrevolution/2026/01/a-more-intelligent-comment-than-most-of-the-emotional-reactions-we-are-seeing.html why do comments have to be “intelligent” per se? you can just say the obvious thing that expresses your values! simple things can be correct.
- https://epirium.com/wp-content/uploads/2025/11/GSA2025_Poster_Epirium_MF-300_Phase-1.pdf that sure is a drug for sarcopenia in phase 1 trials
- https://www.axios.com/2026/01/24/ice-cbp-shooting-minneapolis-democrats-reaction imagine, congress actually putting checks on the executive branch!
- https://writing.antonleicht.me/p/how-ai-safety-is-getting-middle-powers too much inside baseball for me, but probably a useful pointer to what experts think
- https://corticallabs.com/ sounds like a joke? you do not want to do any computation on neurons, they are slow and fragile. (you might want to run brain-inspired algorithms, but on semiconductors!)
- https://centuryofbio.com/p/sid extreme personalized medicine approach to cancer—it worked! after the standard of care failed, single-cell sequencing exposed a fibroblast-like phenotype, and an anti-fibroblast radiotherapy shrank the tumor enough to be operable.
- https://www.sam-rodriques.com/post/the-humanity-project this sounds entirely desirable but insanely expensive. sure, i’d like this, and a pony too. and world peace. can you just do that these days? i’d be delighted if so.
- https://defenseanalyses.org/work/the-maha-pentagon/ ehhh like so much DARC stuff it’s hard to tell if they’re serious. biohacking for soldiers?
  - what I’d really like to see in the health/military space is to take advantage of the DOD’s truly massive biobank resources & make more data accessible to researchers from modern omics techniques.
- https://en.wikipedia.org/wiki/Hanfu_Movement historical costume is big in China.
- https://www.humanlayer.dev/blog/writing-a-good-claude-md keep it simple, stupid (one perspective)

sarahconstantin 21 Jan 2026 18:20 UTC
5 points
0
on: sarahconstantin’s Shortform
links 1/21/26: https://roamresearch.com/#/app/srcpublic/page/01-21-2026
- safety of BPC-157:
  - https://www.sciencedirect.com/science/article/abs/pii/S027323002030091X good results in animals
  - https://clinicaltrials.gov/study/NCT02637284 Phase I trial never published results
- efficacy of BPC-157:
- Shaun Tan’s art: https://www.npr.org/2022/11/03/1131598819/shaun-tan-creature-the-lost-thing-the-arrival
- https://merge.io/blog Merge Labs, the successor to Sumner Norman’s Forest Neurotech. Sam Altman is involved.
  - “entirely new technologies that connect with neurons using molecules instead of electrodes, transmit and receive information using deep-reaching modalities like ultrasound, and avoid implants into brain tissue.”

sarahconstantin 21 Jan 2026 18:18 UTC
5 points
0
in reply to: sarahconstantin’s comment on: What comes after Roam’s renaissance?
as of now, i’m using a Claude Code-generated app for the same logging and diary functions as my private Roam. I had no compelling reason to switch to a different app, but a natural-language-queryable, endlessly reconfigurable interface? yes please. it can do everything I wanted out of Roam plus anything else I want!

sarahconstantin 15 Jan 2026 13:02 UTC
2 points
0
in reply to: papetoast’s comment on: sarahconstantin’s Shortform
no, sometimes it’s short

sarahconstantin 14 Jan 2026 21:41 UTC
5 points
0
on: sarahconstantin’s Shortform
links 1/14/26: https://roamresearch.com/#/app/srcpublic/page/01-14-2026
- https://www.destroyallsoftware.com/screencasts/catalog/functional-core-imperative-shell
- https://blog.spec.tech/p/spectech-lessons-and-updated-hypotheses-d9f Ben Reinhardt on lessons learned at SpecTech in 2025.
- https://www.corememory.com/p/exclusive-fridge-magnet-medicine-nonfiction-labs-cancer drugs whose binding affinity is magnetosensitive—so you can target to tumors and not other organs

sarahconstantin 11 Jan 2026 16:12 UTC
2 points
0
on: sarahconstantin’s Shortform
links 1/11/26: https://roamresearch.com/#/app/srcpublic/page/01-11-2026
- https://www.science.org/doi/full/10.1126/sciadv.aaw7243 TRPM7 is involved in mechanosensing
  - https://karger.com/cpb/article-abstract/19/1-4/1/207934/Direct-Mechano-Stress-Sensitivity-of-TRPM7-Channel
- https://link.springer.com/article/10.1007/s12032-017-0987-1 TRPM7 is involved in the epithelial-mesenchymal transition in multiple tumor types
  - https://journals.sagepub.com/doi/full/10.3233/CBM-190666
  - https://aacrjournals.org/cancerres/article/72/16/4250/575870/TRPM7-Is-Required-for-Breast-Tumor-Cell
  - https://www.sciencedirect.com/science/article/pii/S0006295223004483
- pharmacologically targeting TRPM7 https://www.nature.com/articles/s41467-023-38362-3
- DRP1 and mitochondrial fission in cancer https://www.sciencedirect.com/science/article/abs/pii/S0304419X25001957
- DNML1 in cancer https://pmc.ncbi.nlm.nih.gov/articles/PMC11540555/
- the anti-apoptotic protein FLIP as a cancer target https://febs.onlinelibrary.wiley.com/doi/pdfdirect/10.1111/febs.14523
  - https://www.tandfonline.com/doi/full/10.2147/JIR.S519885
  - https://www.tandfonline.com/doi/abs/10.1517/14728222.2013.736499
- GUK1 and cancer: https://pmc.ncbi.nlm.nih.gov/articles/PMC12148050/
- ELL, involved in the elongation phase in transcription, in cancer: https://www.nature.com/articles/nrc2915
- SEPHS2 & selenium metabolism in cancer: https://www.sciencedirect.com/science/article/pii/S0006497118585554
  - https://link.springer.com/article/10.1007/s00109-025-02563-8
  - https://www.nature.com/articles/s42255-020-0224-7
  - https://www.nature.com/articles/s41598-019-52718-0
- trial of transferrin receptor drug in cancer: https://ashpublications.org/blood/article/144/Supplement%201/1496/531990/Results-of-a-Phase-1-First-in-Human-Study-of-INA03
  - more on transferrin & cancer: https://www.sciencedirect.com/science/article/abs/pii/S0304416511001826
- NMT1, myristoylation, and cancer: https://www.nature.com/articles/s41598-025-13529-8
  - and phase 1 trial of drug with this target: https://link.springer.com/article/10.1007/s10637-024-01448-w
- ADSL and cancer
  - https://onlinelibrary.wiley.com/doi/pdf/10.1002/hep.31685
- PH5A and cancer: https://www.nature.com/articles/s41598-023-44899-6
  - https://link.springer.com/content/pdf/10.1186/s12885-024-12620-z.pdf
  - https://link.springer.com/content/pdf/10.1186/s12967-022-03821-w.pdf
- YRDC and cancer: https://onlinelibrary.wiley.com/doi/full/10.1002/cnr2.70325
  - https://www.nature.com/articles/s43018-024-00748-7
- https://platform.opentargets.org/ Open Targets Platform, info about many drug targets
- https://depmap.org/portal/ DepMap, cancer dependency database
- https://www.hyperdimensional.co/p/among-the-agents how Dean Ball uses Claude Code
- https://marginalrevolution.com/marginalrevolution/2026/01/the-tyranny-of-the-complainers.html most complaints to government agencies come from a few very active complainers
- https://x.com/mollycantillon/status/2008918474006122936?s=20 account (AI-written?) of how someone uses Claude Code for everything. seems potentially appealing
- https://www.alexkesin.com/p/a-better-proxy-than-the-rat-and-dog foundation models for toxicology
- https://www.asimov.press/p/research-forms metascience
- https://archive.ph/20260108173745/https://www.washingtonpost.com/opinions/2026/01/07/ice-minneapolis-shooting-deportaitons-escalation/ this is scary
- https://backofmind.substack.com/p/the-myth-of-merit-in-the-managerial this is probably the best defense of affirmative action—‘you weren’t that good at picking the best anyway, it’s mostly biases and vibes, so you may as well decide to correct historical inequities’. i think even to the extent this is true you could argue that the policies actually enacted were unwise, but i do think it’s a point worth addressing.

sarahconstantin

FW26 Color Stats

Things that Go Boom

Get­ting Back To It

Biotech Startup Stats

Getting Back To It