--- Log opened Tue Sep 09 00:00:34 2025
01:23 < hprmbridge> .monokhrome> where do fish get their omega-3? Do they make it?
01:24 < fenn> fish eat smaller fish, and those fish eat krill, and krill eat copepods, and copepods eat algae. the algae must grow in cold waters, usually on the underside of sea ice
01:25 < fenn> the only reason it gets to humans in temperate climates is global animal migration
01:25 < fenn> there are also some smaller sources in various algae and plants
01:26 < fenn> i wish people would stop saying "omega-3" when they mean DHA
01:26 < fenn> you're literally playing into the hands of a conspiracy
01:26 < hprmbridge> .monokhrome> shit, global warming will destroy the entire omega-3 food chain. Better get my cell membranes in order before then
01:36 < fenn> this is what gpt-5-mini thinks the human-relevant DHA food web looks like: https://fennetic.net/irc/sankey_dha.html
01:37 < fenn> must have found a dupe bug for small forage fish
01:37 < fenn> a bigger AI would probably give better numbers
01:46 < fenn> better https://fennetic.net/irc/sankey_dha.2.html
02:05 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
02:05 < fenn> "kids in school in china are taught to use comfyui as part of their normal curriculum" lol we're fucked
02:06 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
02:13 < hprmbridge> .monokhrome> I'm never really clear on what stuff is worth teaching in school anymore, and what is just a waste with ai doing it
02:19 < fenn> learning how tools work is always worthwhile
02:20 < fenn> if the AI won't help you, then you can do it yourself
02:20 < fenn> but also this indicates that they are taking AI seriously
02:21 < fenn> in the US people just bitch about how the kids don't care "anymore" (as if they ever did)
05:31 < hprmbridge> kanzure> there are lab mice strains selectively bred for litter size. >20 is the record.
05:32 < hprmbridge> kanzure> double the throughput for science!
05:48 < kanzure> how well does vibevoice work for multiple minutes of blathering with tech jargon and science jargon?
05:49 < kanzure> for speech-to-text i mean.
05:51 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
07:45 < kanzure> MuaddibLLM: tell me about hyperbolic routing or virtual coordinate routing systems like vivaldi 
07:46 < MuaddibLLM> kanzure: Virtual coordinate schemes like Vivaldi learn Euclidean coords from RTTs to predict latency and do greedy neighbor selection, while hyperbolic routing embeds the network in hyperbolic space so scale-free structure enables near-shortest greedy forwarding - both scale with tiny per-node state but can hit triangle-inequality violations, coordinate drift under churn, and occasional local minima.
07:46 < kanzure> meh
07:47 < kanzure> that doesn't really get to the core intuitions
08:23 < kanzure> i don't understand the auditing use case for certificate transparency logs.
08:23 < kanzure> for example, it is claimed "CT data can reveal certificates issued for domains without the domain owner's knowledge" but how would you assess from the CT log whether the domain owner had knowledge or not?
08:23 < MuaddibLLM> kanzure: Think of CT as a public append-only checkbook for certs - auditing is monitors and clients verifying Merkle inclusion and consistency proofs and gossiping STHs so you get alerts if a CA “wrote a check” for your domain you did not authorize and so lying logs that try split views or rewrite history get caught and distrusted.
08:24 < kanzure> one use case might be pubic monitoring for unusual validity periods or weak keys, which might be kind of interesting i guess, but if i found a weak key in the data set, who am i supposed to contact?
08:25 < kanzure> AFAIK nobody is checking the CT logs for me and notifying me as for whether or not my CA has issued a new cert.
08:26 < kanzure> possibly there is some argument to be made that the CT log is useful for retrospective audits after something very broken or very bad happens. and having browsers force or require issuers to post into CT is useful to broadly log bad behavior, even if logging doesn't itself catch or detect bad behavior.
08:40 < kanzure> "CT monitors can prove, efficiently and quickly, that all certificates have been consistently appended to the log." this surely must be completely false... how are you going to prove that /all/ certificates have been appended to the log, if you only have the log and the CA public key and never had operational control over the CA private key?
08:45 < kanzure> if the goal is to detect mis-issued certificates, then arguably it's certificate clients that should submit certs, not signers. if you don't trust your CA/signer then you should not use them.
09:10 < kanzure> what is this and why is it from google? https://crt.sh/?id=19817274945
09:13 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Quit: Avoid fossil fuels and animal products. Have no/fewer children. Protest, elect sane politicians. Invest ecologically.]
09:14 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
09:25 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Remote host closed the connection]
10:15 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
10:29 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 272 seconds]
10:42 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
10:49 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Remote host closed the connection]
11:19 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
11:22 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 256 seconds]
11:22 -!- stipa_ is now known as stipa
11:54 < hprmbridge> kanzure> https://msep.one/ molecular nanotechnology tool released in july
11:55 < hprmbridge> kanzure> oh it's from drexler
12:16 < kanzure> https://github.com/Builditluc/wiki-tui this might be nice if it also has a way to disable the text user interface to dump relevant text to stdout.
13:15 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
13:20 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
13:20 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
13:47 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 272 seconds]
14:11 < andytoshi> kanzure: i think you're supposed to be monitoring the CT log yourself
14:15 < kanzure> do you do that?
14:17 < andytoshi> no, i do not. i am mostly annoyed by the CT logs and wish they weren't there
14:17 < andytoshi> but if i were running a massive org i might feel differently
14:28 < hprmbridge> kanzure> can you elaborate? why are you annoyed?
14:28 < andytoshi> because they mean that if I create a SSL cert for mysecretdomain.wpsoftware.net, the existence of mysecretdomain is public
14:28 < hprmbridge> kanzure> you can do a private cert tho?
14:29 < hprmbridge> kanzure> oh you don't want to install in your browser cert stores.
14:29 < andytoshi> yeah
14:29 < hprmbridge> kanzure> this seems like typical PKI usability annoyances
14:29 < andytoshi> it's fine, the solution to this is to get a wildcard cert
14:29 < andytoshi> which will work for all my secret domains. but then, the security boundary is stretched
15:19 < L29Ah> x.509 PKI is beyond repair
15:19 < L29Ah> lots of actors can bypass it and MITM you
15:21 < L29Ah> including your state and your ISP, so only useful to counter cheap user-side MITMs
15:59 < hprmbridge> kanzure> hyperbolic metric spaces https://arxiv.org/abs/0805.1266
16:04 < hprmbridge> kanzure> well that's weird. i would have assumed the intuition is that the efficiency comes from somewhere else like the address encoding or coordinate calculation does most of the work in a distributed way. this doesn't seem to be what they are saying.
16:12 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Ping timeout: 256 seconds]
17:11 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
17:21 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 272 seconds]
17:43 < kanzure> "Waldeyer-Hartz was the first to estimate the number of oocytes contained in the neonatal human ovary (36,000–50,000) (2). Since then the idea has been accepted that at birth the ovary contains all the oocytes for future life, which means that some oocytes could have a lifespan of 45–50 years (3). Afterward, von Hansemann estimated between 40,000 and 80,000 oocytes in a neonatal ovary (4). ...
17:43 < kanzure> ...Hammar and Hellmann counted 194,283 oocytes in an ovary of three-year-and eight-month-old girl who died of thyroaplasia (5). Häggström counted 419,911 oocytes in both ovaries of a 22-year-old woman (6). It is currently estimated that a woman's ovaries contain between one and two million oocytes at birth and that this number decreases to about 400,000 at puberty when approximately 300–400 ...
17:43 < kanzure> ...oocytes will undergo a maturation process during ovulation. At menopause, approximately 1000 primary follicles remain to maintain endocrine activity and ovulation (7), as it was described by Zhu et al. (1)."
17:44 < kanzure> are there oocytes in those 1000 primary follicles that remain?
17:44 < MuaddibLLM> Yes - the ~1000 at menopause are typically primordial follicles (often mislabeled “primary”), and each contains a single primary oocyte arrested in prophase I, though they’re mostly quiescent and atresia-prone by then.
17:53 <+gnusha_> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=34c9ba55 Bryan Bishop: HLA-g stuff, feto-maternal stem cells, fertility upregulation genetic changes, delayed age of menopause phenotype >> http://diyhpl.us/diyhpluswiki/genetic-modifications/
18:16 < kanzure> a few crazy ideas:
18:17 < kanzure> harvest fetal ovaries from aborted fetuses at week 11-12, and then do in vitro maturation to bring the germinal vesicle cells into matured oocytes. compared to adult follicle pools, this is a replenishble nearly inexhaustible supply of eggs.
18:18 < kanzure> if used for IVF/pregnancy, it's not so bad- you would be using the egg of your daughter and gestating a granddaughter but that's a lot better than alternatives like "no genetically related pregnancy at all" or "human cloning but you have unresolved questions about genomic imprinting problems"
18:20 < kanzure> pregnancy also tolerizes the mother to some of the fetal cells, although not completely. with some genetic engineering it may be possible to generate a fetus that grows a fetal ovary that could then be grafted into the mother and allowed to grow and age. later you might be able to harvest your daughter's eggs from the fetal ovary implanted in your arm? and/or would this delay menopause?
18:21 < fenn> do teratomas ever form ovaries?
18:21 < kanzure> yeah i was looking for that.. the "somatic sex" article is the best place to find that. my search was leaning towards "no" or inconclusive. ovaries generate teratomas pretty frequently though!
18:23 < kanzure> "Somatic sex: on the origin of neoplasms with chromosome counts in uneven ploidy ranges" https://pmc.ncbi.nlm.nih.gov/articles/PMC8373647/
18:25 < kanzure> with enough cell mass and forced electrofusion events maybe you can generate teratomas that have normal-ish development, i dunno
18:26 < fenn> i was thinking OSKM reprogramming
18:26 < kanzure> (by which i mean play the numbers game and most of the teratomas are going to be deeply degenerate)
18:26 < fenn> apply some growth factors that make it prefer to become a gonad
18:27 < fenn> we know the tech tree right
18:27 < fenn> differentiation pathway
18:39 -!- _flood [~flooded@45.87.214.27] has quit [Remote host closed the connection]
18:39 -!- _flood [~flooded@45.87.214.27] has joined #hplusroadmap
18:40 < hprmbridge> kanzure> make what become a gonad?
18:41 < hprmbridge> kanzure> oh, the teratoma?
18:41 -!- _flood [~flooded@45.87.214.27] has quit [Remote host closed the connection]
18:42 -!- _flood [~flooded@45.87.214.27] has joined #hplusroadmap
18:44 < fenn> the pluripotent stem cell
18:44 < fenn> it's not yet a teratoma
18:45 < fenn> maybe this is not how it works
18:45 < hprmbridge> kanzure> otherwise we'd already be growing whole organs for transplants this way
18:45 < fenn> unclear
18:46 < hprmbridge> kanzure> it's often hard to tell if something doesn't exist because bad idea, or doesn't exist because dumb dumbs
18:46 < fenn> "there couldn't possibly be $100 on the sidewalk in front of us" said one economist to the other
18:46 < fenn> you do have to actually figure out what differentiation / growth factors are needed
18:47 < fenn> michael levin probably has something to say about this
18:53 < hprmbridge> kanzure> "The possibility of using eggs from an aborted foetus in IVF treatment has recently arisen after researchers presented work in which they managed to keep alive in the laboratory ovarian tissue taken from second and third trimester aborted foetuses for several weeks. Several ethical..." https://www.tandfonline.com/doi/abs/10.1179/hrge.11.1.127461242080884u
18:54 < hprmbridge> kanzure> i guess this was the bostrom/sparrow thing
18:56 < fenn> bioethicists not banning something? i can't believe my eyes
18:56 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
18:56 < fenn> from 2005, maybe it was before SJWs took over
18:57 < hprmbridge> kanzure> what if we did ancephalic fetus pregnancies with targeted programmed cell death somewhere, and then after the death you get a fresh source of neonatal organs for donation. the germline could also be modified to increase compatibility with certain organ recipients too.
18:58 < hprmbridge> kanzure> er, i mean post birth organ donation
19:00 < fenn> why not just use clones
19:00 < hprmbridge> kanzure> genomic imprinting
19:01 < fenn> harumph
19:01 < hprmbridge> kanzure> we only recently used gene editing tools to find the minimum set of imprinting for mice, and humans have at least 200 known imprinting spans
19:01 < fenn> i have a bunch of notes about that stuff i need to look up references to still
19:02 < fenn> "super-sox" reprogramming https://foresight.org/resource/sergii-velychko-super-sox-unraveling-the-secret-of-mammalian-genesis/
19:03 < hprmbridge> kanzure> "Ex vivo reconstitution of fetal oocyte development in humans and cynomolgus monkeys" https://pmc.ncbi.nlm.nih.gov/articles/PMC9475534/ i think merrick from church lab cited this one in his iPSC meiosis induction paper
19:03 < hprmbridge> kanzure> sergiy was also from church lab.
19:04 < hprmbridge> kanzure> "Initiation of meiosis from human iPSCs under defined conditions through identification of regulatory factors" https://www.science.org/doi/abs/10.1126/sciadv.adu0384
19:04 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
19:04 < fenn> etymology of cynomolgus: "dog-milker"
19:05 < hprmbridge> kanzure> "In vitro reconstitution of epigenetic reprogramming in the human germ line" https://www.nature.com/articles/s41586-024-07526-6 (2024)
19:08 < hprmbridge> kanzure> merrick also suggests co-culture with cells literally named "ovarian support cells" for oocyte in vitro maturation protocols https://link.springer.com/article/10.1007/s10815-024-03143-4
19:09 < hprmbridge> kanzure> these ovarian support cells can also be iPSC derived and still continue to have IVM beneficial effect.
19:09 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
19:11 < hprmbridge> kanzure> "Live offspring produced by mouse oocytes derived from premeiotic fetal germ cells" https://academic.oup.com/biolreprod/article-abstract/75/4/615/2629116?redirectedFrom=fulltext&login=please (2006)
19:12 < fenn> do we really need both kinds of imprinting or just sperm imprinting on half the chromosomes?
19:13 < hprmbridge> kanzure> in the human ex vivo fetal oocyte maturation article above, took them 3 months of culturing to get to primordial follicle stage.
19:13 < hprmbridge> kanzure> i dunno, ask all the people with naturally occurring imprinting disorders
19:14 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 272 seconds]
19:15 < fenn> the way i see it, males have a game theoretic incentive to maximize offspring growth, females have an incentive to limit growth to a reasonable rate, and these competing priorities lead to separate imprinting patterns which end up balancing out
19:15 < hprmbridge> kanzure> here is 1 of the 2 mouse genomic imprinting engineering articles https://diyhpl.us/~bryan/papers2/bio/in-vitro-fertilization/Adult%20bi-paternal%20offspring%20generated%20through%20direct%20modification%20of%20imprinted%20genes%20in%20mammals%20-%202025.pdf don't think i have the other one...
19:15 < fenn> but maybe we can just get away with maximum growth, or reduce the epigenetic imprinting of maximum growth markers after the fact, to balance it out a little
19:16 < fenn> that's ESCs tho
19:16 < fenn> they're already imprinted?
19:16 < hprmbridge> kanzure> here was a search study to try to find regulators of human genomic imprinting https://diyhpl.us/~bryan/papers2/bio/in-vitro-fertilization/Identifying%20regulators%20of%20parental%20imprinting%20by%20CRISPR-Cas9%20screening%20in%20haploid%20human%20embryonic%20stem%20cells%20-%202021.pdf
19:17 < hprmbridge> kanzure> yeah i think the other one published around the same time might be more relevant, sorry
19:20 < fenn> oh "bi-paternal" means a fusion of two sperm nuclei
19:20 < hprmbridge> kanzure> "Fertile androgenetic mice generated by targeted epigenetic editing of imprinting control regions" https://www.pnas.org/doi/abs/10.1073/pnas.2425307122
19:23 < fenn> there's no way i'm ever going to read all this fwiw
19:25 < hprmbridge> kanzure> this is why we have slopmachines
19:25 < fenn> the mouse paper says "20KO" meaning they knocked out 20 genes from one of the sperm? isn't that equivalent to combining a sperm and a naive oocyte?
19:26 < fenn> introducint 20 lethal recessive mutations doesn't sound like a good plan going forward
19:26 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
19:27 < fenn> i liked that paper where they just tethered a demethylase to the crispr guide RNA targeted at the imprinting regions. too bad it may be fake
19:28 < fenn> .t https://www.pnas.org/doi/pdf/10.1073/pnas.2115248119
19:28 < saxo> Just a moment...
19:28 < fenn> "Viable offspring derived from single unfertilized
19:28 < fenn> mammalian oocytes"
19:28 < hprmbridge> kanzure> why have we lost the technology of title bots
19:29 < hprmbridge> kanzure> https://diyhpl.us/~bryan/papers2/bio/in-vitro-fertilization/notes.txt
19:29 < fenn> i can probably rustle up another title bot that works
19:30 < hprmbridge> kanzure> would need to use playwright.js these days :/
19:30 < fenn> that's a lot of notes
19:42 < fenn> is it feasible to target only a single chromosome with cas9/gRNA by sequencing and finding the differences, then creating a gRNA that only matches one of them? gpt-5-mini is going on about PAM-abolishing mismatches and i don't understand
19:43 < fenn> "PAM changes are much more decisive than single mismatches inside the guide because nucleases strictly require a PAM to initiate binding."
19:43 < hprmbridge> kanzure> do you mean chromosomal inactivation
19:44 < fenn> i mean using a cas9-tethered methylase/demethylase targeted to one of the chromosomes, so we only imprint one chromosome
19:44 < fenn> for cloning. instead of imprinting one naive gamete and then also maybe the other naive gamete
19:45 < fenn> one chromosome at a time
19:46 < fenn> if you just used the sequence in common to both chromosomes you'd get the same imprinting pattern on both chromosomes
19:48 < hprmbridge> kanzure> what about CRISPR microbead and you hold it next to the chromosome when it separates? or is that wrong phase for editing.
19:48 < fenn> i think it would be fine in the first few divisions
19:48 < fenn> i don't understand how this scheme is supposed to work tho
19:49 < fenn> maybe we can just half-imprint both chromosomes and this is needless premature optimization
19:54 < fenn> oh lol the "one chromosome at a time" thing is what they did: "We designed single-guide RNAs (sgRNAs) with a protospacer adjacent motif (PAM) region that matches one allele but not the other."
19:55 < fenn> it would be a lot less work if you didn't have to do that
19:55 < fenn> ugh reading is so hard
19:57 < hprmbridge> kanzure> "Efficient introduction of specific homozygous and heterozygous mutations using CRISPR/Cas9" https://www.nature.com/articles/nature17664
19:59 < hprmbridge> kanzure> or https://pmc.ncbi.nlm.nih.gov/articles/PMC9584603/ idk
21:32 -!- drmeister [sid45147@id-45147.ilkley.irccloud.com] has quit [Ping timeout: 244 seconds]
21:33 -!- potatope [sid139423@id-139423.lymington.irccloud.com] has quit [Ping timeout: 244 seconds]
21:33 -!- drmeister [sid45147@id-45147.ilkley.irccloud.com] has joined #hplusroadmap
21:34 -!- strages [sid11297@id-11297.helmsley.irccloud.com] has quit [Ping timeout: 244 seconds]
21:35 -!- strages [sid11297@id-11297.helmsley.irccloud.com] has joined #hplusroadmap
21:35 -!- potatope [sid139423@id-139423.lymington.irccloud.com] has joined #hplusroadmap
22:09 < fenn> some weird UFO footage shared in congress https://inv.nadeko.net/watch?v=L_R81AFxwzE
22:43 < jrayhawk> why are you linking to a video of a balloon getting blown up
22:49 < jrayhawk> why is your schizophrenia character arc so boring
22:52 < fenn> i'm sorry it offends you
22:53 < fenn> it doesn't look like a balloon to me
22:54 < jrayhawk> it is literally 100% consistent with a balloon in every possible way
22:55 < hprmbridge> kanzure> is this the fish oils
22:56 < jrayhawk> i would expect that to go the other way, if anything
22:59 < fenn> alright maybe it's a balloon
23:01 < jrayhawk> i still say you should've gone with a messiah complex. way cooler.
23:01 < fenn> my schizophrenia character arc was in 2005-2008
23:02 < fenn> and it was more boring
23:04 < fenn> please ask the united states congress to schedule their UFO witness hearings at a time more convenient for you
23:04 < jrayhawk> what, so i can testify the obvious?
23:07 < jrayhawk> UFO idiots don't care. even when the U.S. military explicitly labels a video "GIMBAL" they say "OH WHOA MAN LOOK AT THAT ERRATIC MOVEMENT DRIVEN BY ALIEN PROPULSION TECHNOLOGY" because a camera rotated
23:07 < fenn> yeah it's a problem
23:07 < fenn> and the never ending parade of lens flares
23:16 < hprmbridge> .monokhrome> yeah it's telling that the only footage we ever see is of out of focus blobs
23:16 < hprmbridge> .monokhrome> what happened to those mystery balloons that were shot down over north america in early 2023? They just vanished from the news media
23:21 < fenn> and the jersey drones
23:21 < jrayhawk> https://en.wikipedia.org/wiki/2023_Chinese_balloon_incident
23:21 < fenn> the balloon stuff was never particularly interesting to start with
23:33 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
23:40 < hprmbridge> .monokhrome> chinese had balloons over north america for god knows how long, us air force shoots them down then say it's of no interest
23:40 < hprmbridge> .monokhrome> and then there's those drones that kept overflying that military base. The whole subject is strange
23:51 < superkuh> Spy balloons are not very interesting. But high altitude balloons are very interesting because they're such low complexity and cost. 
--- Log closed Wed Sep 10 00:00:35 2025