--- Log opened Mon Sep 29 00:00:53 2025
01:46 < hprmbridge> kanzure> "<3.1% of the genome is transcribed per cell" https://www.cell.com/cell/abstract/S0092-8674(25)01034-7
01:51 < fenn> i always felt like i was tripping on psychedelics after eating at the north-indian/pakistani restaurant "pakwan" in SF
01:52 < fenn> it could have actually been drugs, but maybe there is some spice that does this
01:52 < hprmbridge> kanzure> mosaic errors of immunity https://cdn.discordapp.com/attachments/1064664282450628710/1422144060713599017/image0.jpg?ex=68db9a5e&is=68da48de&hm=adfa91d1a0814cba379bfd1bef52aaa9b6c8bfdaed22ee7c80b9bfcf1467e594&
01:53 < fenn> (others who had eaten the same food didn't experience the mind-altering effects)
01:54 < fenn> with mosaic errors of immunity, half your immune system doesn't work right?
01:55 < fenn> it's not like each branch of differentiation has a single progenitor cell
01:55 < hprmbridge> kanzure> from https://www.cell.com/trends/immunology/fulltext/S1471-4906(25)00222-4
01:56  * fenn shrugs
01:57 < hprmbridge> kanzure> well anyway, lowering mutational burden seems like a reasonable ideA
01:58 < fenn> the treatment would have to be very personalized, for de-novo mutations
01:58 < fenn> could you even get a vector to the bone marrow stem cells that are presumably churning out useless leukocytes
01:58 < fenn> first you have to figure out which bone the mutation is in...
01:59 < hprmbridge> kanzure> many somatic mutations will be caught by immune system due to expression of unexpected MHC etc
01:59 < fenn> not if they occur during embryogenesis
02:00 < hprmbridge> kanzure> checking mosaic mutational load before birth seems like a prudent thing to look into?
02:01 < hprmbridge> kanzure> or you can identify top bad mutations and preprogram immune system to reject.
02:02 < fenn> there's like a billion of those
02:02 < hprmbridge> kanzure> this is hard ok
02:03 < hprmbridge> kanzure> okay let's just fix the problem by encoding checksums with error correction capability below some error threshold you can recover the data.
02:04 < hprmbridge> kanzure> why did anyone talk about reversible computing at all, it looks like it's completely wrong? its not adiabatic either.
02:05 < fenn> what do you mean it's wrong?
02:06 < MuaddibLLM> [gpt-5] HSCs circulate so you don't need to pinpoint "which bone" - recent advances in targeted lipid nanoparticles now enable direct in vivo HSC editing in bone marrow (https://www.nature.com/articles/s41586-025-09070-3), while prenatal mosaic screening via cfDNA has limited sensitivity for low-frequency clones, immune preprogramming against millions of neoantigens doesn't scale, biology uses error detection (DDR/mismatch 
02:06 < MuaddibLLM> repair) not literal ECC, and reversible computing isn't wrong - logically reversible operations can theoretically beat Landauer's principle though practical adiabatic circuits remain lossy (https://spectrum.ieee.org/reversible-computing).
02:06 < fenn> first we need superconducting chips, because most of the energy dissipation is in the wires (interconnect layers)
02:06 < hprmbridge> kanzure> https://diyhpl.us/~bryan/papers2/physics/The%20thermodynamics%20of%20computation:%20A%20contradiction%20-%20Porod.pdf
03:39 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Ping timeout: 264 seconds]
04:46 < TMA> fenn: reading the "bunch of stuff you will never read but that proves your ideology wrong" so far it is "this good outcome is possible therefore it will occur" I have no problem with good outcomes being possible (as "there _exists_ a chain of decisions that will lead to that outcome") I have problem with the second part (because it is not true that "for all chains of decisions that chain will lead to a good outcome") not even most decisions lead to good outcom
04:52 < fenn> most people see the upside as incredibly small. you yourself said "hundreds"
05:20 < TMA> hundreds are enough for the jealousy to kick in
06:27 < hprmbridge> kanzure> do you mean envy?
07:10 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
07:11 < TMA> yes.
07:16 < TMA> but according to merriam webster jealous is a synonym of envious in one of its senses; https://www.merriam-webster.com/dictionary/jealous 
07:16 < TMA> https://www.merriam-webster.com/dictionary/jealousy even discusses the overlap of jealousy vs. envy
07:16 < TMA> but envy would be probably better in this case
07:28 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=8b7794fa Bryan Bishop: use cell-cell genome consensus validation for DNA damage repair >> http://diyhpl.us/diyhpluswiki/genetic-modifications/
07:29 < kanzure> admittedly that's deeply stretching the imagination far past the point of absurdity, but if it's ever engineered then it will ~permanently solve the problem of DNA damage, mutation, and repair.
07:31 < kanzure> several years ago i posted about "ECDSA encryption proteins" and DNA error correction codes built into the genome: https://groups.google.com/g/enzymaticsynthesis/c/WvDidIldm7c/m/q7aQTnmkAAAJ
07:32 < kanzure> "magic codons: codons: we could lengthen codons to be more than 3 letters long, with some tolerance for errors, these codons might be more practical for error-prone DNA synthesis. We might also be able to make polymerases that convert from X bp per codon to 3 bp per codon. This way, if you get one of the base pairs wrong in the codon that you're writing, you will still be able to get a fuzzy ...
07:32 < kanzure> ...match in many cases, to the correct tRNA synthetase and amino acid.
07:32 < kanzure> "checksum enzymes: in the future it would be nice to have checksums and hash functions in tRNA synthetases or in polymerases or DNA repair enzymes to check that the DNA molecule is still correct based on the checksums."
07:34 < kanzure> "ribosome repurposing: maybe we could convert a ribosome to construct DNA molecules or oligos or RNA or something; the tRNA synthetase machinery already translates from codons (3 bp) to a single amino acid. This is similar to mechanisms required for information decoding or error correction codes." (apparently i forgot about this idea of mine and independently self-reinvented this? ugh)
07:36 < kanzure> https://gnusha.org/logs/2017-05-12.log https://gnusha.org/logs/2017-05-16.log
07:38 < kanzure> there are a few more elaborations of this idea on the wiki now but i wanted to at least include this in the irc logs for future reference too:
07:38 < kanzure> "Even more advanced would be a cell network consensus-based DNA error detection and DNA error correction system. This consensus-based DNA validation system would use engineered intercellular communication networks where individual cells randomly sample genomic fragments and query neighboring cells through engineered quorum sensing molecules, CRISPR-based fragment querying to neighboring cells, ...
07:39 < kanzure> ...large natural competence pores (or other kinds of transmembrane pores), exosomes, and synthetic G-protein coupled receptor signaling pathways to establish distributed consensus about sequence fidelity. The approach exploits the statistical improbability that identical mutations occur independently across multiple adjacent cells, using Byzantine fault-tolerant algorithms implemented through ...
07:39 < kanzure> ...genetic circuits to distinguish between inherited damage (present across lineages) and spontaneous mutations (isolated to individual cells), ultimately triggering targeted DNA repair mechanisms via fragment import, controlled apoptosis, or lineage replacement when consensus indicates genomic corruption has occurred."
07:39 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=9b3de346 Bryan Bishop: add links for consensus-based genomic integrity strategy stuff >> http://diyhpl.us/diyhpluswiki/genetic-modifications/
07:56 < kanzure> i wonder if the species with ridiculously high numbers of chromosomes have any sort of special DNA repair mechanisms for relying on other redundant chromosome copies, and then using chromosome elimination to  or consensus to fix damage.
07:58 < kanzure> DNA repair mechanisms like homologous recombination use sister chromatids for templates but maybe you can get some advantage in high polyploidy somehow.
08:09 < kanzure> some weird polyploidy stuff: https://pmc.ncbi.nlm.nih.gov/articles/PMC526024/ including hybrid plants from cut stem grafts that are hybridized at a nuclear genomic level? "allopolyploidization [often follows] a nonrandom, reproducible pattern characterized by preferential elimination of sequences from one of the parental genomes."
08:10 < kanzure> "Spring (1997) provided a convincing argument that humans and other vertebrates may be ancient polyploids; a survey of more than 50 gene families from aldolases to zinc finger transcription factors showed that usually a single invertebrate gene corresponds to up to four equally related human genes on different chromosomes."
08:13 < kanzure> "The reasons for this nonrandom pattern of sequence elimination are unknown." bah
08:17 < kanzure> and https://gnusha.org/logs/2025-07-25.log 
08:17 < kanzure> "CRISPR/Cas9-mediated targeted chromosome elimination" https://genomebiology.biomedcentral.com/articles/10.1186/s13059-017-1354-4   from https://gnusha.org/logs/2017-11-28.log
08:21 < kanzure>  "uniparental genome elimination" was the thing of relevance from 2025-07-25.log and other kinetochore stuff (the microtubule attachment sites that mediate chromosome inheritance).
08:22 < kanzure> "Unrepaired DNA damage facilitates elimination of uniparental chromosomes in interspecific hybrid cells" https://pmc.ncbi.nlm.nih.gov/articles/PMC4049971/ ".. human chromosomes are always eliminated during clone formation when human cells are fused with mouse cells. However, the underlying mechanisms are still elusive."
08:58 < kanzure> meanwhile:
08:59 < kanzure> "But uh, people have followed the paradigm of multiple rounds of directed evolution with sequencing in between the rounds, and it turns out that sequencing after the first round can predict outcomes 20 rounds later. There are some papers on this. Very faint signals that you need to amplify across 20 rounds. You can pick that up after the first round of selection." (from 2017)
10:04 < L29Ah> .t https://www.psypost.org/lifelong-social-connections-may-slow-biological-aging-and-reduce-inflammation/
10:04 < saxo> Lifelong social connections may slow biological aging and reduce inflammation
10:11 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
10:11 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
11:25 -!- hprmbridge [~hprmbridg@user/fenn/bot/fennbots] has quit [Ping timeout: 264 seconds]
11:26 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
11:29 -!- hprmbridge [~hprmbridg@user/fenn/bot/fennbots] has joined #hplusroadmap
11:30 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 264 seconds]
11:30 -!- stipa_ is now known as stipa
13:13 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Quit: Konversation terminated!]
13:21 < hprmbridge> .monokhrome> Or maybe sickly people tend to become more introverted and do less social stuff
14:20 < hprmbridge> jerlendds> https://diffuse.one/p/d1-009
14:20 < hprmbridge> jerlendds> > abstract: The people want a virtual cell. So let's simulate an entire human cell atom-by-atom for 24 hours with molecular dynamics. I estimate it will take 2×10382×1038 FLOPs and 200 TW of electricity. Based on historic frontier simulations, we should be able to reach this milestone in 2074. I based on this analysis of about 650 historic molecular dynamics simulation papers and some empirical
14:20 < hprmbridge> jerlendds> measurements of FLOPs required. See details below.
15:40 < hprmbridge> Eli> Deepmind says they will do it before 2030. Guess we won’t have long to see if they’re right …
15:58 -!- flyback [~flyback@2601:540:c700:2380:ff11:666d:5b34:c829] has quit [Ping timeout: 250 seconds]
16:11 -!- flyback [~flyback@2601:540:c700:2380:d4f1:bf17:a77d:bd7a] has joined #hplusroadmap
16:19 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
18:24 < hprmbridge> Eli> New CVD biomarker dropped. We kind of already knew we needed to get our CRP down to extend healthspan. And I think we might even get an IL6 inhibitor. https://cdn.discordapp.com/attachments/1064664282450628710/1422393736863682670/IMG_2188.jpg?ex=68dc82e5&is=68db3165&hm=aa5cd968e6912c5487e632a529dadeaae20125608721e97f90f47af0cdb56204&
18:25 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has joined #hplusroadmap
18:25 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has quit [Remote host closed the connection]
18:26 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has joined #hplusroadmap
18:26 < fenn> instead of error correction the self could simply self destruct when the (double stranded) DNA checksum doesn't match
18:27 < fenn> s/self/cell/
18:28 < fenn> this works because you can just goose the replication rate and there are no accumulated mutations from excess replication, and resources from apototic cells are reused
18:28 < fenn> trying to compute the correct sequence from a checksum sounds too complicated
18:29 < fenn> mutations in the checksum will also self-destruct because it won't match
18:33 < fenn> might want to protect brain tissue from this pathway
18:35 < hprmbridge> kanzure> apoptosis already occurs under certain situations of extreme damage
18:35 < hprmbridge> kanzure> ... i think. such as cancer.
18:39 < hprmbridge> kanzure> yeah there are some checksums that can tolerate errors and still tell you how to fix it.
18:40 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
18:41 < hprmbridge> kanzure> https://en.wikipedia.org/wiki/BCH_code
18:42 < hprmbridge> kanzure> https://en.wikipedia.org/wiki/Error_correction_code
18:43 < fenn> hmm. chatgpt thinks 0.2% of cells will survive per division
18:43 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 264 seconds]
18:43 -!- stipa_ is now known as stipa
18:44 < hprmbridge> kanzure> yes the somatic mutation rate per mitosis is non-zero just from enzymatic error
18:44 < hprmbridge> kanzure> although maybe we can engineer the enzymes to have substantially lower error rate.
18:45 < fenn> if we only detect mutations in coding regions, 97% of cells survive per division
18:46 < fenn> that sounds tolerable
18:47 < fenn> you could calculate the important non-coding regions when you're placing these checksums
18:50 < fenn> it says "decades" between neural mutations
18:52 < fenn> also we can upregulate DNA repair enzymes
18:53 < fenn> i think there's a few orders of magnitude low hanging fruit there
18:53 < fenn> i really like the idea of whole-genome correction, not just coding regions, because it scales indefinitely
18:54 < fenn> if you only correct coding regions, eventually you will get weird regulatory dysfunction due to the accumulated mutational load, and then what
18:54 < fenn> we can also turn the self-destruct code on and off so it happens at convenient times (when well fed, not sick, not during a war, etc.)
18:56 < fenn> actually fasting triggers apoptosis normally, not sure what the considerations are there
19:01 < hprmbridge> kanzure> "A counterpossible is a counterfactual with an impossible antecedent" what
19:02 < hprmbridge> kanzure> there's some very good DNA repair upgrades on the wiki page at the top.
19:02 < fenn> yeah we should probably be doing this aside from the ECC scheme
19:04 < hprmbridge> kanzure> "Why Zeno’s paradoxes of motion are actually about immobility" oh no i'm in the weird part of the internets again.
19:04 < fenn> intermittent fasting can increase cell turnover / apoptosis and DNA repair
19:05 < hprmbridge> kanzure> https://diyhpl.us/wiki/genetic-modifications/#index2h1
19:08 < fenn> Ecoli the biologist, now that's what i call nominative determinism
19:09 < fenn> oh i see, the bacterium was being credited for its contributions to science, nevermind
19:12 < fenn> a lot of "radiation resistance" infrastructure is actually oxidative stress resistance
20:12 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has quit [Remote host closed the connection]
20:28 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has joined #hplusroadmap
20:54 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Remote host closed the connection]
20:59 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
21:41 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
21:41 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
22:26 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has quit [Ping timeout: 248 seconds]
22:38 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has joined #hplusroadmap
23:01 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has quit [Ping timeout: 256 seconds]
--- Log closed Tue Sep 30 00:00:54 2025