--- Log opened Sun Aug 18 00:00:25 2024
00:05 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has quit [Remote host closed the connection]
00:05 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has joined #hplusroadmap
00:47 < hprmbridge> kanzure> ?
00:51 < fenn> dai = big, xiao = small
01:37 < kitaleth> You can just shrink proteins?
01:46 < hprmbridge> kanzure> not without a lot of testing to ensure function wasn't disabled...
01:49 < hprmbridge> kanzure> the cult of the light-controlled polymerase continues to survive rent free in people's hearts https://x.com/LoganTCollins/status/1824979693176496314
01:53 < kitaleth> docl: re overspecialising: compared to what alternative?
02:07 < kitaleth> fenn: How bottlenecked do you think we are in nanotech by lack of good simulation tooling?
02:07 < kitaleth> Quantum DFT seems like it would be too computationally intensitve, but what about classical approximations?
02:12 < kitaleth> Oh, you weren't lying about "non-bio manufacturing just works"
02:41 < fenn> hey i started that cult
02:42 < fenn> ideally you'd have 2 "bits" (light directed control channels) that get mechanically multiplexed in the protein
02:42 < fenn> unpacked
02:42 < fenn> i meant demultiplexed
02:42 < fenn> it's not right. anyway, either you get it or you don't
02:43 < fenn> the clock channel / ratchet / deprotection mechanism is essential
02:45 < fenn> kitaleth: i think it's foolish to build increasingly precise simulations without the real world data to verify them. once you get data you can build a model
02:46 < fenn> simulating millions of molecules from scratch is not gonna work
02:47 < fenn> approximations are needed but "classical" is probably not how to do it
02:47 < kitaleth> fenn: I thought we had physics sorted out enough to be able to simulate candidates?
02:47 < kitaleth> i don't get what makes it computationally complex: do the laws of physics not act as a way to prune the search tree?
02:49 < fenn> i think it's just because every particle affects every other particle, and you can't treat everything as billiard balls because the resonance structures are super important
02:52 < fenn> so for 1 million atoms, for each timestep you have (1 million * (6 electrons + 1 carbon nucleus))^2 = 49 trillion vector calculations, just for the classical model
02:53 < kitaleth> Do you need to model 1M atoms for nanotech stuff that we can build now?
02:53 < fenn> heavier atoms like gold have a lot more electrons, although you can probably ignore the inner orbitals
02:54 < fenn> when a STM probe tip comes in to touch a molecule on a surface, forces are transmitted through the whole structure, so you have to model a big chunk of the tip and the surface, even if you're only poking at a small molecule
02:55 < fenn> probably not 1 million, on
02:55 < fenn> no
02:56 < kitaleth> Even if you did need 1M atoms, 49T vector calcs seems.. fine? RX 7600 is at 21.5 TFLOPS
02:56 < kitaleth> (yes, you *do* have numerics issues, but since we know the domain, I presume we can fix the range beforehand/use tools like Herbie)
02:56 < fenn> a typical biological protein will contain ~50 kDa = ~4000 atoms
02:56 < kitaleth> I get what you mean about the STM probe, though
02:57 < kitaleth> It doesn't seem *too* intractable to model in terms of hardware components, just a matter of not enough good software?
02:57 < kitaleth> s/components/requirements
02:58 < kitaleth> So can a typical biological protein not be modeled ab initio?
02:58 < fenn> i'm not sure, computers are pretty fast these days
02:59 < kitaleth> indeed, yes
02:59 < kitaleth> that's the thing that surprises me: no one seems to have tried yet>
02:59 < kitaleth> ?
03:00 < kitaleth> (seems easier to try building a sim platform based on known laws, validate on existing measurements, and then see if it works?)
03:03 < fenn> i asked someone who works with supercomputers for chemistry simulations, will let you know if there's a good response
03:04 < kitaleth> Thanks!
03:50 < fenn> https://old.reddit.com/r/StableDiffusion/comments/1euqwhr/realism_comparison_amateur_photography_lora_flux/ now with less pizzaz!
03:50 < fenn> z
04:31 -!- galambo [galambo@user/galambo] has joined #hplusroadmap
04:40 < kitaleth> https://rosen.cbe.princeton.edu/
06:26 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 260 seconds]
06:36 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has quit [Remote host closed the connection]
06:36 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has joined #hplusroadmap
06:45 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
06:46 -!- TMA [tma@twin.jikos.cz] has quit [Ping timeout: 252 seconds]
06:49 < hprmbridge> kanzure> https://caseyhandmer.wordpress.com/2024/08/18/antimatter-is-the-best-post-chemical-rocket-propulsion-system/
06:53 -!- TMA [tma@twin.jikos.cz] has joined #hplusroadmap
07:19 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
07:19 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
07:27 < kitaleth> You can produce antimatter via solar power?
07:48 -!- galambo [galambo@user/galambo] has left #hplusroadmap [Leaving]
07:56 -!- TMA [tma@twin.jikos.cz] has quit [Ping timeout: 260 seconds]
08:02 < fenn> antimatter forms naturally in atmospheres by cosmic ray impact
08:03 < fenn> currently it's extremely inefficient to produce as an energy storage medium
09:05 -!- TMA [tma@twin.jikos.cz] has joined #hplusroadmap
09:06 < docl> you can potentially use sunlight to produce the equivalent of cosmic rays at scale with the aerographite sail-shell trick
09:10 < docl> kitaleth: fundamentally-cheap approaches have surprising capabilities sometimes. like the $2 kit for cell-free gene editing. someone should iterate on that to add more capabilities while keeping the cost low
09:10 < docl> .t https://www.nature.com/articles/s41467-024-50767-2
09:10 < saxo> A frugal CRISPR kit for equitable and accessible education in gene editing and synthetic biology | Nature Communications
10:02 < docl> kitaleth: for cell free lysate and similar bioproducts, I'm contemplating a system usiung kombucha as a growth medium. It forms a cellulosic biomat over the top and uses low pH to exclude most bacteria, but the microbiome is in constant flux, so it might be favorable for guided evolution as well if you do it right. The usual tea in a jar system isn't economical enough, but I've had some success growing 
10:02 < docl> it on paper and perlite. we're talking processes not far removed from composting here
10:45 < docl> re antimatter catalyzed fission-fusion https://gnusha.org/logs/2023-12-18.log
11:10 < hprmbridge> docl> kombucha scoby biomat grows on/through paper and around perlite granules just fine (nothing surprising) https://cdn.discordapp.com/attachments/1064664282450628710/1274792653669531668/image.png?ex=66c38aad&is=66c2392d&hm=f1efd8f91ac40f29ceab2173afd65da46b0647f7d384ad4504b94d41f6606e0f&
11:14 < hprmbridge> alonzoc> That white looks like mold tho
11:14 < hprmbridge> alonzoc> Oh wait
11:15 < hprmbridge> alonzoc> those are perlite are they?
11:16 < hprmbridge> kanzure> https://www.overcomingbias.com/p/how-far-might-we-fall
11:21 < kitaleth> docl: for mammalian cells?
11:31 < docl> yeah those are just perlite granules. am considering using perlite-bioleather aggregate for wall coverings or maybe winter wear
11:35 < docl> cheaper ways to grow actual mammalian cells? I'll have to think about it further... they obviously can't handle the low pH of kombucha. maybe you could evolve hela cells to do so?
11:50 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Quit: Konversation terminated!]
11:50 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
12:28 < hprmbridge> alonzoc> docl: I previously mentioned thought emporiums DIY media project.  It's got some interesting results on drinks that can be used to dilute DMEM, one works upto like 90% dilution.
12:28 < hprmbridge> alonzoc> Mammalian cells also need incubation and are trickier to grow in general so you'd really need mammalian cells
12:30 < hprmbridge> alonzoc> I haven't done to much bio lab work but my bio friends from uni complained endlessly about how picky various mammalian cell types were about growing
12:48 -!- Gooberpatrol_66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
12:49 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Ping timeout: 276 seconds]
13:50 -!- Croran [~Croran@user/Croran] has quit [Ping timeout: 258 seconds]
13:52 -!- Croran [~Croran@user/Croran] has joined #hplusroadmap
13:58 -!- abetusk [~abetusk@syn-024-059-059-088.res.spectrum.com] has quit [Ping timeout: 260 seconds]
13:58 -!- abetusk [~abetusk@syn-024-059-059-088.res.spectrum.com] has joined #hplusroadmap
14:33 < docl> I wonder if there's a way to cheaply test how close to mammalian cell compatible medium the brew is and translate to an electrical signal. then e.g. activate a UV light to kill the microbes in the solution, but keep reducing the intensity the closer it gets to optimal
17:01 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has quit [Ping timeout: 272 seconds]
20:51 -!- mxz__ [~mxz@user/mxz] has joined #hplusroadmap
20:51 -!- mxz [~mxz@user/mxz] has quit [Ping timeout: 248 seconds]
20:51 -!- mxz_ [~mxz@user/mxz] has quit [Ping timeout: 252 seconds]
20:51 -!- mxz__ is now known as mxz
22:02 < docl> .t https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9605694/
22:02 < saxo> Better under stress: Improving bacterial cellulose production by Komagataeibacter xylinus K2G30 (UMCC 2756) using adaptive laboratory evolution - PMC
22:06 < docl> I wonder if the bacterial cellulose nanostructure can be tweaked by directed evolution?
22:10 < docl> to get it to evolve in a mammalian cell compatible manner, maybe you could have two layers separated by a semipermeable membrane or thin ceramic. measure where the cells do the best, then deliver nutrients to the microbes immediately on the other side of the membrane.
22:13 < docl> for assaying the mammalian cells, you could give them a pigment or maybe GFP, and check with a camera. or you could use something to measure Na+/K+ pumping action as a sign of cell health, whatever's cheap/easy
22:19 < docl> so you can have two layers, the bottom is the cell culture on transparent glass, with a dead scoby as a scaffold to grow on. there's a buffered solution that flows slowly. on top of that is a porous layer that lets the kombucha-like brew through in small amounts. then you have a live scoby on top of that, with barely the nutrients it needs to stay alive. a camera/sensor moves along the bottom to check 
22:19 < docl> where the cells are doing well, and a dispenser of nutrient solution moves along top and dribbles the sugar water wherever it seems to do the best.
22:58 -!- mxz_ [~mxz@user/mxz] has joined #hplusroadmap
23:27 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
23:27 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
23:58 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has joined #hplusroadmap
--- Log closed Mon Aug 19 00:00:26 2024