--- Log opened Thu Mar 06 00:00:36 2025
00:02 -!- WizJin__ [~Wizzy@2402:a00:184:8354:9dd9:e60b:e3eb:fe0b] has joined #hplusroadmap
00:05 -!- WizJin_ [~Wizzy@2402:a00:184:8354:9dd9:e60b:e3eb:fe0b] has quit [Ping timeout: 268 seconds]
01:50 -!- gl00ten2 [~gl00ten@193.147.150.204] has joined #hplusroadmap
02:57 -!- gl00ten2 [~gl00ten@193.147.150.204] has quit [Read error: Connection reset by peer]
02:57 -!- gl00ten [~gl00ten@193.147.150.204] has joined #hplusroadmap
03:02 -!- WizJin [~Wizzy@2402:a00:184:8354:9dd9:e60b:e3eb:fe0b] has joined #hplusroadmap
03:02 -!- WizJin [~Wizzy@user/WizJin] has changed host
03:04 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Remote host closed the connection]
03:04 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
03:05 -!- WizJin__ [~Wizzy@2402:a00:184:8354:9dd9:e60b:e3eb:fe0b] has quit [Ping timeout: 248 seconds]
03:30 -!- Netsplit *.net <-> *.split quits: justanotheruser
03:31 -!- gl00ten [~gl00ten@193.147.150.204] has quit [Remote host closed the connection]
03:53 -!- gl00ten [~gl00ten@193.147.150.204] has joined #hplusroadmap
04:14 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Remote host closed the connection]
04:14 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
04:46 -!- rafspiny [~raffaele@a120210.upc-a.chello.nl] has joined #hplusroadmap
05:02 -!- WizJin_ [~Wizzy@2402:a00:184:8354:9dd9:e60b:e3eb:fe0b] has joined #hplusroadmap
05:05 -!- WizJin [~Wizzy@user/WizJin] has quit [Ping timeout: 245 seconds]
05:34 -!- gl00ten [~gl00ten@193.147.150.204] has quit [Remote host closed the connection]
05:36 -!- gl00ten [~gl00ten@193.147.150.204] has joined #hplusroadmap
05:44 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
05:45 < kanzure> wall-mounted 3d printed slice of a babbage machine, doesn't have to be the full machine, just a portion of the architecture
05:45 < kanzure> with manual turn wheel to manually power the gears or something. that would be nice.
10:00 < RangerMauve> fenn: Reminds me of the "source code to the universe" in the Fine Structure book by qntm. https://qntm.org/structure
10:14 -!- andytoshi [~apoelstra@user/andytoshi] has quit [Ping timeout: 245 seconds]
10:16 -!- andytoshi [~apoelstra@user/andytoshi] has joined #hplusroadmap
10:21 -!- WizJin_ [~Wizzy@2402:a00:184:8354:9dd9:e60b:e3eb:fe0b] has quit [Quit: Leaving]
10:22 -!- WizJin [~Wizzy@user/WizJin] has joined #hplusroadmap
11:04 -!- Redendimon [~Redendimo@49.149.138.57] has joined #hplusroadmap
11:15 < fenn> deepseek-r1-equivalent performance on a single consumer GPU: https://qwenlm.github.io/blog/qwq-32b/
11:17 < fenn> notably missing knowledge-heavy benchmarks
11:48 < fenn> https://www.extremetech.com/computing/nvidia-might-re-launch-rtx-4090-with-up-to-96gb-of-vram
11:51 < fenn> hmm nm bad journalism, that's a hacked card in the "leak" tweet
11:59 -!- Redendimon [~Redendimo@49.149.138.57] has quit [Quit: Leaving]
12:06 < kanzure> .title https://www.youtube.com/watch?v=Ip6kabvaIpk
12:06 < saxo> Jeremy Barton | A Path to Atomically Precise Manufacturing @ Paths to Progress
12:07 < hprmbridge> tylorsama> Hi everyone. That's me.
12:08 < kanzure> hello
12:09 < kanzure> anyone who discusses molecular nanotechnology should be required to experiment with AFM/SEM
12:09 < kanzure> i think the metric that matters is total time between experiments, and then number of experiments per ~day
12:09 < hprmbridge> Jeremy> Yep. The suffering builds character.
12:15 < hprmbridge> Jeremy> The biggest challenge with 'experiments per day' is that if you do it right the SPM often isn't your bottleneck. Setting up the experiment relies on lots of pre-work. Getting those lined up so that the SPM is your bottleneck requires a lot of planning and effort, but can be done.
12:16 < kanzure> sounds fine to me as long as the bottlenecks can be identified in mortal bounded time
12:18 < hprmbridge> Jeremy> Amen to that. Getting to full on self-replicating inorganic nanofactories likely requires solving lots of problems we can't currently anticipate. I'm happy to dive deeper on the projected challenges if there is interest.
12:20 < kanzure> what are your thoughts about theory/simulation people in this area, versus just trying stuff out?
12:20 < kanzure> reaction chemistry sometimes only works under very specific conditions that you won't know about except through theory or simulation
12:21 < kanzure> and yeah you can simulate a few electron transfer reactions maybe but not anything large, and not quickly
12:21 < hprmbridge> Jeremy> You benefit by having both. Actually identifying what is going on, what you are seeing, is really hard with these methods. Getting an accurate simulation of your experimental observalbe and how that correlates with your atomic structure is extremely valuable when you are trying stuff out. Similarly, identifying viable chemistry is a lot cheaper in simulation, given the cost and time of
12:21 < hprmbridge> Jeremy> synthesizing these molecules.
12:22 < kanzure> a bench chemist can run circles around simulation though
12:23 < hprmbridge> Jeremy> Not always, particularly if you are trying to build something hard. It can take many months to synthesize particularly hard molecules.
12:23 < kanzure> been there, done that
12:24 < hprmbridge> Jeremy> Then getting those molecules into a usable orientation and density on a surface can be really slow if you are using SPM as your sole surface chemistry technique. So interpreting other techniques also benefits from simulation to help interpret the data.
12:26 < hprmbridge> Jeremy> There exist molecular tool designs and principles that enable mechanosynthesis. That cuts down on a lot of the simulation burden, as you have something to work from, rather than trying to find a poorly designed needle in a search-space. Someday the validation of those principles should be public... 😕
12:27 < hprmbridge> Jeremy> Someday soon, the rumor runs.
12:33 < hprmbridge> somewhereville> If I knew where you lived, I'd ask you to step outside.
12:37 < kanzure> inside is where all the progress happens
12:56 < hprmbridge> Jeremy> So for context, my general perspective right now is that advances in protein folding and nanoelectronics give us the opportunity to build(lower performance) nanomachines without leveraging mechanosynthesis. That opens up lots of early opportunities, and gives us a platform for iteratively improving nanomachines until they can do mechanosynthesis. So I'm starting that process, establishing
12:56 < hprmbridge> Jeremy> relationships and fundraising. The core technology should be open-sourced since there is little reason to try to protect it. Instead you protect and commercialize the applications of that technology to solve real problems. That lets more brains work with the tech to solve more problems and maximize the impact. That's the initial model right now.
12:59 < kanzure> there will be enough fortunes made to go around for everyone, i wouldn't worry about commercialization
13:00 < hprmbridge> Jeremy> Yep. But to raise money I benefit from framing a model that supports commercialization. 🙂
13:00 < kanzure> for a long time i was focused on nanoelectrode interfacing to proteins for protein-level control
13:00 < kanzure> in particular as a method to control polymerase enzymes, but nobody knows how to do that
13:01 < kanzure> that said, polymerase has been interfaced to nanoelectrodes for the purposes of DNA sequencing, a very cool application, but i wanted synthesis not sequencing
13:02 < hprmbridge> Jeremy> Yeah, that would be very very hard. I'm thinking we go more 'brute force' right now. De novo proteins engineered just as sliding alpha-helices with an embedded charge as an electrostatic actuator. De novo kinematic components attached to make nanomachines.  Set it up in way that you can put multiple electrostatic actuators to the same self-assembled protein device, and you've got a basic
13:02 < hprmbridge> Jeremy> nanomachine.
13:02 < hprmbridge> Jeremy> With multi-axis control.
13:02 < kanzure> btw jeremy did you see roche's sequencing chemistry announcement the other day? i only learned about their technique a few months ago: https://www.biorxiv.org/content/10.1101/2025.02.19.639056v1
13:03 < kanzure> why stop there, might as well ask for a protein nanomechanical babbage machine or numeric control with protein linkages
13:04 < hprmbridge> Jeremy> The polymerase nanopore sequencer code is very very cool, but yeah, signal to noise. I'd heard a reference or two to the sequencing by expansion tricks in the last year, but hadn't dug into it yet. That looks at first glance like a very elegant mechanism to improve S/N.
13:04 < hprmbridge> Jeremy> And sure, why not? Once the basic tech exists, we could try for numeric control.
13:04 < hprmbridge> Jeremy> Particularly important would be 'nanomachines that can assemble other nanomachines out of parts that don't intrinsically self-assemble'.
13:05 < hprmbridge> Jeremy> That gets you an iterative improvement loop, building progressively better nanomanipulators.
13:06 < kanzure> is ralph going to follow through on his "freeze by date" and do we need to stop him from doing that? pretty sure it was circa 2025...... and it's 2025 now....
13:06 < hprmbridge> Jeremy> Regarding that protein/nanoelectronic interface, check the recent DARPA MECO program announcement. It's somewhat related.
13:07 < hprmbridge> Jeremy> No idea what Ralph's plans are these days...
13:08 < hprmbridge> Jeremy> Signing off for a bit. Time to go buy toddler birthday presents.
13:13 < hprmbridge> nmz787> The grad student, nature's original quantum computer
13:20 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
13:20 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
13:22 < hprmbridge> fodagut> Hi Jeremy, welcome
13:36 < kanzure> https://www.youtube.com/watch?v=4r9L3lVpGdw&t=3m25s
15:13 < L29Ah> https://en.wikipedia.org/wiki/Starship_flight_test_8#Flight_timeline now that's a good stream
16:41 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Ping timeout: 272 seconds]
17:02 -!- WizJin_ [~Wizzy@2402:a00:184:8354:9dd9:e60b:e3eb:fe0b] has joined #hplusroadmap
17:05 -!- WizJin [~Wizzy@user/WizJin] has quit [Ping timeout: 248 seconds]
19:00 < geneh2> @Jeremy, check this out: https://arxiv.org/abs/2408.16519
19:02 < geneh2> two sheets of hexagonal boron nitride can be made to switch between nitrogen on top of boron to boron on top of nitrogen under applied electric charge. This means the two sheets move one lattice distance with respect to each other. This has been experimentally demonstrated and is being investigated as a means to make non volatile memories. This paper above proposes exploiting this effect to make an 
19:03 < geneh2> electrostatic stepper motor of sorts
19:11 < geneh2> It should be possible to step with something like 1.44 angstrom steps. It's not a very great way to make a stepper motor: the sheets have to be placed between two gold electrodes, one has to be strained to a nigh ridiculous 3-5%. There are probably better ways to exploit hBN or similar 2d material sheets being able to controllably move a single atomic distance under electrical control.
19:13 < geneh2> I think some people here were interested in synthesizing DNA, DARPA want's to synthesize DNA remotely in cells https://sam.gov/opp/bd7c123684854471abc45fa68a6237a6/view
19:19 < geneh2> more on the protein nanoelectronic interface: https://www.darpa.mil/research/programs/microsystem-induced
19:44 < fenn> "541714 - Research and Development in Biotechnology (except Nanobiotechnology)" <- sure sounds like nanobiotechnology to me...
22:02 -!- Gooberpatrol66 [~Gooberpat@2603:8080:4540:7cfb::176d] has quit [Quit: Konversation terminated!]
22:03 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
22:13 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Ping timeout: 252 seconds]
22:24 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
--- Log closed Fri Mar 07 00:00:37 2025