--- Log opened Tue Mar 21 00:00:16 2023
00:47 < hprmbridge> eleitl> fenn: Moore's law has nothing to do with ops/s.
01:21 < fenn> whatever dude
01:23 < fenn> nobody cares if their cat videos are rendered by relays or or rod logic
01:24 < fenn> turing complete compute is about the most fungible thing there is
01:25 -!- Guest67 [~Guest67@152.red-2-139-223.staticip.rima-tde.net] has joined #hplusroadmap
01:25 -!- Guest67 [~Guest67@152.red-2-139-223.staticip.rima-tde.net] has quit [Client Quit]
02:29 < hprmbridge> eleitl> When you're trying to discuss scaling limits it's difficult to use cat videos as a metric. Moore is easy to measure and has been for the past 58 years.
03:29 < hprmbridge> eleitl> Alpaca in the HN news https://news.ycombinator.com/item?id=35241787
04:27 < hprmbridge> kanzure> "restore from wayback machine" https://archivarix.com/
04:57 < hprmbridge> kanzure> we are doomed https://cdn.discordapp.com/attachments/1064664282450628710/1087706425360396338/image.png
04:58 < hprmbridge> Perry> Context?
04:59 < hprmbridge> kanzure> it's a mailing list of chemistry informatics librarians wondering where the younger adults in their profession have gone to, and so i added a reply into the thread where they were complaining about degradation of chemistry search engines
05:00 < hprmbridge> kanzure> what is particularly alarming is that i specifically said that my email was a work of fiction and generated by chatgpt, and haubenreich saw that line, and still thought it might be a joke
05:02 < hprmbridge> Perry> Chemists still almost completely hide their papers behind paywalls.
05:02 < hprmbridge> kanzure> from the hippocratic oath: "... To hold my teacher in this art equal to my own parents; to make him partner in my livelihood; when he is in need of money to share mine with him; to consider his family as my own brothers, and to teach them this art, if they want to learn it, without fee or indenture."
05:03 < hprmbridge> kanzure> arguably paywalls are against the hippocratic oath
05:03 < hprmbridge> kanzure> we should just get elon or billy gates to buy elsevier and be done with the current era of academic publishing
05:04 < hprmbridge> Perry> in chemistry, the problem is ACS etc. want it this way.
05:14 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0:4ce4:65bf:609c:13fc] has joined #hplusroadmap
05:23 < hprmbridge> lachlan> is scihub still not updating?
05:48 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has joined #hplusroadmap
06:27 < muurkha> I think Billy is in favor of paywalls
06:27 < hprmbridge> Perry> “Billy”?
06:27 < muurkha> he started up his own pseudo-open-access textbook project whose main advantage over the real open access ones is that it's under a proprietary license
06:36 < hprmbridge> kanzure> billy == bill gates
06:36 < hprmbridge> Perry> He’s an IP maximalist.
06:37 < hprmbridge> kanzure> ugh. i guess that explains why his cohort created Intellectual Ventures.
06:47 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer]
08:00 < hprmbridge> eleitl> You're still reading cheminf-l? I stopped years ago.
08:02 < muurkha> to be perfectly fair I'm not sure the textbook project has a paywall currently, but I don't know why else they'd choose a proprietary license if it isn't enshittification
08:08 < hprmbridge> eleitl> Anyone running your own solar PV? I've finally bit the bullet and started with a 800 Wp (a couple 420 Wp panels and a Hoymiles microinverter), to be upgraded to 1.6 kWp. ROIs very fast, since I'm at 0.67 EUR/kWh currently.
08:09 < muurkha> congratulations, eleitl!  that's awesome!
08:10 < muurkha> you're in Germany, right?  so not much sun, lots of clouds?
08:10 < hprmbridge> eleitl> Thanks. Average in Germany is 1 MWh/1 kWp per year. I can probably do slightly better, since I'm in the south.
08:11 < hprmbridge> eleitl> Glas-glas panels, so should be good for 30+ years.
08:11 < hprmbridge> eleitl> Microinverters will probably last about 15 years, hopefully next-gen is more reliable.
08:12 < muurkha> that's about an 11.4% capacity factor?  I thought the overall average capacity factor for utility-scale solar in Germany was like 10% or something
08:12 < muurkha> I guess that's the same to one significant figure
08:13 < muurkha> you could probably build a microinverter for a much longer lifetime pretty easily, but it isn't going to happen through the market; the feedback time is too long
08:13 < hprmbridge> eleitl> It's a rule of thumb. Depends on local microclimate, some places are extremely sunny. And, of course, local shadowing. I'll have to measure the harvest factor over the year.
08:14 < muurkha> (I'm Kragen Sitaker, btw.  Glad to reconnect with you.)
08:14 < hprmbridge> eleitl> I understand that currently capacitors are a bottleneck.
08:14 < hprmbridge> eleitl> Oh hi Kragen! It's been a very long while. Are you still in Argentina?
08:15 < muurkha> yup, still am
08:15 < muurkha> getting married next week
08:15 < hprmbridge> eleitl> Congratulations! Buenos Aires still good, despite mondo inflation?
08:16 < muurkha> well, the economy is kind of collapsing, and people are more and more desperate
08:16 < muurkha> but day to day things still look pretty much the same.  the newspaper stand guy still sells newspapers, singing along to US pop music.  the halal butcher still sells lamb and beef.  you can still flag down a taxi
08:17 < hprmbridge> eleitl> You're only slightly ahead of the rest of us. Europe is in deep doo-doo, but not many are fully realizing it yet.
08:17 < muurkha> although every restaurant now has colorful signs outside announcing which bicycle-delivery cellphone applications they're signed up with, and the remiserías have largely closed as the drivers have gone to work for Uber, Cabify, and Didi
08:18 < hprmbridge> eleitl> How is the grid reliability? kWh price?
08:19 < muurkha> we've been having a lot of small power outages over the last week, and on March 1 we had a blackout covering 40% of the country for several hours
08:19 < muurkha> I live on the third floor and over the past couple of days I've been taking the stairs
08:20 < muurkha> a week ago I was going downstairs with my girlfriend (not the one I'm getting married to on Monday) and the power went out.  we met a neighbor coming up the stairs who barely escaped being trapped in the elevator
08:21 < muurkha> that outage lasted about 12 hours, which would have been a long time to be trapped in the elevator, especially at 31°C
08:21 < hprmbridge> eleitl> Not good. This is bad news for any industry depending on constant processes, which is pretty much any producing industry.
08:22 < hprmbridge> eleitl> I've heard about your heat wave, too. We've been spared from a real heat dome so far, but it's only a question of time until we get 40+ deg C.
08:22 < muurkha> my fiancée and I were evaluating event spaces for the wedding reception, and one of the spaces we looked at advertised that they had two permanently installed (250 kVA?) diesel generators
08:23 < muurkha> but that one wanted to charge us US$43 per person for catering, so we went with one that cost half as much
08:24 < hprmbridge> eleitl> Are you running a large UPS privately? I'm considering going something in 1 kWh range next year, to prepare for when our grid starts to crap out.
08:24 < muurkha> industry is really on the ropes here; the fake exchange rate makes it really hard to export or to compete with imports, unless you can smuggle your products out of the country
08:25 < muurkha> I wish, but I'm in abject poverty due to lack of getting shit done
08:25 < muurkha> so when the power went out I didn't even have a fan.  well, except a fake silk hand fan I bought in Chinatown
08:26 < hprmbridge> eleitl> Industry in Germany is being busily murdered by high energy prices. It will just take a while.
08:26 < muurkha> yeah, how did you end up with such high energy prices?
08:27 < muurkha> at least non-energy-intensive, non-labor-intensive industries in Germany are still very competitive internationally
08:27 < muurkha> things like high-precision machining, which uses a lot of energy per product but not per euro of product
08:28 < hprmbridge> eleitl> Gas deliveries went to spot market prices instead of long-term contracts, so speculation drove up the price. Then we imposed sanctions on pretty much everything important coming from Russia. Including helpful allies blowing up our gas pipelines in the North Sea.
08:29 < muurkha> I think you misspelled "colonizers"
08:29 < hprmbridge> eleitl> BASF is shutting down, pretty much everything energy-intensive is shutting down. Aluminium and glas are just the canaries in the coal mine, it affects everything, e.g. bread baking.
08:29 < muurkha> although I hear that within Germany that's a taboo point of view because the ultra-right-wing parties agree with it
08:30 < muurkha> wow, BASF is shutting down?
08:30 < hprmbridge> eleitl> Yeah, helpful allies was meant sarcastically.
08:30 < muurkha> I don't think the US counts as even an unhelpful ally
08:30 < hprmbridge> eleitl> Not everything, not all at once, but they've stopped some production, are firing some people, and investing where energy is still cheap. Like U.S., for instance.
08:31 < muurkha> that makes sense.  do they have plants in other EU countries with better energy prices?
08:31 < hprmbridge> Perry> That’s crazy even for me. And I am paying a lot.
08:31 < muurkha> Perry: what, €0.67/kWh?
08:32 < hprmbridge> eleitl> Nitrogen fixation (ammonia) stopping are of course having an impact on fertilizer production, so expect lower harvests. And we're in a chronical draught, so I'm at the outskirts of the Alps where it's slightly wetter.
08:32 < muurkha> presumably BASF pays less, but still much more than the €0.03/kWh routinely delivered by PV in less polar countries
08:33 < hprmbridge> eleitl> Yeah, 0.67 EUR/kWh. I'll be switching to a 0.37 EUR/kWh contract, but that's in a year. So having some 1.4 kWp PV with a meter that can (illegally) run backwards pays for itself in less than 2 years.
08:34 < hprmbridge> eleitl> It's not just electricity, BASF needs natural gas, and that is also not cheap.
08:34 < hprmbridge> Perry> Sounds like a great time to shut down more nuclear plants.
08:35 < hprmbridge> Perry> I had no idea Buenos Aires had such unreliable power. One of my friends is down there, and he’s never mentioned it.
08:35 < jrayhawk> nordstream 1 was already shut down, and half of nordstream 2 appears to bave been a failure of workmanship. the other half of nordstream 2 remains operational but unutilized. "helpful allies" do not, in this specific context, seem to be a relevant factor.
08:36 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
08:37 < muurkha> Perry: it varies a lot by neighborhood; when I lived in San Telmo I never saw power outages, just pickpockets
08:38 < hprmbridge> eleitl> You might have missed these seismic events of 1 TNT equivalent each which ended 6 pipes. A certain Sy Hersh has published his leaked version, which I'm not entirely buying. It doesn't matter whether Norway was involved, or not.
08:38 < hprmbridge> Perry> ah.
08:38 < muurkha> when I lived in the Abasto we had a three-week power outage
08:38 < muurkha> not really a poorer neighborhood, more that it was a neighborhood without political influence
08:38 < hprmbridge> Perry> Seymour Hersh burned his credibility a long time ago. It’s possible he is correct of course.
08:39 < muurkha> yeah, Hersh might be wrong, but he also does have a history of sometimes being right when nobody else is
08:39 < hprmbridge> eleitl> The facts on the ground (under water, rather) speak for themselves. Somebody has been very thorough. I wonder about these two pipes which didn't blow up, and whether somebody took care of these charges which failed to detonate.
08:40 < muurkha> can't you import fertilizer from places that aren't sanctioning Russia?
08:41 < hprmbridge> eleitl> Anyways, we're now still buying Russian gas, but now LNG. And more expensive.
08:42 < muurkha> most fertilizer is cheaper, easier, and safer to transport than LNG, though there have been some spectacular fertilizer accidents
08:43 < hprmbridge> eleitl> Most places stopped producing ammonia since gas prices are global. Fracking gas is still relatively cheap in the U.S. and they probably have nitrogen fixation plants there. However, the fracking thing is a flash in the pan, so it doesn't change the long-term situation. We're getting energy-poor globally, fast.
08:43 < muurkha> but things like phosphates and sulfates of ammonia are inert
08:43 < muurkha> oh, that makes sense
08:43 < hprmbridge> eleitl> Liquid ammonia is also easy to transport. There are even pipelines for it.
08:44 < hprmbridge> Perry> I will bet against that.
08:44 < muurkha> getting energy-poor globally would be a pretty big departure from my predictions in Dercuano.  have you read Dercuano, eleitl?  you'd probably be among the people who would most enjoy it
08:44 < hprmbridge> Perry> I don’t think we’re going to go energy poor at all.
08:44 < hprmbridge> Perry> Germany might.
08:45 < hprmbridge> eleitl> https://dercuano.github.io/ ? Thanks, I'll check it out.
08:45 < muurkha> yeah, I think we're likely to see unprecedented energy abundance when PV takes over, but we're still almost a decade away from that
08:45 < muurkha> yeah, though I'd recommend downloading a local copy: http://canonical.org/~kragen/dercuano/
08:45 < hprmbridge> eleitl> Perry, in terms of net energy per capita we're declining, and due to EROEI the future situation is worse.
08:46 < muurkha> uh, http://canonical.org/~kragen/dercuano
08:46 < muurkha> without the trailing slash
08:46 < hprmbridge> Perry> maybe you are in Germany. The US is not going to, at least not over long periods.
08:47 < hprmbridge> eleitl> Got that, Kragen. Thanks.
08:47 < muurkha> the US is hardly a paragon of enlightened infrastructure development
08:47 < hprmbridge> Perry> and I will happily bet that the long-term price of both natural gas and energy in general goes down in real dollars.
08:47 < muurkha> PRC and UAE might be in a better position
08:48 < muurkha> and, I don't know, Indonesia or something, despite the clouds
08:48 < hprmbridge> eleitl> Perry, the geology of tight resource extraction is unfortunately not favorable. Which explains why U.S. is so interested in foreign fossil resources.
08:48 < hprmbridge> Perry> I am happy to make an equivalent of Julian Simon’s bet.
08:49 < muurkha> how would you go about betting on the long-term price of natural gas and energy?  CME futures contracts only go 12 years out; is that long enough?
08:50 < hprmbridge> eleitl> I don't bet. I read graphs. Renewable unfortunately is neither cheap enough nor happening fast enough at scale. Right now renewable is a way to stretch cheap, dirty coal.
08:51 < hprmbridge> Perry> Bet on the prices of both in inflation adjusted dollars in N years. I presume Eugen is willing to offer odds?
08:51 < muurkha> if N < 12 you can definitely do that with CME
08:51 < muurkha> I suppose you have a trading account since you were working at Shearson Lehman Hutton when I met you 41 years ago :)
08:52 < hprmbridge> Perry> https://www.econlib.org/archives/2012/05/the_bettors_oat.html
08:52 < muurkha> uh, 31
08:53 < muurkha> eleitl: what do you think are the critical limiting factors for the growth of renewables?  They certainly seem cheap enough in sunny or windy areas
08:53 < kanzure> i am trying to get speaker-segmentation and whisper.cpp to work together-- so far whisper doesn't produce transcripts as beautiful as mine https://gist.github.com/kanzure/233b7af168c7546fca0614d04c213482
08:53 < hprmbridge> Perry> It was just Lehman internally even though we were an American express subsidiary at the time.
08:53 < muurkha> heh, I didn't realize that
08:53 < kanzure> (unfortunately pyannote is giving me a bunch of trouble with installation- no time to sort this out)
08:54 < hprmbridge> Perry> I understand it’s not as good at that as one would like. Wait a couple of months I suppose.
08:54 < kanzure> well it could just be that nobody is thinking about this use case; pyannote looks like it should be able to do speaker separation for me. i don't need speaker labeling.
08:55 < hprmbridge> Perry> We were shearson.com when I arrived but everyone answered the phone as Lehman and we were lehman.com very soon even externally.
08:55 < muurkha> yeah, ISTR pmetzger@shearson.lehman.com
08:56 < hprmbridge> Perry> everyone is thinking of that use case. We’ve had the same problem at my start up. I won’t comment beyond that.
08:56 < hprmbridge> Perry> We never did that.
08:56 < hprmbridge> eleitl> It is difficult to summarize the issues we're facing, so I can only point to https://escholarship.org/uc/energy_ambitions which does a very thorough job of it.
08:56 < hprmbridge> Perry> I was at shearson.com for a while, and then at lehman.com.
08:57 < hprmbridge> eleitl> I've landed at Capgemini at the moment. Meh, it's a living.
08:57 < muurkha> human memories do strange things over the decades
08:58 < hprmbridge> Perry> I ran email among 30 other things so I remember that well.
08:58 < muurkha> yes, but evidently I don't remember it that well :)
08:58 < muurkha> eleitl, thanks for the link to Tom Murphy's book; I didn't realize he'd made a book of it.  I'll read it carefully
08:59 < hprmbridge> eleitl> In case you find any errors, please mention these.
09:00 < muurkha> hopefully we can escape from the finite planet soon
09:01 < hprmbridge> eleitl> The meta problem is that we're a population of evolutionary agents, which makes their ensemble behaviour almost deterministic. Breaking out from that requires something extremely disruptive, and fast-growing.
09:01 < hprmbridge> eleitl> People who believe that AGI is round the corner think the problem is already solved.
09:01 < kanzure> well, satoshi announced bitcoin to the extropians and cypherpunks and we kind of blew it
09:02 < muurkha> it just has to be faster than evolution, which isn't too fast
09:02 < kanzure> nearly missed the financial singularity
09:02 < muurkha> blew it how?
09:02 < kanzure> well, it was an opportunity to implement the electronic cash dreams and then fund the necessary projects
09:03 < muurkha> people do seem to be using it for that
09:03 < hprmbridge> eleitl> p2p cryptocurrency alone isn't all that disruptive. Putting a few GUSD into the hands of the right people potentially is, but look how that money is spent currently.
09:04 < kanzure> not really. i mean maybe jed mccaleb with his space habitat company and astera institute stuff?
09:04 < muurkha> Tim May predicted it would be a lot more disruptive than it's turned out to be so far
09:04 < kanzure> well, too much disruption and it's global ruine anyway
09:05 < muurkha> yeah, that was my concern
09:05 < hprmbridge> eleitl> How about really cheap thin-film solar PV? Or electrocatalysts for O2 and H2 production. That would really give us a little more time.
09:05 < hprmbridge> eleitl> Look how long it's taking to just get sodium batteries.
09:06 < kanzure> part of my concern is that a lot of tech seems to require a whole "social movement" just to convince people that it should be funded, because most things are just slightly out of reach of the low-end budgets
09:06 < muurkha> traditional silicon PV is already pretty cheap: https://www.solarserver.de/service-tools/photovoltaik-preisindex.html certainly quite cheap enough to beat fossil fuels in most of the world
09:06 < muurkha> and without the scaling limits of fossil fuels
09:07 < kanzure> and, after being consistently right about technology, by now someone should have the funding scale to invest in the really interesting projects we should have been doing all along.... or something.
09:07 < muurkha> and hydrofracking is also cheap enough for "enhanced" geothermal to compete with fossil fuels on even grounds, again without the scaling limits (though there's the risk of provoking a social movement in response, as happened in Basel)
09:10 < muurkha> (note, though, that the Preisindex has been frozen at €0.19/Wp for four years now, with an excursion to €0.15 during the pandemic quarantines; production is not scaling up and prices scaling down as they should be)
09:10 < hprmbridge> eleitl> Right now PV panels are less than half the price of new installations. Inverters like power electronics don't scale so fast -- we will need SiC and GaN stuff to make low-loss power electronics. The storage issue can potentially be solved with hydrogen, but then if you look how much methane you need to substitute, while we're at some 4% with PV for energy. And even without storage <0.1 EUR/kWh
09:10 < hprmbridge> eleitl> isn't here. But do read Tom Murphy, he treats it thoroughly.
09:11 < hprmbridge> eleitl> I need to do some work, so bbl.
09:11 < muurkha> I will, thanks
09:11 < muurkha> kanzure: what's your top three list of necessary projects?
09:12 < muurkha> I need to go meet up with someone who unexpectedly became rich from cryptocurrency in 20 minutes
09:12 < muurkha> though not at GUSD scale
09:12 < kanzure> biological intelligence acceleration and enhancement, germline engineering, mechanical self-replication, extremely good DNA synthesis, molecular nanotechnology
09:13 < kanzure> human embryo genetic engineering
09:13 < muurkha> by mechanical self-replication do you mean specifically non-MNT?
09:13 < kanzure> i'd be pretty happy with either!
09:15 < kanzure> to be fair we already have biological self-replication and we might have to live with that, but the computer people don't like this answer.
10:00 < kanzure> https://www.wired.com/story/i-saw-the-face-of-god-in-a-tsmc-factory/
10:01 < kanzure> "A disassembled EUV takes up 40 shipping containers"
10:01 < kanzure> "Shipping it (mostly to Asia) takes 20 trucks and 3 Boeing 747s"
10:01 < kanzure> "ASML teams must be on-the-ground to maintain them" "The min spend to house EUVs is $1B"
10:02 < kanzure> "ASML can only produce 50 EUV machines a year"
10:02 < kanzure> "speciality parts like the Zeiss lens takes 40 weeks to produce"
10:04 < kanzure> another source says more than 40 weeks: "And then. . . it takes more than 12 months to make the lens.""
10:05 < kanzure> "EUV: 5,000 suppliers provide 100,000 parts, 3,000 cables, 40,000 bolts and 2 kilometers of hosing. The tool weighs about 180,000 kilograms (200 tonnes), and ships in 40 containers spread over 20 trucks and three cargo planes."
10:05 < hprmbridge> lachlan> “In spite of the fact that he himself trained as an engineer at MIT and Stanford, Morris Chang, who founded TSMC in 1987, has long maintained that American engineers are less curious and fierce than their counterparts in Taiwan. At a think-tank forum in Taipei in 2021, Chang shrugged off competition from Intel, declaring, "No one in the United States is as dedicated to their work as in Taiwan."
10:05 < kanzure> "Inside an EUV machine, 50,000 droplets of molten tin fall through a chamber at its base each second. A pair of lasers zap every drop, creating a plasma that in turn releases light of the desired wavelength [Shorter wavelengths allow the etching of smaller components/more powerful chips]. The mirrors guiding this light, made of sandwiched layers of silicon and molybdenum, are ground so ...
10:06 < kanzure> ...precisely that, if scaled to the size of Germany, they would have no bumps bigger than a millimeter. Because EUV light is absorbed by almost anything, including air, the process must take place in a vacuum."
10:06 < kanzure> "ASML boasts its EUV system controls beams of light so accurately that it is equivalent to shining a light torch from the earth and hitting a 50 eurocent coin placed on the moon."
10:09 < kanzure> i don't know, i really don't see why someone else isn't trying to build these machines
10:10 < hprmbridge> lachlan> There is another company
10:10 < hprmbridge> lachlan> I forget their name
10:10 < kanzure> yes they are frightening numbers but there is clear demand and well-known physics. venture capital should be lining up.
10:10 < hprmbridge> lachlan> I imagine intellectual property is also an issue
10:10 < kanzure> modify the idea, then keep it in house and don't talk about it. i dunno. there are many things to do.
10:11 < kanzure> maybe these guys? https://astrileux.com/index.html
10:12 < hprmbridge> lachlan> I don’t think so
10:12 < hprmbridge> lachlan> Iirc they were using FELs to generate EUV
10:12 < hprmbridge> lachlan> Free electron lasers
10:13 < kanzure> here is someone at the lithography conference writing some notes https://lithoguru.com/life/?p=696
10:13 < hprmbridge> lachlan> I’m not sure though, I don’t think that kind of tech is approachable right now
10:13 < kanzure> (from chris mack i think)
10:32 < L29Ah> > How about really cheap thin-film solar PV?
10:32 < L29Ah> they promised those 20 years ago, and they are still more expensive than silicon per watt
10:38 < hprmbridge> Perry> photovoltaics are already cheap enough that the main expenses are the packaging and installation. There is no particular need for any technological breakthroughs on cells.
10:41 < kanzure> twitter spaces coming up in a few minutes where i will be telling them how to figure out anti-aging genetic interventions
10:41 < kanzure> https://twitter.com/samira_kiani1/status/1638233742312636422
10:50 < L29Ah> except that 20% efficiency kinda sucks
11:12 -!- superz [~superegg@user/superegg] has quit [Ping timeout: 246 seconds]
11:14 -!- codaraxis [~codaraxis@user/codaraxis] has joined #hplusroadmap
11:17 < kanzure> for some reason i'm not able to speak in the twitter spaces- if someone else would like to jump in my place, please join
11:24 -!- cthlolo [~lorogue@77.33.23.154.dhcp.fibianet.dk] has joined #hplusroadmap
12:05 < hprmbridge> nmz787> Heh, that's the site of Chris Mack (the gentleman). I was in a meeting with him last week
12:08 < hprmbridge> Perry> 20% isn’t bad. Getting close to the single absorber limit inexpensively isn’t easy, and getting multiple absorbers really requires MNT.
12:09 < hprmbridge> Perry> BTW, the UI of the IRC gateway sucks.
12:09 < hprmbridge> Perry> or rather the UX.
12:14 < kanzure> one of the questions was "what if this creates a permanent underclass of slave labor?"
12:14 < kanzure> my response was "well even human slaves in the past were pretty good at escaping, i don't think that's going to be a problem"
12:15 < kanzure> pmetzger: you could just use an IRC client if you wanted to; discord is a new experiment we are trying. irc.libera.chat #hplusroadmap
12:16 < L29Ah> the cheap thin film ones would allow e.g. coating buildings and vehicles with photovoltaic w/o much thought
12:17 < L29Ah> we aren't there yet: the silicon ones are heavy and fragile and very flat
12:17 < L29Ah> so they need annoying frames and protective glasses
12:17 < L29Ah> can't take them with me on a hike rolled into a thin tube
12:18 < hprmbridge> Perry> I wouldn’t be here at all if it wasn’t for discord.
12:18 < L29Ah> mfw discord > transhumanism
12:19 < hprmbridge> Perry> I don’t think that that would be a particularly easy or cheap thing to do given the state of materials science. With molecular manufacturing maybe, but not with the way that we produce photovoltaic cells right now. And at the moment, the current manufacturing methods are really cheap and efficient.
12:20 < hprmbridge> Perry> thin film ones will be even more fragile. Cleaning the surface will destroy them.
12:22 < L29Ah> a thin film cell painted onto a wall could use a layer of transparent epoxy or such
12:22 < hprmbridge> Perry> and why can’t you use that for current cells? I can give you some hints.
12:22 < L29Ah> or what those car finishes are made of?
12:23 < hprmbridge> Perry> you can’t even use normal glass, you have to use low iron glass in order to have it sufficiently transparent in the infrared.
12:23 < L29Ah> because i can't paint pure silicon on a wall, nor glue a film to it
12:24 < hprmbridge> Perry> there aren’t magic materials here, and if there were, people would probably be using them already. Scratches on the surface reduce absorption. You need something that will be remains scratch free, can be cleaned, and is transparent throughout the necessary region of spectrum. There are other materials that can be used, but they aren’t cheap.
12:24 < L29Ah> *nor have it glued
12:24  * L29Ah returns to thumb twiddling
12:24 < hprmbridge> Perry> most epoxies photodegrade and photo age. None that I know of are transparent through the necessary portions of the spectrum.
12:25 < hprmbridge> Perry> all would have much worse aging characteristics than glass.
12:25 < L29Ah> seems like PVs RoI is ~10 years here implying total power use (that's not going to happen since the grid doesn't accept power from consumers, and batteries are too expensive)
12:26 < hprmbridge> Perry> anyway, current photovoltaic technology is not going to get displaced particularly quickly. It’s really good. It has excellent characteristics, like it ages well, it’s cheap to manufacture, it’s straightforward to install, etc.
12:26 < L29Ah> after the grid company gets fixed, it would be awesome
12:33 < hprmbridge> Perry> I did a study for an investor around 2008 on the future of photovoltaics. My general prediction was that conventional monocrystalline silicon cells were going to win without any question, and if there was no particular role for other technologies. So far, everything that I said would happen seems to have.
12:34 < hprmbridge> Perry> scaling of the manufacture of monocrystalline Si production was easy, and the cells age well in the field. Existing efficiencies were more than good enough.
12:35 < hprmbridge> Perry> we will have access to much more efficient technologies given MNT, but I think that for at least 20 years to come, absent MNT, monocrystalline Si is it.
13:06 < kanzure> nmz787: pmetzger: qurtz ingot vs furnace refining better discussed here i think.
13:08 < kanzure> .. quartz, rather.
13:29 < hprmbridge> nmz787> I just never considered something mined to also be called sand
13:46 < muurkha> it's hard to program biological self-replicators to make other things on demand
13:48 < muurkha> Perry: "packaging" in the case of photovoltaics is, as I understand it, mostly gluing them to a sheet of glass with transparent EVA
13:49 < muurkha> as L29Ah points out, it would be nice to get 30% or 40% efficient multijunction cells to be cheap
13:49 < hprmbridge> Perry> It would also be nice for us to have good aging reversal tech. And someday we'll have it.
13:49 < hprmbridge> Perry> A triple absorber should be able to do 65% or more. Someday.
13:49 < hprmbridge> Perry> Right now, it's far cheaper and easier to use monocrystalline Si.
13:49 < kanzure> i think de-programmed aging is more likely to be achieved first before aging reversal
13:50 < hprmbridge> Perry> My point wasn't about aging.
13:50 < muurkha> quite aside from the potential greater portability etc. of flexible thin-film solar cells, gluing the cells to protective glass makes them cost at least as much as the glass
13:50 < hprmbridge> Perry> Yes, so? The "thin film" stuff won't survive weather or cleaning without glass either.
13:51 < hprmbridge> Perry> If you have a magic polymer protection layer you can use it with monocrystalline Si right now too.
13:52 < hprmbridge> Perry> Thin film ideas were there because people believed they'd be cheaper. Same with chalcogen cells, organic cells, loads of other ideas.
13:54 < hprmbridge> Perry> All of them were attempts to lower costs, most of them weren't actually a good way to lower costs. Organics were a really bad idea.
13:55 < muurkha> L29Ah: by "RoI 10 years here" do you mean that in Montenegro PV returns 10% on investment every year, valued at the cost of grid power?  that sounds pretty decent already.  https://solargis.com/maps-and-gis-data/download/montenegro says PVOUT for Montenegro is in the range 1200–1500 kWh/year/kWp (13.7%–17.1% capacity factor) while California I think averages 29% capacity factor
13:55 < muurkha> and places like Perú, Egypt, and the UAE are higher still
13:55 < hprmbridge> Perry> (Organics were a crazy bad idea. Any organic chromophore capable of lofting an electron from HOMO to LUMO so you can harvest it is also capable of photobleaching, and will, quickly.)
13:56 < hprmbridge> Perry> (Why doesn't monocrystalline Si have this problem? When you break a bond between adjacent Si atoms, there's no where for that bond to go but reforming. When you have organics, you do a rearrangement reaction most of the time.)
13:56 < muurkha> Perry: I think what has actually won is not conventional *monocrystalline* silicon cells but conventional *polycrystalline* silicon cells (and I agree with you that thin film is not a viable competitor at least for the next ten years)
13:56 < hprmbridge> Perry> Monocrystalline is the bulk of the stuff being sold.
13:57 < muurkha> Really? I thought the low-cost (16% efficient) stuff was polysilicon
13:57 < kanzure> "Folding@home: achievements from over twenty years of citizen science herald the exascale era" https://arxiv.org/abs/2303.08993
13:58 < kanzure> just a review article (probably because folding@home decided to close down/suspend for the moment)
13:58 < hprmbridge> Perry> Not to my knowledge. So far as I know the bulk is really cheap monocrystalline. It's all from scale-up of the manufacturing of the raw ingots.
13:58 < hprmbridge> Perry> It went from a couple thousand tons a year to millions. Huge scale up opportunity.
13:59 < muurkha> L29Ah: even 10% annual RoI sounds great compared to the stock market, but there's the chance that those returns evaporate quickly if the scale-up of renewables drives prices down (as Perry bets it will, and eleitl and Tom Murphy think it won't)
14:00 < muurkha> hmm, how would we find out whether the bulk of shipped solar cells are polycrystalline?  I don't have any utility-scale solar installations near here I can visit
14:00 < hprmbridge> Perry> There were papers 20 years about how you could reduce costs of the Si refining if you were doing it at 10x or 100x the scale. That happened.
14:00 < hprmbridge> Perry> Look at the products people will sell you. They're mostly monocrystalline.
14:01 < muurkha> the difference is really obvious when you look at the cells
14:01 < muurkha> the products on Amazon may not be representative of what utilities are installing
14:01 < hprmbridge> Perry> So far as I know the mass installations are (as I said) monocrystalline.
14:01 < muurkha> right, but how would we find out?
14:02 < hprmbridge> Perry> look for a gartner report or something.
14:02 < muurkha> heh
14:02 < muurkha> which Gartner Bullshit Quadrant is CIGS in?
14:02 < L29Ah> muurkha: yup, now .me prices are 0.13€/kWh, and will likely rise further, catching up to western europe; the shore has plenty of sunshine all year round
14:02 < muurkha> Perry: you might be interested in my notes on the topic from 2008, despite the embarrassing errors in them
14:03 < muurkha> https://dercuano.github.io/notes/solar-economics.html
14:05 < muurkha> L29Ah: 0.13€/kWh sounds a bit high, but at least it's not Germany.  10% annual RoI is enough to pay for quite a bit of battery.  do you think thermochemical energy storage for HVAC will become economically significant?
14:05 < L29Ah> but anyway, the PV prices don't include installation and inverters
14:05 < muurkha> yeah
14:05 < muurkha> and the rest of "balance of plant"
14:06 < muurkha> Perry: 2008 was right before the polysilicon pricing bubble popped and sunk Evergreen, which might mean it was the last time you could get any kind of usable information about the economics of solar energy from EDGAR
14:06 < L29Ah> idk about thermochemical energy storage, sounds complex; i'm more interested in solar water heating panels tbh, since it's much cheaper and is more useful
14:06 < kanzure> https://www.gatesnotes.com/The-Age-of-AI-Has-Begun
14:07 < L29Ah> anyway i don't even have any real estate (very ₿ maxi)
14:07 < hprmbridge> Perry> I think batteries are the right thing. IMHO, best current chemistry is NaS.
14:07 < muurkha> solar water heating panels are pretty great but, nowadays, usually aren't actually cheaper than PV and a resistor
14:08 < hprmbridge> Perry> There's a French company called DualSun that does both.
14:08 < muurkha> which I find astounding
14:08 < hprmbridge> Perry> PV + water heating, same panel.
14:08 < hprmbridge> Perry> Drives up efficiency a great deal for a high latitude user.
14:08 < muurkha> what are the key obstacles to mass deployment of NaS batteries?
14:09 < hprmbridge> Perry> For some reason only NGK is really selling them. I think the big problem is fabrication of the beta alumina electrolytes.
14:09 < muurkha> that's weird, β-alumina is supposed to be super easy to make
14:10 < hprmbridge> Perry> It isn't. You need the right structure to use it in NaS batteries at 300˚C.
14:10 < muurkha> not quite as easy as α-alumina, which you can do by accident
14:10 < hprmbridge> Perry> I read a bunch of papers on the issues years ago.
14:10 < hprmbridge> Perry> Haven't cared since my battery storage startup was stillborn.
14:11 < muurkha> my girlfriend got annoyed at the bottle of Al(OH)₃ in the sink, though, and I don't have XRD or DTA, so I haven't actually tried making different alumina phases myself
14:12 < hprmbridge> Perry> GE had a huge plant for making zebra batteries in upstate NY, they shuttered it prematurely IMHO.
14:12 < hprmbridge> Perry> But they're not the same company they were at one pont.
14:12 < hprmbridge> Perry> point
14:12 < muurkha> obviously, Thomas Edison died
14:12 < muurkha> what sunk the startup?
14:13 < hprmbridge> Perry> De facto, lack of cohesion among the founders.
14:13 < hprmbridge> Perry> I still think our business plan was good but I'm not really interested in it any more.
14:13 < muurkha> that's usually the proximate cause; what did the lack of cohesion stem from?
14:13 < muurkha> I mean I guess usually the proximate cause is that people quit, but one step back from that is lack of cohesion
14:13 < hprmbridge> Perry> Long story and I think Bryan keeps transcripts and I don't want to annoy people.
14:13 < muurkha> he does, yes
14:14 < kanzure> "The 1967 experiment that proved that anyone could design a nuclear bomb" https://www.amusingplanet.com/2023/03/the-1967-experiment-that-proved-anyone.html
14:14 < muurkha> sorry, I didn't mean to pry
14:14 < muurkha> am I confusing my phases of alumina?  maybe β-alumina is one of the more delicate ones instead of the second easiest one to make
14:15 < kanzure> "nth country experiment"
14:17 < muurkha> oh, https://en.wikipedia.org/wiki/Beta-alumina_solid_electrolyte says it's not actually a phase of alumina (but was thought to be in the 01930s, so I guess it's not *that* hard to make)
14:21 < hprmbridge> Perry> So far as I know NGK is the main company that has commercialized production of beta alumina electrolytes; FIAMM, GE and some others have had production but don't currently.
14:21 < muurkha> I don't think anyone has commercialized production of cacodyl either
14:21 < muurkha> or sodium triiodide
14:22 < hprmbridge> Perry> I don't think those are relevant here.
14:22 < muurkha> well, they do have bigger cons
14:23 < muurkha> but my point is that there are things that are easy to synthesize that nobody sells
14:23 < hprmbridge> Perry> BASE isn't easy to do, according to the papers. Things like porosity etc have to be carefully controlled.
14:24 < hprmbridge> Perry> There was a group in Shanghai IIRC that published a bunch of materials science papers overviewing the problem. I haven't touched this in over a decade though.
14:25 < muurkha> yeah, that's the kind of thing that could make it hard to have a market for β-alumina
14:25 < muurkha> if all your users have strange requirements for things like porosity or isotopic composition or whatever, they might be better off synthesizing it themselves instead of buying it from you
14:34 -!- cthlolo [~lorogue@77.33.23.154.dhcp.fibianet.dk] has quit [Quit: Leaving]
15:45 -!- codaraxis__ [~codaraxis@user/codaraxis] has joined #hplusroadmap
15:45 -!- codaraxis [~codaraxis@user/codaraxis] has quit [Ping timeout: 276 seconds]
16:43 -!- codaraxis [~codaraxis@user/codaraxis] has joined #hplusroadmap
16:44 -!- codaraxis__ [~codaraxis@user/codaraxis] has quit [Ping timeout: 255 seconds]
17:01 -!- codaraxis__ [~codaraxis@user/codaraxis] has joined #hplusroadmap
17:03 -!- codaraxis [~codaraxis@user/codaraxis] has quit [Ping timeout: 264 seconds]
17:42 -!- superz [~superegg@user/superegg] has joined #hplusroadmap
17:43 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
17:44 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 246 seconds]
17:44 -!- stipa_ is now known as stipa
17:50 -!- AMG [ghebo@user/amg] has quit [Ping timeout: 252 seconds]
18:39 < muurkha> eleitl: it's surprising that Murphy thinks geothermal "isn't worth serious attention, due to [its] small scale"; this suggests he perhaps hasn't kept up with developments in enhanced geothermal systems
18:39 < muurkha> the geothermal energy in the crust is several orders of magnitude larger than that of fossil fuels
18:40 < kanzure> what do you make of this approach: https://www.quaise.energy/
18:40 < kanzure> "gyrotron-powered drilling platform vaporizes boreholes through rock"
18:41 < kanzure> "high-power millimeter waves"
18:43 < muurkha> we talked about this in here a year or two ago, didn't we?
18:43 < kanzure> .llmgrep geothermal
18:44 < muurkha> search for gyrotron
18:44 < hprmbridge> lachlan> llmgrep ?
18:45 < kanzure> as a general rule bots should do what i say as well as what i mean
18:45 < muurkha> it seems like a reasonable approach but it's not required to unlock geothermal energy; oil drillers are good enough at making boreholes through rock already
18:49 < kanzure> oh, it's not an unsolved technology problem?
18:49 < muurkha> well, the unsolved technology problem is turning the heat into electricity cheaply enough
18:50 < muurkha> also there's earthquakes, but frackers say they've solved those already
18:51 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0:4ce4:65bf:609c:13fc] has quit [Quit: Leaving]
19:09 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
19:11 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 255 seconds]
19:11 -!- stipa_ is now known as stipa
19:36 < fenn> i wasn't too impressed with murphy's book. he devotes an entire chapter to saying, basically, "Space is big. Really big. You just won't believe how vastly hugely mind-bogglingly big it is" and then concludes that therefore space colonization is impossible and we should all be ashamed for being interested in it, because it's distracting everyone from being doomers
19:37 < fenn> also that wired article on TSMC was shit
19:37 < kanzure> howcome there's just so much "only ASML has the will to do this" marketing going around
19:37 < fenn> mumble mumble biden xi cross-strait tensions
19:38 < fenn> because AI software can't be export controlled, only AI hardware
19:38 < fenn> if anybody could build their own fab, it would diminish USA soft power
19:39 < fenn> (nevermind that ASML is dutch)
19:39 -!- codaraxis__ [~codaraxis@user/codaraxis] has quit [Ping timeout: 276 seconds]
19:42 < fenn> ASML's monopoly doesn't necessarily have to have a basis in fact; if everyone with capital believes the myth, that's sufficient to stop any unauthorized fabs from being built
19:43 < fenn> whether it's true or not, i have noidea
19:43 < muurkha> it's unlikely that Satoshi Nakamoto will build secret EUV machines
19:43 < fenn> i'm confused by the part where they're using lenses instead of mirrors
19:45 < fenn> i guess lenses are old tech and people write about it as if it's relevant to EUV for some reason
19:47 < hprmbridge> lachlan> as much myth as anything else
20:08 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer]
20:11 -!- codaraxis [~codaraxis@user/codaraxis] has joined #hplusroadmap
20:42 < hprmbridge> nmz787> Kanzure the thing is you /could/ build a fab without ASML equipment. But the market economics show people want chips (or products powered by such chips) that currently require a process that only ASML provides.
20:47 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
20:50 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 268 seconds]
20:50 -!- stipa_ is now known as stipa
21:04 < hprmbridge> nmz787> Hmm, to buy a remote controlled lawn/brush mower or not...
21:05 < hprmbridge> nmz787> Tracked drive nonetheless
21:12 < hprmbridge> nmz787> I'm tempted to just start a mini+knockoff digger land
22:27 -!- AMG [ghebo@2605:6400:c847:1449::9441] has joined #hplusroadmap
22:32 -!- AMG [ghebo@2605:6400:c847:1449::9441] has quit [Ping timeout: 265 seconds]
22:44 -!- cthlolo [~lorogue@77.33.23.154.dhcp.fibianet.dk] has joined #hplusroadmap
22:52 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has quit [Ping timeout: 256 seconds]
23:01 -!- cthlolo [~lorogue@77.33.23.154.dhcp.fibianet.dk] has quit [Read error: Connection reset by peer]
23:16 -!- stipa [~stipa@user/stipa] has quit [Read error: Connection reset by peer]
23:16 -!- stipa [~stipa@user/stipa] has joined #hplusroadmap
--- Log closed Wed Mar 22 00:00:17 2023