--- Log opened Thu Nov 04 00:00:23 2021 07:44 -!- gnusha [~gnusha@user/gnusha] has joined #hplusroadmap 07:44 -!- Topic for #hplusroadmap: biohacking, nootropics, transhumanism, open hardware | pardon the dust | this channel is LOGGED: http://gnusha.org/logs | https://diyhpl.us/wiki | not quite sponsored by george church | banned by the MIT media lab and the FDA 07:44 -!- Topic set by fenn [~fenn@user/fenn] [Wed May 19 15:37:47 2021] 07:44 [Users #hplusroadmap] 07:44 [ ^ditto ] [ cpopell ] [ fltrz ] [ juri_ ] [ otoburb ] [ sgiath ] 07:44 [ abetusk ] [ darsie ] [ gnusha ] [ kanzure ] [ pasky ] [ sivoais ] 07:44 [ acertain ] [ deltab ] [ greenz1[m]] [ L29Ah ] [ potatope ] [ sknebel ] 07:44 [ andytoshi] [ docl ] [ heath ] [ livestradamus] [ ptrcmd ] [ srk ] 07:44 [ archels ] [ dr-orlovsky] [ hellleshin] [ lsneff ] [ redlegion ] [ strages ] 07:44 [ balrog ] [ dustinm ] [ Hooloovoo ] [ maaku ] [ rndhouse ] [ streety ] 07:44 [ berndj ] [ EnabrinTain] [ HumanG33k ] [ Malvolio ] [ RubenSomsen] [ superkuh] 07:44 [ catalase ] [ faceface ] [ Jenda ] [ mgxm ] [ s0ph1a ] [ TMA ] 07:44 [ chiastre ] [ FelixWeis ] [ johest ] [ nmz787 ] [ saxo ] [ xaete[m]] 07:44 [ Codaraxis] [ fenn ] [ jrayhawk ] [ nsh ] [ SDr ] [ yuanti ] 07:44 -!- Irssi: #hplusroadmap: Total of 60 nicks [0 ops, 0 halfops, 0 voices, 60 normal] 07:44 -!- Channel #hplusroadmap created Wed May 19 06:51:36 2021 07:45 < kanzure> get back in here 07:46 -!- Irssi: Join to #hplusroadmap was synced in 128 secs 07:52 < docl> oh, that's xentrac? cool! 07:52 < docl> he had good ideas 07:57 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0::f2f0] has joined #hplusroadmap 08:32 -!- mrdata [~mrdata@user/mrdata] has joined #hplusroadmap 09:23 -!- CryptoDavid [uid14990@id-14990.uxbridge.irccloud.com] has joined #hplusroadmap 10:39 < nsh> xentrac is only as muurkha atm 10:39 < nsh> *online 10:39 < nsh> have passed on an invitation 11:41 -!- muurkha [~kragen@adjuvant.canonical.org] has joined #hplusroadmap 11:41 < muurkha> hi, I'm kragen/xentrac. just a muurkha 11:42 < muurkha> nsh said someone wanted me? 13:09 -!- mrdata [~mrdata@user/mrdata] has quit [Ping timeout: 268 seconds] 13:42 -!- spaceangel [~spaceange@ip-89-176-181-220.net.upcbroadband.cz] has joined #hplusroadmap 14:00 -!- Llamamoe [~Llamamoe@public-gprs352816.centertel.pl] has joined #hplusroadmap 14:25 < docl> muurkha: I started a thread on twitter and your work came up 14:25 < docl> .title https://twitter.com/lsparrish/status/1456100227103039492 14:26 < saxo> No title found 14:26 < docl> .tw 14:26 < saxo> Anyone know of any good technical proposals for self replicating robotic systems? I'm looking for anything recent // https://twitter.com/jaromczyk/status/1456043697812946950 (@lsparrish) 14:39 < docl> I didn't really have specific questions but thanks for writing it, I'm enjoying the read 15:03 -!- Llamamoe [~Llamamoe@public-gprs352816.centertel.pl] has quit [Quit: Leaving.] 16:08 -!- CryptoDavid [uid14990@id-14990.uxbridge.irccloud.com] has quit [Quit: Connection closed for inactivity] 16:48 < muurkha> which notes in particular are you finding useful? 16:51 < muurkha> I'm glad you're enjoying it 16:57 -!- spaceangel [~spaceange@ip-89-176-181-220.net.upcbroadband.cz] has quit [Remote host closed the connection] 17:00 < muurkha> http://canonical.org/~kragen/dercuano is the Dercuano homepage 18:06 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has quit [Ping timeout: 256 seconds] 18:23 < lsneff> The self-replicating factory concept gets exponentially harder the fewer humans are in the loop. 18:23 < L29Ah> hyperloop? 18:24 < L29Ah> unemployment is record high! get some humans in your loop now! MONEY! 18:24 < lsneff> Once you decide it needs to be able to take a design file, manufacture all the parts, and assemble them without explicit instructions of how to do that, it becomes a hard problem in my eyes. 18:24 < muurkha> lsneff: what are the immediate hard problems to solve to do that? 18:25 < muurkha> from your point of view 18:25 < muurkha> docl: I'm also interested to hear what other things you've found on the subject 18:26 < lsneff> In my view, generating a pathway to assemble a product out of the individual parts is one of the hardest problems. 18:27 < lsneff> Potentially, it could require the factory to generate and use intermediate tools that are bespoke for one kind of product 18:27 < muurkha> Do you mean the toolpaths for the assemblers? 18:27 < muurkha> like, move screwdriver to x=30 y=53 z=100 and then z=50 before starting to rotate? 18:28 < muurkha> I feel like that kind of thing could be part of the design file? 18:28 -!- mrdata [~mrdata@user/mrdata] has joined #hplusroadmap 18:29 < muurkha> also though it sounds pretty tractable to solve automatically 18:29 < lsneff> I guess you could include all those individual instructions I suppose. 18:29 < muurkha> but maybe I'm not understanding the kind of "assemble a product out of the individual parts" you mean 18:30 < lsneff> Well, for instance: suppose you want a cnc machine 18:30 < muurkha> ok 18:32 < muurkha> like the MPCNC, Root CNC, or PrintNC, say 18:33 < lsneff> To make a cnc machine, you need to make motors, and to make motors, it needs to wrap wire. The factory could include a machine to wrap wire, but it’s an example of how many niche tools are needed. There’s a combinatorial explosion of tools and I’d suspect the design of an object would need to include instructions to make and use a tool to wrap wire or do whatever niche thing. 18:34 < muurkha> how does having humans in the loop help with that? 18:34 < lsneff> Sorry, this is less insurmountable the more i think about it 18:35 < lsneff> Humans in the loop would be able to assemble things, but I think you’re right that it’s computationally solvable 18:36 < muurkha> Humans in the loop still end up using a wire-winding tool to make coils though 18:36 < muurkha> https://www.youtube.com/watch?v=ngvfoc2Jo7Y shows a highly developed example of this 18:37 < muurkha> but in some cases people chuck up a nail in their drill or something 18:38 < muurkha> the part about CNC machines that seems especially difficult to me right now is the C 18:40 < muurkha> but maybe that's just because it's what I know the most about :) 18:42 < lsneff> Oh, that’s certainly really hard. Actually impossible right now afaik to fit a way to do semiconductors in a self-contained factory 18:42 < lsneff> I think there are a lot of logistic issues. A CNC is just a good example because it has all the things: computers, motors, lots of structural elements 18:43 < lsneff> Do you design the cnc to have a particular model of microprocessor, does the ISA matter? What if the factory doesn’t have the capability to make that processor? Does the cnc design need to include the design for the microprocessor? Stuff like that 18:48 < muurkha> You might be able to do it without semiconductors or with low-performance semiconductors that aren't economically competitive with silicon 18:49 < muurkha> To make an autotrophic self-replicating factory you do need to include the design for the processor, yes, but it doesn't have to be a microprocessor 18:51 < muurkha> it turns out that there are dozens and dozens of inorganic semiconductors and an unlimited number of organic ones 19:04 < lsneff> There was some news recently about an arm chip that was printed on a plastic semiconductor—I wonder whether that process could be made self-contained enough 19:05 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0::f2f0] has quit [Quit: Leaving] 19:05 < lsneff> I think another problem that I suspect will be harder than it seems will be just how parts get picked up and moved. 19:06 -!- CryptoDavid [uid14990@id-14990.uxbridge.irccloud.com] has joined #hplusroadmap 19:08 < fenn> i guess not providing an online html version is one way to spread your memes around the dark corners of a large number of hard drives... 19:10 < lsneff> Browsers cache pages, an html version would still spread memes to hard drives 19:11 < fenn> nah that stuff gets deleted constantly 19:11 < fenn> this tarball will go into a downloads folder where it will sit for the rest of known time 19:27 < fenn> executing a stored series of movement commands from a data tape doesn't require very much computational power 19:28 < fenn> it's sort of a compression algorithm. the more high level and abstract your movement commands, the more computation you need to do, but the less space it takes up on a tape 19:29 < fenn> at the lowest level you are just recording the train of pulses going into the motors 19:30 < fenn> "oh but i want closed loop feedback!" well there are lots of forms of feedback that _could_ be implemented in manufacturing, but aren't 19:30 < fenn> leather workers have to cut around the brand marks and holes in hides 19:31 < fenn> with a shortage of intelligence available you just have to make do with blind unfeeling processes that rely on uniform inputs in a known state 19:32 < fenn> anyway, early NC machines were really simple 19:33 < fenn> from a control perspective 19:42 < lsneff> More advanced computing capabilities would certainly be necessary down the line though. Perhaps not for basic CNC, but an automated factory would need advanced computing 20:42 -!- johest [~johest@user/johest] has quit [Quit: Ping timeout (120 seconds)] 20:42 -!- johest [~johest@user/johest] has joined #hplusroadmap 20:47 < fenn> not at all obvious. factorio gets by with real dumb automata 20:47 < fenn> i guess it's a matter of designing things with restoring forces built in 20:48 < fenn> lego brick assemblies with friction studs gradually get loose and imprecise, but the snap pin style constructions self maintain in a known state 20:49 < fenn> the principle generalizes 20:50 < fenn> i think humanity underestimates the power of stupid automation 21:21 < lsneff> Factorio hides the hard problems. It has assembler blocks which can basically make anything from parts, including itself. Clearly, that requires computation. 21:24 < fenn> please explain why it requires computation 21:24 < fenn> 99% of the time in factorio you set up the assembler to do one task, and never touch it again 21:24 < fenn> the only unrealistic thing is that it doesn't break down ever 21:25 < fenn> and no bad parts produced 21:35 < Malvolio> (see also #reprap for more self replication except with 3d printers) 23:02 -!- CryptoDavid [uid14990@id-14990.uxbridge.irccloud.com] has quit [Quit: Connection closed for inactivity] --- Log closed Fri Nov 05 00:00:23 2021