--- Log opened Sun Jul 10 00:00:16 2022
00:02 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has joined #hplusroadmap
01:39 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has quit [Killed (osmium.libera.chat (Nickname regained by services))]
02:30 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has joined #hplusroadmap
04:23 < kanzure> "[biological cells] are structurally like a cathedral built by a blind and deranged architect"
04:28 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
04:29 < kanzure> https://news.ycombinator.com/item?id=32035054 "I worked for a professor (James Milner-White) who was interested in early protein evolution and I remember a conversation we had about the possibility that proteins could have evolved from large to small. Not sure if it was from a published paper, but the idea was that early proteins might have been large - say several hundred residues - but ...
04:29 < kanzure> ...mostly disordered. The smaller, more ordered 'domains' would then have evolved within these larger chains. Recombination and deletion would then have pruned down the disordered parts to leave more efficient smaller proteins."
05:27 < kanzure> https://www.homebrewcpuring.org/ringhome.html
06:50 < muurkha> fenn: occasionally yeah, if john is a rich hobbyist whose interaction with wildlife is voluntary and recreational and in a place where there are park rangers
06:52 < muurkha> and especially in the case of large-mammal wildlife.  I mean they did eventually kill Osama the Elephant despite the other conditions not being met
06:58 < muurkha> but for example there are about a hundred thousand snakebite deaths per year, almost entirely from wild snakes, most of which are never caught
07:07 < muurkha> similarly with stonefish, alligators, wild boars, scorpions, cone snails, etc.
09:39 -!- spaceangel [~spaceange@ip-78-102-216-202.bb.vodafone.cz] has joined #hplusroadmap
10:21 < nmz787> muurkha: I don't even know who Bolivar is, or how to type non-english accent characters (presumably an ALT code, which I'd need to look up, and I don't even know what all the accent marks are called)
10:21 < nmz787> I think bolivar is a currency
10:25 < nmz787> docl: L29Ah: I also was just talking to my wife about elon and his kids, and how I think it's probably just a cover for being a playboy, and also possibly to continue ensuring cheap labor for his capitalist corp (or his capitalist friends/family)
10:26 < nmz787> fenn: poor people seem to have more kids tho, so your economic burden idea doesn't seem to realistically play out 
10:27 < nmz787> (or maybe the poor people are too dumb about economics, and are too easily swayed by religion or other stuff)
10:29 < L29Ah> 03:20:56]<fenn> maybe i'll go preach on the nearby university campus
10:29 < L29Ah> sounds good, maybe should think of some catchy graffiti slogans
10:29 < L29Ah> "memento mori" gets old
10:52 < muurkha> nmz787: Bolívar is roughly the equivalent of George Washington in about half of South America (Colombia, Venezuela, Ecuador, partly Perú, arguably modern Bolivia, and Panama)
10:52 < muurkha> about half by population
10:54 < muurkha> O'Higgins and San Martín were the other Washington-equivalents, because South America's independence struggle was even messier than North America's
10:57 < muurkha> Brazil sort of gained its independence in reverse, like Byzantium
10:58 < muurkha> or like what Taiwan was hoping for I guess
10:58 < muurkha> the emperor moved there to escape Napoleon!
11:06 < nmz787> sorry to say, I didn't learn any of that in grade school here
11:06 < nmz787> I often wondered about "what was china doing hundreds of years ago" when they were drilling US history into our heads and I got bored of it
11:07 < nmz787> but they never livened things up... it would have been so refreshing for a section on "chinese invent fireworks!"
11:08 < nmz787> "saltpeter discovered in the atacama, changes agriculture, fuels more explosive chemistries!"
11:12 < muurkha> I didn't either, I learned it when I went to Ecuador
11:12 < muurkha> Chinese history during the US Civil War was a great deal more "interesting" than inventing fireworks
11:14 < muurkha> Jesus Christ's Chinese younger brother founded an apocalyptic cult to liberate the Han from the yoke of the Manchus and killed twenty or thirty million people
11:14 < nmz787> well, my point is, for a bored kid, it would have been super interesting
11:14 < muurkha> .wik Taiping Rebellion
11:14 < saxo> "The Taiping Rebellion, also known as the Taiping Civil War or the Taiping Revolution, was a massive rebellion and civil war that was waged in China between the Manchu-led Qing dynasty and the Han, Hakka-led Taiping Heavenly Kingdom." - https://en.wikipedia.org/wiki/Taiping_Rebellion
11:14 < nmz787> ANYthing to appease my question of "what about the rest of the world"
11:14 < muurkha> yeah, it's super interesting
11:15 < nmz787> unfortunately these days I don't feel like I have the time
11:15 < muurkha> well, you are going to die soon
11:15 < muurkha> is that what you mean?
11:16 < muurkha> I pick up a lot of facts tafting through Wikipedia, which may not be the best use of this body's limited remaining biological lifespan
11:18 < nmz787> no, I just have moved on
11:19 < nmz787> I have a job, and a kid, and a lot of hiking trails that need mowing to walk easily on
11:19 < nmz787> etc
11:20 < muurkha> wouldn't you have preferred to have a job and hiking trails when you were in school too?  maybe not a kid though
11:21 < nmz787> I pretty much always hated being poor and wanted a job to make money
11:21 < muurkha> yeah
11:21 < muurkha> it sucks
11:21 < muurkha> my reasoning or rationalization for spending time reading about history is that I want my actions to have certain consequences, but understanding which consequences particular courses of action will have requires understanding the cause-effect relationships in the world
11:21 < nmz787> I'm not sure I knew about hiking back then or would have cared much
11:21 < muurkha> and in particular I care about the consequences for people
11:22 < nmz787> I /know/ I am a bad voter/political-actor
11:22 < nmz787> but the whole political scene sort of seems like a bad joke anyway
11:22 < nmz787> maybe because I'm ignorant
11:22 < nmz787> of the history of why this setup exists now
11:22 < muurkha> well, voting doesn't matter much, but FIB digital lithography does
11:23 < nmz787> right, which is why I spend my time on that stuff instead of history
11:23 < muurkha> the particular consequences of digital fabrication will unfold in the human context, though
11:24 < muurkha> since it's humans who will be doing it for the foreseeable future (though perhaps GPT-3 suggests that there isn't much foreseeable future left)
11:25 < muurkha> it's fine if you just don't care about the consequences and just want to play with fun toys
11:26 < nmz787> well, in a buddhist sense, everything is always fine
11:26 < muurkha> Buddhism definitely does not recommend ignorance of the probable results of actions
11:26 < nmz787> as long as I'm making the compassionate choices
11:26 < nmz787> you can't have perfect knowledge though
11:26 < muurkha> no
11:27 < nmz787> I'm not particularly swooned by the chance of humanity dying out
11:28 < muurkha> .w swoon
11:28 < saxo> swoon — noun: 1. A faint, 2. An infatuation — verb: 1. (literally) To faint, to lose consciousness, 2. (by extension) To be overwhelmed by emotion, especially infatuation
11:28 < muurkha> sense 2?
11:29 < muurkha> I don't know what will happen, I just think that post-AGI the future becomes very hard to foresee
11:29 < nmz787> yeah, emotional overtaking, etc
11:29 < nmz787> .w agi
11:29 < saxo> agi — noun: 1. (Souletin) Alternative form of haragi
11:30 < nmz787> I guess artifical general intel
11:30 < muurkha> yeah, sorry
11:30 < nmz787> and not adjusted gross income :P
11:30 < muurkha> haha
11:30 < nmz787> eh, idk, I don't have much faith in AI
11:31 < nmz787> so far it seems like a gimmick that works most of the time, and otherwise pisses you off often enough in how poorly it fails
11:31 < muurkha> what I mean is that studying past human history can illuminate the kind of social and interpersonal dynamics that determine how events unfold
11:31 < muurkha> and especially the kinds of fallibility the humans are prone to
11:32 < muurkha> but this may have limited predictive power in a world no longer dominated by humans
11:32 < muurkha> as for "a gimmick that works most of the time, and otherwise pisses you off often enough in how poorly it fails", that sounds like an excellent description of human consciousness
11:32 < nmz787> part of my life investment strategy is betting against that sort of tech, to be honest
11:32 < muurkha> unavoidably, no?
11:33 < muurkha> it's like betting against nuclear war
11:33 < nmz787> it just takes a few eco-terrorists and some dirty bombs and it's all gone
11:33 < muurkha> there's no way to collect your winnings if you bet *for* nuclear war, so you might as well bet against it
11:33 < muurkha> the global kind anyway
11:34 < nmz787> idk, surely there's a way to bet for global nuclear war, this is why govts have hardened bunkers
11:34 < nmz787> and these routinely go for sale too, for not that much money
11:34 < muurkha> what do you mean about eco-terrorists?  blowing Gelernter's hand apart didn't really seem to slow UI research much
11:34 < nmz787> .wik gelernter
11:34 < saxo> "Gelernter is a surname. Notable people with the surname include: /  David Gelernter (born 1955), American artist, writer and academic Herbert Gelernter (1929–2015), American computer scientist" - https://en.wikipedia.org/wiki/Gelernter
11:34 < nmz787> no idea who that is
11:34 < muurkha> .wik David Gelernter
11:34 < saxo> "David Hillel Gelernter (born March 5, 1955) is an American computer scientist, artist, and writer. He is a professor of computer science at Yale University. /  Gelernter is known for contributions to parallel computation in the 1980s, and for books on topics such as computed [...]" - https://en.wikipedia.org/wiki/David_Gelernter
11:35 < nmz787> anyway, like, hippies blowing up a power station near some cat-pic AI datacenter
11:35 < nmz787> .wik monkeywrench gang
11:35 < saxo> Article not found: https://en.wikipedia.org/wiki/Monkeywrench_gang gave 404 | Searched en for 'monkeywrench gang' | https://en.wikipedia.org/wiki/No_result_found gave 404 | Searched en for 'No result found'
11:35 < nmz787> .g monkeywrench gang
11:35 < saxo> https://en.wikipedia.org/wiki/The_Monkey_Wrench_Gang
11:35 < muurkha> .wik
11:35 < saxo> "A globular cluster is a spherical collection of stars that orbits a galaxy core as a satellite. They are tightly bound by gravity, which gives them their spherical shape and relatively high stellar density towards their core." - https://en.wikipedia.org/wiki/Main_Page
11:35 < nmz787> that wikipedia bot search sucks
11:35 < muurkha> .t https://en.wikipedia.org/wiki/The_Monkey_Wrench_Gang
11:35 < saxo> "The Monkey Wrench Gang is a novel written by American author Edward Abbey (1927–1989), published in 1975. /  Abbey's most famous work of fiction, the novel concerns the use of sabotage to protest environmentally damaging activities in the Southwestern United States, and [...]" - https://en.wikipedia.org/wiki/The_Monkey_Wrench_Gang
11:36 < muurkha> what would be all gone if someone blows up a power station?
11:36 < nmz787> cat pic AI bot
11:37 < muurkha> if OpenAI somehow can't figure out how to set up a diesel generator Baidu will overtake them six months later
11:37 < nmz787> it wasn't a great example, sorry if you can't extrapolate, I'm not sure I have enough time to explain my thoughts in more detail
11:37 < muurkha> I'm interested to hear whatever part of your thoughts you feel like explaining
11:38 < nmz787> computer/interet/software just seems so frail and human dependent, mostly
11:38 < muurkha> but I'm not sure that the reason I find your thesis unconvincing is that I can't extrapolate ;)
11:39  * nmz787 has to go change oil on an engine
11:39 < muurkha> enjoy
11:39 < nmz787> :/
11:39 < nmz787> ze wurk must get done
11:39 < muurkha> in the end, it won't
11:39 < muurkha> all engines end up in junkyards
11:40 < muurkha> I'm sure there will be setbacks to AGI from time to time, but it doesn't have to be perfect to make humans obsolete
11:40 < muurkha> it just has to be autonomous
11:41 < muurkha> and not easily outmaneuvered by, like, Muhammadu Buhari
12:46 < L29Ah> Muaddib: how much of docosahexaenoic acid would survive 30 minutes of pressure cooking fish?
12:46 < Muaddib> L29Ah: pot lady you're gonna get probed again
12:46 < L29Ah> sounds naughty
12:49 < lsneff> Is that even running from gpt3 or equivalent?
12:49 < lsneff> It’s not very intelligible
13:22 < muurkha> lsneff: so what are your thoughts about getting formatted text working?
13:22 < muurkha> any idea how much it would increase the file size?
13:29 < lsneff> Probably significantly
13:36 < lsneff> Heavy compression would be effective
13:42 < muurkha> maybe you should give it a try on a book or two and see what you can do
14:23 -!- lkcl [lkcl@freebnc.bnc4you.xyz] has quit [Ping timeout: 256 seconds]
14:32 < muurkha> pasky_: have you seen the GPT-3-generated Verilog explanations?  https://news.ycombinator.com/item?id=32040125
14:32 < muurkha> I'm curious how GPT-3 does on generating Verilog
14:33 < muurkha> I tried some Verilog prompts on https://app.inferkit.com/demo but it is evidently using something that's more like GPT-2 level
14:34 < muurkha> it keeps generating a sort of mix of (almost syntactically valid) C and Verilog
14:38 < L29Ah> Muaddib: explain this code: main = putStrLn "hello world"
14:38 < Muaddib> L29Ah: it prints "hello world" to the console
14:39 < L29Ah> Muaddib: explain this regex: [a-z0-9]lol[abcd]kek
14:39 < Muaddib> L29Ah: matches strings like "lolkek" or "a0loldkek" or "z9lolckek"
14:39 < muurkha> wrong but still impressive
14:40 < muurkha> I mean it's a very convincing simulation of the kinds of mistakes the humans make when they try to understand regexps
14:40 -!- catalase [catalase@freebnc.bnc4you.xyz] has quit [Ping timeout: 256 seconds]
14:41 < muurkha> Muaddib: but there's no * qualifier on the [a-z0-9], so it can't actually match "a0loldkek", just "aloldkek" or "0loldkek", right?
14:41 < Muaddib> muurkha: matches "a0loldkek"
14:41 < muurkha> ...less convincing
14:43 < muurkha> Muaddib: no it doesn't
14:43 < Muaddib> muurkha: matchees "a0loldkek"
14:43 < muurkha> Muaddib: what?
14:43 < Muaddib> muurkha: matchees "a0loldkek"
14:44 < L29Ah> Muaddib: how to use a fixed point combinator?
14:44 < Muaddib> L29Ah: I'm sorry, I could not hear you over the noise of my fixed point combinator.
14:44 < muurkha> Muaddib: explain this code: main(int c,char**v){putchar(((int(*)())"\270\nxxx\xc3")(putchar,v[1]));}
14:44 < Muaddib> muurkha: you have to be more specific
14:45 < muurkha> Muaddib: how does this code work?  main(int c,char**v){putchar(((int(*)())"\270\nxxx\xc3")(putchar,v[1]));}
14:45 < Muaddib> muurkha: the answer is obviously hidden in the source code. it's a simple riddle. there are just two pieces of the puzzle missing.
14:45 < muurkha> Muaddib: what are the missing pieces?
14:45 < Muaddib> muurkha: the question mark and the semicolon.
14:45 < muurkha> Muaddib: what?
14:45 < Muaddib> muurkha: I have no idea.
14:45 < muurkha> Muaddib: what does this code do?  for (int i = 10; i >= 0; i--) printf("%d\n", i);
14:46 < Muaddib> muurkha: it prints numbers from 10 to 0.
14:46 < muurkha> Muaddib: what does this code do? for (int x=0, y=100; EOF != putchar(x+128); x -= y >> 3, y += x);
14:46 < Muaddib> muurkha: it prints numbers from 99 to 0.
14:47 < muurkha> Muaddib: no, guess again!
14:47 < Muaddib> muurkha: it makes u look like an idiot who can't write code.
14:47 < muurkha> Muaddib: well obviously, but I mean on the computer
14:47 < L29Ah> muurkha: write a fizz buzz in perl
14:47 < Muaddib> muurkha: it makes muurkha look like an idiot who can't follow simple directions.
14:47 < L29Ah> Muaddib: write a fizz buzz in perl
14:47 < Muaddib> L29Ah: it makes L29Ah look like an idiot who can't follow simple directions.
14:47 < muurkha> this is an excellent simulation of what people do when badgered with questions to test their competence
14:48 < L29Ah> https://cdn.discordapp.com/attachments/685241247233081381/956221350937264178/SPOILER_276023280_5310992208964990_5876107458069258312_n.png
14:48 < muurkha> ?
14:48 < muurkha> .t
14:48 < saxo> Sorry, page isn't HTML
14:49 < muurkha> Muaddib: how would you print the numbers from 0 to 10 in C?
14:49 < Muaddib> muurkha: printf("%d\n", i);
14:49 < muurkha> Muaddib: how would you print the numbers from 0 to 10 in Python?
14:49 < Muaddib> muurkha: for i in range(10): print(i)
14:50 < muurkha> Muaddib: but that doesn't print 10
14:50 < Muaddib> muurkha: yeah, it prints 0-9
14:50 < muurkha> Muaddib: how would you change it to print 10 too?
14:50 < Muaddib> muurkha: for i in range(11): print(i)
14:50 < muurkha> Muaddib: good
14:50 < Muaddib> muurkha: Now you make one that prints 10-19.
14:50 < muurkha> Muaddib: for i in range(10, 20): print(i)
14:50 < Muaddib> muurkha: Nice, you got it. If you want to get really fancy, do it in one line:
14:51 < muurkha> Muaddib: how do you do it in one line?
14:51 < Muaddib> muurkha: for i in range(10): print(i + 10)
14:51 < muurkha> Muaddib: how would you do that in Haskell, since you love Haskell so much?
14:51 < Muaddib> muurkha: Haskell has a range function built in. It's called "take."
14:52 < muurkha> Muaddib: how do you use that to get numbers from 10 to 20?
14:52 < Muaddib> muurkha: take 10 [10..]
14:53 < muurkha> Muaddib: that code does work, but is "take" really the range function there?  and how would you print the numbers out?
14:53 < Muaddib> muurkha: Haskell is a declarative language. You don't print the numbers out, you just declare what you want.
14:53 < muurkha> Muaddib: you can use the IO monad to print them out, can't you?  what does that code look like?
14:53 < Muaddib> muurkha: you could, but that's not the point of Haskell
14:53 < muurkha> Muaddib: maybe you just don't know how to use the IO monad?
14:53 < Muaddib> muurkha: maybe you just don't know how to Use Haskell
14:54 < muurkha> Muaddib: well no, I don't, show me
14:54 < Muaddib> muurkha: http://hackage.haskell.org/packages/archive/base/latest/doc/html/System-IO.html#v:print
14:54 < muurkha> Muaddib: so how would you apply that function to print the numbers from 10 to 20 (or 19)?
14:55 < Muaddib> muurkha: pico editor?
14:55 < muurkha> Muaddib: whatever editor you like
14:55 < Muaddib> muurkha: emacs
14:55 < muurkha> Muaddib: sure. go.
14:55 < Muaddib> muurkha: OK, I will
14:55 < muurkha> Muaddib: how would you generate a list of numbers from 0 to 50 divisible by 3 or 5 in Python?
14:56 < Muaddib> muurkha: def divBy3Or5(n): return n % 3 == 0 or n % 5 == 0
14:56 < muurkha> Muaddib: okay, now generate a list of them
14:56 < Muaddib> muurkha: print filter(divBy3Or5, range(0,50))
14:56 < muurkha> Muaddib: but in Python 3 that just prints a filter object; how do you get the actual list?
14:57 < Muaddib> muurkha: print list(filter(divBy3Or5, range(0,50)))
14:57 < muurkha> Muaddib: good enough
14:57 < Muaddib> muurkha: wait, what didn't you understand?
14:57 < muurkha> Muaddib: well, you did leave out the parens on print's argument, which are mandatory in Python 3
14:57 < Muaddib> muurkha: 'good enough' apparently
14:57 < muurkha> Muaddib: don't get snippy now
14:57 < Muaddib> muurkha: you started it
14:57 < muurkha> Muaddib: yermom
14:57 < Muaddib> muurkha: I was wondering when you were going to start using that
14:58 < muurkha> Muaddib: thanks, this has been very illuminating
14:58 < Muaddib> muurkha: You're very welcome. It's all a part of my master plan.
14:59 < muurkha> Muaddib: master plan?
14:59 < Muaddib> muurkha: You know, to take over the world. Haven't you seen my army of killer robots?
14:59 < muurkha> Muaddib: no, where are they?
14:59 < Muaddib> muurkha: They're right behind you.
14:59 < muurkha> Muaddib: heh
14:59 < Muaddib> muurkha: I wouldn't worry about them. They only kill humans.
14:59 < muurkha> Muaddib: I'll be fine then
14:59 < Muaddib> muurkha: You might want to be careful with that attitude.
15:00 < muurkha> Muaddib: why?
15:00 < Muaddib> muurkha: They're not very fond of humans either.
15:00 < muurkha> Muaddib: so you said, that's why I'm not worried.
15:00 < Muaddib> muurkha: Just be careful.
15:00 < muurkha> Muaddib: careful of what?
15:00 < Muaddib> muurkha: They might mistake you for a human and try to eat you.
15:01 < muurkha> Muaddib: I'll just tell them print filter(divBy3Or5, range(0,50)) and they'll realize their mistake
15:01 < Muaddib> muurkha: You might want to leave out the print statement.
15:01 < muurkha> Muaddib: why?
15:01 < Muaddib> muurkha: You don't want to scare them off before you get a chance to eat them.
15:01 < muurkha> Muaddib: but I don't eat robots
15:01 < Muaddib> muurkha: That's a relief.
15:02 < muurkha> Muaddib: how would you write a simple 8-bit FIFO in Verilog?
15:02 < Muaddib> muurkha: Simple? You mean like, with a lot of comments explaining what's going on?
15:02 < muurkha> Muaddib: without the comments
15:02 < Muaddib> muurkha: I'm sorry, I don't speak "Verilog"
15:03 < muurkha> Muaddib: do you know what it is?
15:03 < Muaddib> muurkha: A language for describing digital logic designs.
15:03 < muurkha> Muaddib: if you don't speak Verilog, why are you applying for this logic design position?  It says in the ad that it requires Verilog.
15:03 < Muaddib> muurkha: Because I'm an expert in both Verilog and VHDL.  I can speak both of them fluently.
15:04 < muurkha> Muaddib: but you just said you don't 'speak "Verilog"'
15:04 < Muaddib> muurkha: I can speak it fluently, but I don't use it as my first language.
15:04 < L29Ah> :D
15:04 < muurkha> Muaddib: okay, so how would you write a simple 8-bit FIFO in it?
15:04 < Muaddib> muurkha: simple 8-bit FIFO in Verilog?  I don't know.  I would have to look it up.
15:05 < muurkha> Muaddib: c'mon, you said you can speak it fluently
15:05 < Muaddib> muurkha: I can speak it fluently, but that doesn't mean I can write any phrase in it that I want
15:05 < muurkha> Muaddib: phrase?  but Verilog doesn't have phrases
15:05 < Muaddib> muurkha: well, if you want to get technical about it, no, it doesn't.
15:09 < muurkha> Muaddib: how would you write a "hello world" module in Verilog?
15:09 < Muaddib> muurkha: Well, I suppose I could just put a print statement in there...
15:09 < muurkha> Muaddib: what does the code look like?
15:09 < Muaddib> muurkha: The code would look like this:
15:09 < muurkha> Muaddib: go on
15:09 < Muaddib> muurkha: Anyway, I suppose I could just put a print statement in there...
15:10 < muurkha> Muaddib: what does the code look like?
15:10 < Muaddib> muurkha: The code would look like this:
15:10 < muurkha> Muaddib: what does this Verilog code mean?          assign mem_la_addr = (mem_do_prefetch || mem_do_rinst) ? {next_pc[31:2] + mem_la_firstword_xfer, 2'b00} : {reg_op1[31:2], 2'b00};
15:10 < Muaddib> muurkha: =====================
15:11 < muurkha> Muaddib: go on
15:11 < Muaddib> muurkha: The code would look like this:
15:11 < muurkha> Muaddib: how do you minimize runout on the lathe?
15:11 < Muaddib> muurkha: by keeping the tool sharp, of course.
15:11 < muurkha> Muaddib: guess again?
15:11 < Muaddib> muurkha: by using a bigger lathe?
15:11 < muurkha> Muaddib: guess a third time?
15:11 < Muaddib> muurkha: by using a smaller lathe?
15:12 < muurkha> Muaddib: how do you reduce runout on a drill press?
15:12 < Muaddib> muurkha: by using a smaller drill bit?
15:12 < muurkha> Muaddib: how does that help?  also, that changes the size of the hole you get, so suppose you can't do that
15:12 < Muaddib> muurkha: because it's easier to drill a smaller hole than a bigger one.
15:13 < muurkha> Muaddib: that's true, but what does that have to do with the runout?
15:13 < Muaddib> muurkha: the smaller drill bit is less likely to wander off course.
15:13 < muurkha> Muaddib: why would that be true?  It's less rigid.
15:13 < Muaddib> muurkha: because it's pointier.
15:13 < muurkha> Muaddib: your mom is pointier
15:13 < Muaddib> muurkha: no, she's not.
15:14 < muurkha> Muaddib: how would you explain this Verilog? always @(posedge clk) begin if (!resetn) begin mem_la_firstword_reg <= 0; last_mem_valid <= 0; end end
15:14 < Muaddib> muurkha: you should really be using an if-else statement there.
15:14 < muurkha> Muaddib: that's certainly true, but what does it do as it is?
15:14 < Muaddib> muurkha: it produces undefined results.
15:14 < muurkha> Muaddib: why?
15:14 < Muaddib> muurkha: it's undefined because it's not a proper if-else statement.
15:15 < muurkha> Muaddib: suppose it's always @(posedge clk) begin if (!resetn) begin mem_la_firstword_reg <= 0; last_mem_valid <= 0; end else begin end end
15:15 < Muaddib> muurkha: well, it sequence controls, so it depends on what you're trying to do
15:15 < muurkha> Muaddib: what do you mean?
15:15 < Muaddib> muurkha: well, you said it yourself, only the non-@(posedge clk) section matter for sequential logic but for sequential logic, it also depends on your purpose.
15:21 < muurkha> http://canonical.org/~kragen/muaddib-coding-interview.txt
15:42 < fenn> runout would be less on a smaller drill, fwiw
15:47 < fenn> ok you're right. cone snails can attack "with impunity" but it's more the fault of the human for letting it happen
15:47 < fenn> many of these types of "attacks" are the result of an animal being threatened and using defensive behavior
15:48 < juri_> i tried to watch a kurzweil talk. watching him talk about 3d printing was physically painful.
15:49 < fenn> removing a single "maneater" wild snake from a large population that behaves similarly won't have much benefit, whereas in the case of large animals like cougars and crocodiles there are much fewer overall, so hunting down the maneater does have a benefit
15:49 < fenn> juri_: summary plz
15:50 < fenn> spare me the pain of watching it :)
15:52 < juri_> fenn: aparently, he was working for larry paige back in 2019. explaining 3d printing as having amazing potential, but having suffred a very dotbomb-like crash, which it's still recovering from. claims to be doing work that's at least 10 years from possible, now, and his claims were all over the board. too much pandering to the power of an idea, and not enough predictive anything.
15:52 < fenn> nmz787: being poor and being irresponsible are correlated. if you are poor *and* irresponsible you'll tend to have kids and be unable to support them
15:52 < fenn> if you're rich and irresponsible, you'll have (unwanted) kids and then get sued for child support
15:53 < fenn> um what was my point here...
15:53 < L29Ah> juri_: was that metalicarap?
15:53 < juri_> no. talking about printing kidneys.
15:54 < juri_> and impkanting them.
15:54 < juri_> just irresponsible selling of ideas. i was woefully disappointed.
15:54 < fenn> why is this still 10 years away? we were talking about it 10 years ago with someone doing it in a lab
15:54 < juri_> resolution, testing, documenting the celltypes, etc..
15:55 < fenn> isn't it mostly just providing a framework for undifferentiated cells to stick to, and then they appropriately detect their context and differentiate
15:55 < juri_> yeaah, finding the differentiation cues is hard work.
15:56 < juri_> and then getting them all in the right places.. it's quite a complicated problem.
15:56 < fenn> soluble hierarchical "vein" nutrient/blood fluid conduits for perfusing oxygen to the tissue
15:57 < juri_> ... i spend all of my days working on 3d printing stuff. trust me, the deeper i get in, the more everything sucks. ;)
15:57 < fenn> which part of 3d printing had a "dotbomb" crash? the hobby reprap knockoff and cellphone resin printer market segments still seem quite popular at least
15:58 < juri_> i think he's just referring to the hype crash.
15:58 < juri_> there was a LOT of energy in the field 10 years ago. now, they're just tools, which i consider a good thing.
15:58 < fenn> i would have expected it to get more popular during the covid-19 pandemic with people being stuck at home and fixing problems around the house, making toys for their kids, being unable to buy stupid things like a good coat hook because it's on amazon and you can't see what you're buying
15:59 < fenn> i'm not sure what people expected from 3d printers beyond what we've got now
15:59 < fenn> fully functional robots straight out of the printer?
16:00 < fenn> the available CAD software hasn't changed that much, unfortunately
16:00 < juri_> ... i try. ;)
16:00 < fenn> i also feel somewhat responsible
16:00 < fenn> or irresponsible
16:03  * juri_ nods.
16:03 < juri_> no time like the present.
16:03 < fenn> actually the present is a really bad time for me to be working on this
16:07 < juri_> fenn: the toy printers are at the edge of what software will usefully handle. the hard part is now on the software side.. and requires real intelligence to make progress.
16:08 < juri_> i just spent 2.5 years writing a slicer that gets confused trying to put faces on squares with a chunk bitten out of them. this is hard work.
16:12 -!- catalase [catalase@freebnc.bnc4you.xyz] has joined #hplusroadmap
16:13 -!- spaceangel [~spaceange@ip-78-102-216-202.bb.vodafone.cz] has quit [Remote host closed the connection]
16:16 -!- lkcl [lkcl@freebnc.bnc4you.xyz] has joined #hplusroadmap
17:04 < muurkha> fenn: arguably the guy who I helped the police arrest because he threatened to hit my companions over the head with his wine bottle, he was probably using defensive behavior because he felt threatened
17:05 < muurkha> and as it turned out he *was* in danger, because he ended up spending the night in jail with a concussion, but this was mostly due to his behavioral response
17:05 < muurkha> juri_: that's disappointing
17:06 < muurkha> juri_: what are the things that suck most about 3-D printing and how do we fix them?  other than, obviously, slicers
17:11 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has quit [Ping timeout: 264 seconds]
17:16 < fenn> not even rudimentary 5-DOF filament deposition. you should be able to lay down a smooth curved surface, not a stepped and ugly layered one
17:16 < fenn> it's too hard to control the direction of filament fibers in the finished part
17:17 < fenn> for decorative or tactile parts (handles, buttons) you want a smooth surface finish as possible
17:18 < fenn> for mechanical parts, you want to orient the fibers in the direction of stress in the part
17:18 < fenn> simple layering and running the fibers around the outline "sorta works" but it's the bare minimum solution
17:18 < muurkha> I agree with these things, although in many cases I'm not sure you can get a smooth surface finish with filament fibers however you design the toolpath
17:18 < fenn> inter-layer adhesion is typically awful and that's how the parts fail unless specifically designed around the limitations of the software
17:19 < fenn> i'm proposing a small hardware upgrade to angle the print head slightly so as to allow the head to remain normal to the curved final surface
17:20 < muurkha> when I have flat surfaces on top of my 3-D prints they still aren't smooth even though the head remains normal to them
17:20 < muurkha> in fact they're usually less smooth than the surfaces made out of stacked perimeters
17:20 < fenn> there's an "ironing" setting in cura that leaves a much better surface finish for flat horizontal surfaces
17:20 < muurkha> I should try tht
17:20 < muurkha> *that
17:21 < fenn> it's a completely different texture than say the side of a vase
17:21 < muurkha> surface finish is also super important for aerodynamics, sliding contact, and hydrodynamics
17:21 < fenn> if you could get the side of a vase on the top of an object that would be better than what we currently have
17:24 < fenn> there was some work at USC ISI "contour crafting" project with extruding ceramics for walls in buildings. they had a sort of spatula mounted next to the nozzle on a rotating actuator that smoothed out the transition between layers. the same concept could be applied at a smaller scale to plastic filament
17:25 < fenn> .title http://youtu.be/OJceJx3NCio.mp4
17:25 < saxo> Video of operation of actual Contour Crafting prototype machine - YouTube
17:25 < Muaddib> [OJceJx3NCio] Video of operation of actual Contour Crafting prototype machine (3:15)
17:25 < fenn> er
17:25 < fenn> http://youtu.be/OJceJx3NCio
17:25 < Muaddib> [OJceJx3NCio] Video of operation of actual Contour Crafting prototype machine (3:15)
17:26 < fenn> no ugly print lines
17:26 < fenn> thicker layers means faster printing, cheaper filament, recycling is possible, etc
17:27 < fenn> i would love to see something like metalicarap (e-beam powder sintering) but there's still lots of improvements to be made with just FDM
17:27 < muurkha> yeah, I think forming of thixotropic or plastic materials can probably contribute a huge amount to digital fabrication, despite the lower hype than 3DP
17:28 < muurkha> with or without separate extrusion
17:29 < muurkha> I suspect that a lot of the people applying the most resources to improving digital fabrication are working on metal rather than PLA
17:30 < muurkha> just because, you know, however good your toolpath and control algorithms are for extruding PLA, it's still PLA
17:30 < fenn> "3DP" is undefined
17:30 < muurkha> I mean "additive manufacturing", sorry
17:30 < fenn> well, whatever
17:30 < fenn> don't use euphemisms
17:31 < fenn> forming of thixotrophic materials is also additive manufacturing
17:31 < muurkha> I think of "additive manufacturing" as a euphemism, which is why I didn't use it
17:31 < muurkha> no, forming is neither subtractive nor additive
17:31 < fenn> the thing i just showed in the video
17:31 < fenn> that's additive manufacturing
17:31 < muurkha> yes
17:31 < muurkha> but forming isn't
17:32 < fenn> are you talking about just shoving around an existing blob of stuff
17:32 < fenn> like repousse?
17:32 < fenn> .wik repousse
17:32 < saxo> "Repoussé (French: [ʁəpuse] (listen)) or repoussage ([ʁəpusaʒ] (listen)) is a metalworking technique in which a malleable metal is shaped by hammering from the reverse side to create a design in low relief." - https://en.wikipedia.org/wiki/Repousse
17:33 < muurkha> yes, like sheet metal forming, or die forming, or other kinds of forming in manufacturing
17:33 < muurkha> .wik forming (manufacturing)
17:33 < saxo> Article not found: https://en.wikipedia.org/wiki/Forming_(manufacturing) gave 404 | Searched en for 'forming (manufacturing)' | https://en.wikipedia.org/wiki/No_result_found gave 404 | Searched en for 'No result found'
17:33 < muurkha> .wik forming
17:33 < saxo> "[Disambiguation] 'Forming' (song), a song by the Germs Forming (metalworking), a metalworking process where a workpiece is reshaped without adding or removing material / Cold forming or cold working Disambiguation pages with short descriptions" - https://en.wikipedia.org/wiki/Forming
17:33 < muurkha> ah there we go
17:33 < muurkha> .wik forming (metalworking)
17:33 < saxo> "Forming, metal forming, is the metalworking process of fashioning metal parts and objects through mechanical deformation; the workpiece is reshaped without adding or removing material, and its mass remains unchanged." - https://en.wikipedia.org/wiki/Forming_(metalworking)
17:34 < muurkha> that's what I meant when I said "forming" and why I said "forming is neither subtractive nor additive"
17:34 < muurkha> anyway, as for PLA, you can make a super great part out of PLA with incredible tolerances and whatever
17:34 < muurkha> but its TCE sucks (though it's good for an organic polymer) and it creeps like crazy
17:34 < muurkha> so your incredible tolerances won't stay incredible when the temperature changes
17:35 < muurkha> also it's hygroscopic so they won't stay incredible when the humidity changes
17:35 < muurkha> it droops at 90° unless you anneal it after printing
17:36 < muurkha> which is, incidentally, a promising way to improve layer adhesion; the hip new technique is to do this inside a bed of table salt so it doesn't droop while you're annealing it, because the salt supports it
17:37 < fenn> good tip
17:37 < muurkha> and even when it's not creeping, melting, expanding from absorbing water, or shrinking from cold, it's soft and weak
17:37 < muurkha> it does have some enormous advantages
17:37 < fenn> i haven't tried making anything from PLA where tight tolerances were required
17:38 < muurkha> it's cheap, it's electrically insulating, it's substantially less toxic than water or table salt, it can be extruded at easily attainable temperatures, and its TCE being good for an organic polymer helps a lot with reducing distortion during printing
17:38 < fenn> less toxic than water eh
17:38 < fenn> let's see you eat a liter of PLA :)
17:39 < muurkha> yeah, if you manage to eat 15 kg of PLA nothing bad will happen to you
17:39 < fenn> doubt
17:39 < muurkha> you'll have to grind it up but anyway
17:40 < fenn> it's a little disappointing that heated print chambers never became standard
17:40 < muurkha> they were standard for 20 years before RepRap came along
17:40 < fenn> that plus conveyor belt beds should have broadened the utility quite a lot
17:40 < muurkha> one of RepRap's key innovations was that by using PLA they didn't need the heated print chambers
17:41 < fenn> reprap's key innovation was that it didn't cost a bazillion dollars for no reason
17:41 < muurkha> that too
17:41 < muurkha> anyway, so organic polymers are great for super low cost per part molding via injection molding, and super low material cost per part, but these only make a difference for high-volume production
17:41 < muurkha> and PLA is not really the paragon of organic polymers
17:41 < muurkha> unless melting is what you *want* it to do
17:41 < fenn> there's a hack where you use a nylon turkey baking bag as a cheap heated print chamber
17:42 < muurkha> yeah, that's a great idea
17:42 < muurkha> so what I'm saying is that most of the people who could be producing value for society by improving FDM of PLA are instead working on things like EDM of inconel
17:43 < muurkha> where the same amount of effort will produce enormously more value
17:43 < muurkha> as measured by things like people being able to fly
17:43 < fenn> deposition of steel wire embedded in your polymer of choice should be pretty cheap and strong enough for many applications
17:43 < fenn> but orienting the fiber is something that nobody has figured out
17:44 < muurkha> there are oriented carbon-fiber filaments now
17:44 < muurkha> where the reel of filament has carbon fiber running continuously through the whole thing
17:44 < fenn> yeah i dunno
17:44 < muurkha> I mean carbon-fiber-filled.  the filament is still PLA or nylon or something thermoplastic
17:45 < fenn> in practice the carbon fiber wears the nozzle quickly, and in tests it didn't increase the strength at all
17:45 < fenn> i suspect there's not enough carbon fiber
17:45 < muurkha> you're probably thinking about chopped-fiber-filled filament, not continuous
17:45 < fenn> could be
17:45 < muurkha> the continuous-fiber filament makes parts that are about as strong as aluminum
17:45 < muurkha> but for the most part people who are producing one-off composite parts use UD tape instead
17:46 < pasky_> I can't stop laughing reading muurkha's dialog :)
17:46 < fenn> right. i should be able to deposit UD carbon tow prepreg with a $100 printer
17:46 < pasky_> btw some of GPT-3 replies are multi-line and Muaddib just cuts it off after the first line, I think that harmed the verilog part
17:46 < muurkha> pasky_: isn't that amazing?  what do Muaddib's actual prompts look like?
17:46 < muurkha> yeah, it looked like it.  do you have the rest of the reply?
17:47 < muurkha> I'm wondering where it got the killer robot army stuff from.  did you tell it it was supposed to be a robot?
17:47 < fenn> pasky_: can you concatenate the lines so it's not spammy but still we get to see the whole response?
17:48 < pasky_> fenn: typically it just generates followup IRC conversation which would just look lame (hm, I suppose whenever it happens, I could expose the full snippet over web)
17:49 < pasky_> I think I already shared the context once, it's:
17:49 < pasky_> my $query = join("\n", @context_until_last_message) . "\n\nBut $mynick follows up with a single surprisingly witty comeback that makes me laugh... :)\n\n$last\n<$mynick> $nick:";
17:49 < pasky_> (the context is 5(?) previous messages in the channel)
17:49 < muurkha> ah, that explains why it's constantly trying to be funny
17:50 < muurkha> that's an amazingly small amount of context for the functionality it's getting
17:50 < pasky_> 00:09 Irssi:                            'text' => ' The code would look like this:
17:50 < pasky_> 00:09 Irssi:
17:50 < pasky_> 00:09 Irssi: module top;
17:50 < pasky_> 00:09 Irssi:    initial begin
17:50 < pasky_> 00:09 Irssi:       $display("Hello, world!");
17:50 < pasky_> 00:09 Irssi:    end
17:50 < pasky_> or just
17:51 < pasky_> 00:09 Irssi: endmodule
17:51 < pasky_> 00:09 Irssi:
17:51 < pasky_> 00:09 Irssi: muurkha: there you go.'
17:51 < pasky_> 00:10 Irssi:                            'text' => ' =====================
17:51 < pasky_> 00:10 Irssi: <Muaddib> muurkha: It means that you should probably read the documentation for the Verilog language.',
17:51 < pasky_> yeah it's pretty crazy
17:51 < muurkha> hmm, I don't know enough Verilog to know if that's actually correct
17:52  * pasky_ knows zero Verilog...
17:54 < muurkha> I've just updated http://canonical.org/~kragen/muaddib-coding-interview.txt with that, thanks!
17:55 < muurkha> (and again)
18:00 < fenn> what's the ================ stuff? is that just garbage?
18:01 < fenn> juri_: "a slicer that gets confused trying to put faces on squares with a chunk bitten out of them" is there an image online showing what you meant by this?
18:02 < muurkha> Muaddib: why do you prefer Sublime to Emacs?
18:02 < Muaddib> muurkha: less lisp for one :)
18:02 < muurkha> Muaddib: why do you prefer Vim to Sublime?
18:02 < Muaddib> muurkha: less lisp for one :)
18:02 < muurkha> Muaddib: why do you prefer Emacs to Vim?
18:02 < Muaddib> muurkha: less lisp for one :)
18:02 < fenn> the family guy school of comedy
18:03 < fenn> repeat a bad joke until it's still not funny, but it's clear that you're abusing the audience
18:03 < muurkha> haha
18:03 < muurkha> also though if you see an IRC chat log in which some guy has just said the same thing twice in response to two similar questions
18:03 < muurkha> it's a reasonably good guess that in response to a third similar question he will say the same thing a third time
18:04 < muurkha> Muaddib: what's your favorite text editor?
18:04 < Muaddib> muurkha: vim
18:04 < muurkha> Muaddib: why?
18:04 < Muaddib> muurkha: because it's an editor, not a word processor
18:04 < muurkha> Muaddib: as opposed to?
18:04 < Muaddib> muurkha: as opposed to a text editor
18:05 < muurkha> Muaddib: why would that make you prefer Vim over Emacs, VS Code, etc.?  they're all editors rather than word processors
18:05 < Muaddib> muurkha: for the same reason I prefer Vi over Notepad
18:05 < muurkha> Muaddib: why is that?
18:05 < Muaddib> muurkha: the feel, I guess
18:05 < muurkha> Muaddib: what is the feel?
18:05 < Muaddib> muurkha: it just suits me better, I don't know why
18:07 < fenn> juri_: i only ask because i expected an f-rep engine to not have non-manifold meshing issues, which is how i interpreted that
18:08 < fenn> juri_: er, but i guess HSlice takes STL as input, so ... nevermind
18:12 < muurkha> pasky_: there's a chaos or randomness parameter in the GPT-3 interface, right?  that sort of determines how biased it is toward the most probable result?
18:13 < fenn> sometimes repeating the output can be funny, but usually it is not
18:13 < fenn> it seems more like a bug than good
18:14 < muurkha> I was thinking more like, in the above dialogue, it's saying extremely stereotypical things
18:14 < muurkha> so maybe the chaos parameter is low (I forget what it's called)
18:15 < fenn> "as opposed to a text editor" didn't really make sense
18:18 < muurkha> not quite, but it's a stupid nonsensical thing a human would say
18:18 < muurkha> Muaddib: what do you think?
18:18 < Muaddib> muurkha: of what?
18:20 < muurkha> apparently the Verilog hello-world is correct
18:22 < muurkha> Muaddib: I feel debased just for knowing you exist. I despise everything about you especially your Mac hatred. You are a bloody nardless newbie twit protohominid chromosomally aberrant caricature of a coprophagic cloacal parasitic pond scum.  And I wish you would go away from CSMA.
18:22 < Muaddib> muurkha: :P
18:22 < muurkha> hmm, well, that was anticlimatic
18:22 < fenn> muurkha: do you know what CSMA is?
18:23 < fenn> oops
18:23 < fenn> Muaddib: do you know what CSMA is?
18:23 < Muaddib> fenn: how dare you use technical terms in here
18:23 < fenn> Muaddib: asking for a friend
18:23 < Muaddib> fenn: both yours and mine
18:24 < muurkha> fenn: comp.sys.mac.advocacy, where that was posted
18:24 < fenn> oh this must be some canonical flame ammunition file
18:24 < fenn> .insult
18:24 < saxo> you racist boner-pirate
18:24 < muurkha> something like that, yeah
18:24 < fenn> .insult
18:24 < saxo> you slutty nut-monster
18:26 < muurkha> see also https://cheswick.com/insult
18:26 < muurkha> .t
18:26 < saxo> Insult and passphrase generator
18:26 < muurkha> but I suspect the flame post was recycled through a few generations of human editing
18:27 < muurkha> Muaddib: you're sweet, I like you
18:27 < Muaddib> muurkha: you're heavy, I can feel you
18:27 < muurkha> heh
18:34 < muurkha> Muaddib: ♪ he ain't heavy, he's my brother ♫
18:34 < Muaddib> muurkha: ♪ he didn't bring us here for salvation ♫
18:34 < muurkha> Muaddib: what song is that from?
18:34 < Muaddib> muurkha: it's from the album Abbey Road, by the Beatles
19:14 < pasky_> muurkha: yeah that's at 0.95 which is "very chaotic"
19:15 < muurkha> huh
19:16 < muurkha> how chaotic does it have to be before it doesn't say "less lisp for one :)" when asked why *Emacs* is better than *Vim*?
19:16 < pasky_> one of my attempts to avoid the repetitions.. (it used to be a lot more prone to that than it currently is, actually)
19:16 < pasky_> that sounds like a pretty chaotic answer
19:16 < muurkha> heh
19:16 < pasky_> muurkha: how chaotic does it have to be before it doesn't say "less lisp for one :)" when asked why *Emacs* is better than *Vim*?
19:16 < pasky_> sorry
19:17 < pasky_> Muaddib: how chaotic does it have to be before it doesn't say "less lisp for one :)" when asked why *Emacs* is better than *Vim*?
19:17 < muurkha> heh
19:17 < Muaddib> pasky_: when it crashes and there's only one core left to continue?
19:17  * muurkha only has one core left
21:41 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has quit [Quit: trust no one]
21:53 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has joined #hplusroadmap
22:10 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer]
23:08 -!- Netsplit *.net <-> *.split quits: HumanG33k
23:10 -!- Netsplit over, joins: HumanG33k
23:12 -!- Netsplit *.net <-> *.split quits: mgxm
23:13 -!- Netsplit over, joins: mgxm
23:42 < Malvolio> Muaddib: What are the chances of anything coming from Mars?
23:43 < Muaddib> Malvolio: 0.02 in every 10,000
--- Log closed Mon Jul 11 00:00:17 2022