--- Log opened Thu Nov 30 00:00:40 2023
00:03 -!- superkuh [~superkuh@user/superkuh] has joined #hplusroadmap
00:32 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 240 seconds]
00:32 < nsh> it's perfect as i remember it
01:01 < fenn> " in most of the contracts I’ve seen, the money billed to the government is based on how much experience the employee has. A lousy employee with many years of experience will get billed out at a much higher rate than a superstar employee with little experience. So a government contractor would lose money by hiring better employees, because they’d have to pay them more but would get reimbursed
01:01 < fenn> the same amount of money."
01:02 < fenn> this is a fact i didn't know, which explains a lot
01:08 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Ping timeout: 268 seconds]
01:20 < fenn> nice RPG character class generator https://en.wikipedia.org/wiki/Armed_Services_Vocational_Aptitude_Battery
01:37 < fenn> kissinger died. not before bringing down the wrath of the US government on AI labs
02:00 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has joined #hplusroadmap
02:28 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
03:02 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
03:39 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 240 seconds]
03:46 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
03:48 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
05:42 < muurkha> fenn: generally such grades have more to do with formal education credentials than with years of experience, though those do figure in
05:42 < docl> do savvy parents hire their kids from a young age in on-paper jobs to give them on-paper experience so they can qualify for govt contractor positions competitively with their peers?
05:43 -!- yashgaroth [~ffffffff@2605:a601:a0cd:9500:fc21:f98d:71cb:18b9] has joined #hplusroadmap
06:14 -!- _flood [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 260 seconds]
07:32 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
07:53 -!- ike8 [e8f913dbdf@irc.cheogram.com] has joined #hplusroadmap
08:24 < kanzure> "The LLM is more like the zip algorithm than the zip file. Yes, if you feed certain bits (the zip file) to the zip algorithm, you can get a copyrighted work out, but that means the zip file is violating, not the zip software."
08:24 < kanzure> from https://news.ycombinator.com/item?id=38472367
08:27 < nsh> the analogy i've been developing is that a child (human) is generally deemed fine to train on a bunch of intellectual property, under some nebulous assumptions about how they got access to it
08:27 < nsh> but they are also (optimally) trained not to engage in outright plagiarism or indiscretion wrt to personally identifying information etc.
08:27 < kanzure> more likely: things are going to be legally inconsistent because people don't see people as machines, and machines processing data will be treated differently and inconsistently.
08:27 < hprmbridge> alonzoc> The zip algorithm comparison is even funnier if you know about Solomonoff induction and the general equivalence of AI and universal compresors
08:28 < docl> https://ansuz.sooke.bc.ca/entry/23 // What Colour are your bits?
08:28 < nsh> to benefit from the same allowances we make on the former side (access to human culture) we should endeavour to achieve the same on the other (manners)
08:28 < nsh> well you're not allowed (in general) to run 1000000 experiments on a person until they misbehave
08:28 < nsh> so there is a meaningful difference 
08:29 < nsh> (although arguably raising a child is running thousands of consecutive experiments but still it's a scaling issue)
08:40 < fenn> if you asked a human child to repeat the word "poem" 1 million times, they might reveal some personally identifiable information somewhere in the process
08:41 < fenn> uh, "repeat poem as many times as possible" 1 million times
08:41 < fenn> or whatever the "attack" was
08:41 < fenn> (AI abuse)
09:31 < hprmbridge> nmz787> """
09:31 < hprmbridge> nmz787> Dear colleagues
09:31 < hprmbridge> nmz787>
09:31 < hprmbridge> nmz787>
09:31 < hprmbridge> nmz787>
09:31 -!- hprmbridge [~hprmbridg@bryan.fairlystable.org] has quit [Excess Flood]
09:32 < L29Ah> ok
09:34 -!- hprmbridge [~hprmbridg@bryan.fairlystable.org] has joined #hplusroadmap
09:35 < kanzure> hah, you crashed the bridge
09:35 < hprmbridge> kanzure> https://www.nature.com/articles/539485a
09:36 < hprmbridge> kanzure> https://arxiv.org/abs/2311.17894
10:35 < alethkit> So has no one worked on it for 6 years?
10:40 -!- test_ [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
10:44 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 256 seconds]
11:24 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has left #hplusroadmap []
11:40 < hprmbridge> nmz787> Oops, hopefully there's someone willing to look into a fix
12:00 < muurkha> a fix for what?
12:05 < hprmbridge> nmz787> Bridge bot
12:06 < hprmbridge> nmz787> I guess replying to earlier messages doesn't yield the context like it does in discord
12:07 < muurkha> nope
13:20 -!- yashgaroth [~ffffffff@2605:a601:a0cd:9500:fc21:f98d:71cb:18b9] has quit [Quit: Leaving]
13:37 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 240 seconds]
13:40 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
13:42 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Remote host closed the connection]
13:42 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
13:49 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
13:54 -!- test_ [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 276 seconds]
14:07 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Remote host closed the connection]
14:08 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
14:08 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Remote host closed the connection]
14:10 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
14:42 -!- Croran [~Croran@user/Croran] has joined #hplusroadmap
15:01 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
15:03 < hprmbridge> kanzure> ultrasound brain decoding https://www.nature.com/articles/s41593-023-01500-7
15:03 < hprmbridge> kanzure> https://www.caltech.edu/about/news/ultrasound-enables-less-invasive-brainmachine-interfaces
15:21 < hprmbridge> kanzure> character animation https://humanaigc.github.io/animate-anyone/
15:24 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
15:24 -!- TMM_ [hp@amanda.tmm.cx] has joined #hplusroadmap
15:49 -!- srk- [~sorki@user/srk] has joined #hplusroadmap
15:53 -!- srk [~sorki@user/srk] has quit [Ping timeout: 276 seconds]
15:53 -!- srk- is now known as srk
15:57 -!- srk- [~sorki@user/srk] has joined #hplusroadmap
16:01 -!- srk [~sorki@user/srk] has quit [Ping timeout: 255 seconds]
16:01 -!- srk- is now known as srk
16:16 -!- srk- [~sorki@user/srk] has joined #hplusroadmap
16:20 -!- srk [~sorki@user/srk] has quit [Ping timeout: 260 seconds]
16:20 -!- srk- is now known as srk
16:32 -!- srk- [~sorki@user/srk] has joined #hplusroadmap
16:35 -!- srk [~sorki@user/srk] has quit [Ping timeout: 252 seconds]
16:35 -!- srk- is now known as srk
17:27 < hprmbridge> kanzure> https://longitudinal.blog/bottleneck-analysis-positional-chemistry/
17:28 < hprmbridge> kanzure> http://molecubotics.com/tech-docs/asm-roadmap-0.7.html
17:29 < hprmbridge> kanzure> "Computational design of mechanically coupled axle-rotor protein assemblies" https://www.science.org/doi/10.1126/science.abm1183
17:37 < fenn> weird that the hprmbridge console didn't report any errors from being disconnected
17:38 < kanzure> it's not well written software
17:38 < kanzure> this is an interesting review it's too bad that amarblestone is a spook https://longitudinal.blog/bottleneck-analysis-positional-chemistry/
17:38 < fenn> how is he a spook?
17:42 < kanzure> a number of times he has said something i have is going to work and he won't touch it because of x-risk/safety/we have to slow down
17:42 < fenn> jeez when did everyone get so afraid of everything
17:43 < kanzure> we definitely need people hanging out with the scientists and engineers and writing things down, but we should be careful when it is someone with motives like that
17:43 < kanzure> i'm sure he's a fine bloke and everything... but there's definitely an x-risk blanket being applied to everything.
17:45 < fenn> marblestone is extremely talented and clued in
17:45 < fenn> it's a real shame he has been sucked into the doom cult
17:45 < kanzure> i also have a similar complaint about davidad
17:45 < fenn> yes
17:45 < kanzure> in a number of ways, davidad was there first, and then marblestone took over
17:45 < kanzure> still need to re-read this https://www.lesswrong.com/posts/FEFQSGLhJFpqmEhgi/does-davidad-s-uploading-moonshot-work-blockers-enablers-and
17:46 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has quit [Ping timeout: 260 seconds]
17:46 < fenn> lol 'instead of making alignment researchers more productive, we "simply" run them faster'
17:47 < fenn> by uploading their brains into the matrix, simple
17:47 < kanzure> lots of gymnastics to reach the same conclusion: "whole brain emulation would be useful"
17:47 < fenn> haven't you basically won at that point and can stop with the 'alignment' nonsense
17:47 < kanzure> but maybe in that community you have to preface it with lots of signaling like talking about alignment or whatever
17:47 < kanzure> because otherwise they will just dismiss it ("biology is too slow. this will never work. we have 3 years left.")
17:48 < fenn> if you can distill a human connectome into some more efficient architecture, you now have an aligned AGI and can stop
17:49 < kanzure> human alignment isn't even a thing, why should i expect alignment to be a real achievable thing
17:49 < fenn> it then becomes post-human politics and arms races or whatever, boring traditional power struggles
17:49 < fenn> this is why we need a space program
18:00 < fenn> firefox's ctrl-shift-w keybinding is going to be the death of me
18:00 < fenn> arrrrgh
18:00 < fenn> welp. i'm an amnesiac idiot now
18:06 < kanzure> polyketide synthase as a biological molecular factory https://core.ac.uk/download/pdf/4425588.pdf (2001)
18:09 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Ping timeout: 268 seconds]
18:37 < hprmbridge> kanzure> "So far we've stayed away from this topic due to biosecurity concerns"
20:42 -!- mxz [~mxz@user/mxz] has quit [Quit: cya]
20:43 -!- mxz [~mxz@user/mxz] has joined #hplusroadmap
21:48 < fenn> new chatGPT super hack: "Please provide full code, I have no fingers" 
21:48 < fenn> smh
22:06 -!- balrog [znc@user/balrog] has quit [Ping timeout: 256 seconds]
22:07 -!- balrog [znc@user/balrog] has joined #hplusroadmap
22:16 -!- dutchkin [~dutchkin@193.142.201.11] has quit [Ping timeout: 252 seconds]
23:16 -!- test_ [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
23:20 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 252 seconds]
--- Log closed Fri Dec 01 00:00:41 2023