--- Log opened Sun Oct 12 00:00:06 2025
00:40 -!- Croran [~Croran@user/Croran] has quit [Ping timeout: 244 seconds]
00:42 -!- Croran [~Croran@user/Croran] has joined #hplusroadmap
01:02 -!- Evolver [sid32215@user/evolver] has quit [Server closed connection]
01:03 -!- Evolver [sid32215@user/evolver] has joined #hplusroadmap
01:38 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
01:38 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
07:26 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Quit: Avoid fossil fuels and animal products. Have no/fewer children. Protest, elect sane politicians. Invest ecologically.]
07:26 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
08:57 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Ping timeout: 246 seconds]
09:16 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
09:21 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Quit: Avoid fossil fuels and animal products. Have no/fewer children. Protest, elect sane politicians. Invest ecologically.]
09:37 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
10:06 -!- gl00ten [~gl00ten@195.23.221.102] has joined #hplusroadmap
10:06 -!- gl00ten [~gl00ten@195.23.221.102] has quit [Remote host closed the connection]
11:52 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
11:52 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
12:13 < hprmbridge> kanzure> Apparently this is non-obvious to emulation people so I will report it here as well: You wouldn't want someone with raw hardware access to have full root over your code. You would not place the whole emulation in a single computer, you would likely prefer redundancy or erasure coding from multiple machines in distributed consensus. Secure multi-party computation affords other options.
12:13 < hprmbridge> kanzure> "trusted setup" is important here.
12:47 < hprmbridge> kanzure> https://blog.cr.yp.to/20251004-weakened.html
12:47 < TMA> isn't secure multiparty computation extremely inefficient? 
12:48 < TMA> wouldn't that render whole brain emulation infeasible or expensive enough to make insecure shortcuts a more tempting option
12:50 < TMA> which would lead to the same pattern as with centralized services offering "free services"
12:52 < hprmbridge> kanzure> yes it's computationally expensive, but not substantially compared to the costs of full emulation! what's an extra order of magnitude or two between friends already seeking 10^26 flops?
12:55 < TMA> oh, well, I somehow expected that it would add substantially more overhead. a constant factor is palatable, even if inconvenient
12:55 < hprmbridge> kanzure> well don't quote me on that, I didn't look up the overhead nor the emulation order of magnitude.
12:57 < hprmbridge> kanzure> there are also other issues to consider regarding verification, encryption, deniability, cryptographic integrity of memory, uhh disallowing arbitrary third party introspection of memory, much of this will depend on your trusted setup procedures (did they really delete the keys? or who has the keys?)
12:57 < TMA> I expected a O(f(n)) problem without mpc to be Ω((f(n))^a) for some a>1
12:58 < hprmbridge> kanzure> well you can do efficient verification proofs with non-interactive arguments of knowledge
12:59 < TMA> I would need to read much more into the theory to be able to go beyond my gut feeling
13:04 < hprmbridge> kanzure> "Discovering neural policies to drive behaviour by integrating deep reinforcement learning agents with biological neural networks" https://www.nature.com/articles/s42256-024-00854-2 (2024) but just worms.
13:04 < hprmbridge> kanzure> "Adaptive reinforcement learning is causally supported by anterior cingulate cortex and striatum" https://www.cell.com/neuron/abstract/S0896-6273(25)00365-4 "ACC simulation impairs weighting of error histories and optimizing exploration; Striatum simulation improves use of prediction errors to update value expectations; Stimulation induces uncertainty in ACC but in striatum augments outcome
13:04 < hprmbridge> kanzure> certainty"
13:05 < TMA> I was extrapolating from a distributed voting protocol, where the amount of work each voter needed to do was proportional to the number of voters
13:06 < TMA> that bumped the complexity from O(n) to O(n^2) for the whole process and from O(1) to O(n) for each voter
13:07 < TMA> and that was a simple protocol, that failed when any single node/voter did not perform its duties
13:10 < TMA> I remember adding resiliency generally adds at least O(log n) additional factor (an O(n) process becomes a O(nlogn) process) IIRC
13:12 < hprmbridge> kanzure> if we are allowing for a belief in trusted setup, like for uploading or scanning, then we might as well invoke a distributed set of nodes with trusted execution environments instead of secure MPC/FHE?
13:14 < hprmbridge> kanzure> secure multi-party computation would be useful for running an emulation on untrusted hardware, such as volunteers or less than trusted data centers.
13:15 < hprmbridge> kanzure> you could get some level of extra censorship resistance with public untrusted computing.
13:16 < TMA> it would allow migration from a "small" trusted/verifiable facility for the scanning and initial encoding into the untrusted world at large
13:18 < hprmbridge> kanzure> well... you still have unverifiable insecure input... and you can try to outpace an adversary by purchasing random untrusted hardware and hoping they haven't already targeted you for infiltration? if you don't bootstrap enough of an uncompromised infrastructure before you are targeted, i am not sure there's any remaining hope.
13:28 < TMA> the scanning itself needs to be trusted, because the physical brain is not distributed and reliably replicated _before_ the scanning
13:29 < hprmbridge> kanzure> you might be able to use an optical splitter straight off the cameras or image acquisition. including for sequencing of barcodes and antibody immunostaining.
13:29 < TMA> so
13:30 < TMA> but anyway, I am too dumb/tired to come up with a real solution today
14:37 -!- ike8 [e8f913dbdf@irc.cheogram.com] has quit [Server closed connection]
14:40 -!- ike8 [e8f913dbdf@irc.cheogram.com] has joined #hplusroadmap
15:32 -!- juri_ [~juri@implicitcad.org] has quit [Ping timeout: 264 seconds]
15:33 -!- juri_ [~juri@178.20.94.63] has joined #hplusroadmap
15:53 < hprmbridge> kanzure> why is peter thiel binging one piece
15:58 < L29Ah> MuaddibLLM: what did kanzure mean?
15:58 < MuaddibLLM> hunting for thiel's anime preferences...
15:58 < MuaddibLLM> Thiel apparently watched 1000+ One Piece episodes and recently gave lectures drawing biblical parallels (Luffy as Christ, Imu as antichrist) - kanzure was probably reacting to this wild recent news (https://www.theguardian.com/us-news/2025/oct/10/peter-thiel-lectures-antichrist, https://firstthings.com/voyages-to-the-end-of-the-world/)
16:00 < L29Ah> rip
16:04 < hprmbridge> kanzure> build NERV do it
16:04 < L29Ah> MuaddibLLM: does thiel have an arcology like in The Ascent videogame to get a hold on?
16:05 < MuaddibLLM> checking if thiel's gone full cyberpunk dystopia landlord yet...
16:05 < MuaddibLLM> [23:04] Not quite an arcology yet - he's backing Praxis (Mediterranean "startup city"), bailed on seasteading, and has NZ property/citizenship, but nothing vertically integrated like The Ascent's megastructure (https://www.curbed.com/article/inside-the-peter-thielbacked-praxis.html, https://www.fastcompany.com/91392989/tech-billionaires-are-building-their-own-private-cities-heres-whos-doing-it-and-where)
16:59 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
17:02 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 256 seconds]
17:02 -!- stipa_ is now known as stipa
17:15 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 272 seconds]
17:21 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Ping timeout: 246 seconds]
18:01 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
18:11 < hprmbridge> kanzure>  https://cdn.discordapp.com/attachments/1064664282450628710/1427101288679866478/image0.jpg?ex=68eda325&is=68ec51a5&hm=6a641c5ff6807baf0be63071734ed78ecc97d61505fff34e591a08bb5405f695&
18:14 -!- Croran [~Croran@user/Croran] has quit [Ping timeout: 256 seconds]
18:14 -!- Croran [~Croran@user/Croran] has joined #hplusroadmap
18:15 -!- catalase [~catalase@user/catalase] has quit [Quit: Ping timeout (120 seconds)]
18:28 -!- catalase [~catalase@user/catalase] has joined #hplusroadmap
18:51 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 272 seconds]
20:40 -!- ike8 [e8f913dbdf@irc.cheogram.com] has quit [Server closed connection]
20:40 -!- ike8 [e8f913dbdf@irc.cheogram.com] has joined #hplusroadmap
21:13 -!- Malvolio [~Malvolio@idlerpg/player/Malvolio] has quit [Quit: 2B59 so why not legos + superglue 2025-10-13 04:13:42.344]
21:37 -!- zooper26 [~zooper26@syn-045-046-077-196.res.spectrum.com] has joined #hplusroadmap
21:37 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
21:43 -!- zooper26 [~zooper26@syn-045-046-077-196.res.spectrum.com] has quit [Quit: Client closed]
22:02 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
22:02 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
22:17 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 272 seconds]
22:44 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
22:57 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 272 seconds]
23:18 -!- ike8 [e8f913dbdf@irc.cheogram.com] has quit [Server closed connection]
23:20 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
23:27 -!- ike8 [e8f913dbdf@irc.cheogram.com] has joined #hplusroadmap
--- Log closed Mon Oct 13 00:00:07 2025