--- Log opened Tue Apr 30 00:00:41 2024 02:19 -!- millefy6 [~Millefeui@91-160-78-132.subs.proxad.net] has joined #hplusroadmap 02:21 -!- millefy [~Millefeui@91-160-78-132.subs.proxad.net] has quit [Ping timeout: 260 seconds] 02:21 -!- millefy6 is now known as millefy 03:05 < fenn> https://twitter.com/JonAskonas/status/1784994234065359208 03:05 < fenn> .m https://twitter.com/JonAskonas/status/1784994234065359208 03:05 < AugustaAda> ​twitter: We are already seeing an explosion of AI regulation that is designed to ban open source while claiming to be neutral. ␤ ␤ SB 1047 designates a "hazardous capability" to include what a third party can show with infinite fine-tuning and re-training. Meanwhile, closed models get points… https://pbs.twimg.com/media/GMWRObwXEAAVv9B.jpg https://pbs.twimg.com/media/GMWROcMWkAAd3lQ.jpg 03:05 < AugustaAda> ​https://pbs.twimg.com/media/GMWSytVXkAEwopk.png [https://twitter.com/Scott_Wiener/status/1784964934851289492 The thread also contains fear mongering about open source. ␤ ␤ Unless you think catastrophic harms are going to happen left & right with open source projects, it’s absolutely false to say this “de facto criminalizes open source.” ␤ ␤ + these are civil sanctions, not criminal liability! ] 03:09 < fenn> california bill to ban open source AI https://legiscan.com/CA/text/SB1047/id/2919384 03:27 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has joined #hplusroadmap 03:42 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has quit [Remote host closed the connection] 03:42 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has joined #hplusroadmap 04:34 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Quit: Gateway shutdown] 04:39 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap 05:22 -!- gptpaste [~x@yoke.ch0wn.org] has quit [Ping timeout: 272 seconds] 05:22 -!- AugustaAda [~x@yoke.ch0wn.org] has quit [Ping timeout: 260 seconds] 06:07 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.] 06:08 -!- TMM_ [hp@amanda.tmm.cx] has joined #hplusroadmap 08:48 -!- Guest36 [~Guest64@190.213.148.10] has joined #hplusroadmap 09:56 -!- Guest36 [~Guest64@190.213.148.10] has quit [Quit: Client closed] 10:25 -!- tinwhiskers [~tinwhiske@user/tinwhiskers] has quit [Ping timeout: 268 seconds] 10:26 -!- tinwhiskers [~tinwhiske@user/tinwhiskers] has joined #hplusroadmap 11:34 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer] 12:13 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap 12:25 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has left #hplusroadmap [] 12:26 < hprmbridge> Eli> I’m thinking about doing chemo for my Austin Texas allergies. Downsides? 12:28 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap 13:17 < hprmbridge> Eli> “He used this one weird trick to annihilate his allergies. Doctors hate him” 13:56 < ike8> https://www.osnews.com/story/139462/microsoft-intends-to-record-everything-you-do-on-your-pc-for-ai-processing/ 13:56 < ike8> "why is my laptop baking bread in my backpack?" 14:11 < geneh2> google's android keyboard 'processes' everything you type on it. But don't worry google isn't evil 14:12 < geneh2> and they don't want one user having too much influence on the model they are training and they try to 'preserve privacy' 14:13 -!- juri_ [~juri@implicitcad.org] has joined #hplusroadmap 15:22 < hprmbridge> gourneau> A The Singularity is Nearer book launch event in Menlo Park in June, I will be there 15:22 < hprmbridge> gourneau> https://www.eventbrite.com/e/ray-kurzweil-with-amy-kurzweil-tickets-885376002457?aff=eemailordconf 15:27 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has quit [Remote host closed the connection] 15:28 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has joined #hplusroadmap 15:54 < nsh> websim.ai is pretty much most the singularity-convincing thing i've seen 15:54 < nsh> https://twitter.com/search?q=websim.ai&f=live 15:54 < nsh> people are literally imagining websites magically 16:52 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has quit [Ping timeout: 260 seconds] 17:05 < fenn> i can't get most things you've linked from websim to actually run, only the tetris game 17:06 < fenn> what am i missing 17:23 < fenn> quantified mouse panopticon https://oldenlabs.com/ 17:26 < fenn> the hardware is unreasonably expensive 17:35 < hprmbridge> Lev> what happened to tagging the animals 17:35 < hprmbridge> Lev> like am i missing something? that'd be a lot more reliable 17:35 < hprmbridge> Lev> and this is a case where reliability is kind of a big deal 17:36 < hprmbridge> bootstrap3141> Probably designed by techies who don’t have a good grasp on biotech space 18:12 < fenn> no it's by someone who was doing genetic engineering for longevity and handling mice all the time 18:12 < hprmbridge> Lev> yeah but like 18:12 < fenn> they do tag the mice with ear tags 18:12 < hprmbridge> Lev> stick some RFID tags and use that 18:12 < hprmbridge> Lev> Much more reliable and accurate 18:13 < hprmbridge> Lev> Want to have position data for your identical mice as they run around a cage? bet 18:17 < fenn> common practice is to clip their toenails to make a code, but it's hard to read 18:19 < fenn> RFID won't give you movement data, at least not in a mouse sized wearable package 18:19 < fenn> active tags with accelerometers need a big heavy battery 18:19 < hprmbridge> Lev> You shouldn't need accelerometers 18:19 < fenn> why not 18:19 < hprmbridge> Lev> most RFID systems aren't constantly actively scanning 18:20 < hprmbridge> Lev> Do that, deal with overlapping signals, and a bit of triangulation between readers 18:20 < fenn> i agree it's possible in principle, but the set of tag technologies that currently exist won't do that 18:20 < fenn> the 900 MHz tags are too big for a mouse 18:20 < hprmbridge> Lev> I mean isn't that what LANDMARC does etc? 18:21 < hprmbridge> Lev> not exactly what i mean since those are also NN approaches 18:21 < hprmbridge> Lev> but still 18:23 < hprmbridge> Lev> Like given a RFID/NFC like protocol that gives me some signal strength-like metric I think I could whip up something for this fairly quickly 18:23 < hprmbridge> Lev> reliability might be iffy without additional stuff like some explicit linearization of path histories, but still 18:28 < fenn> consider a simple colored eartag as a passive optical transciever that can be triangulated, and you're back to cameras and computer vision. simply going to RF doesn't solve any of the hardness 18:28 < hprmbridge> Lev> It does allow me to relatively explicitly say "this signal comes from x and no other" 18:29 < fenn> you could do that with uniquely colored plastic tags though 18:29 < fenn> i guess it fails with IR night vision 18:30 < hprmbridge> Lev> i mean i'm more saying i think that position noise is less problematic than "oops it's the wrong mouse" noise 18:30 < hprmbridge> Lev> so i find it an odd choice cuz of that 18:32 < fenn> well any kind of automated data collection is orders of magnitude better than doing it manually 18:33 < hprmbridge> Lev> fair 18:39 < fenn> lol kurzweil wants $30 for permission to attend his book signing 18:40 < hprmbridge> Lev> vodka is cheaper and better 18:41 < fenn> i'll construct a robot kurzweil with AI and have it sign the book instead 20:40 -!- mxz__ [~mxz@user/mxz] has joined #hplusroadmap 20:40 -!- mxz_ [~mxz@user/mxz] has quit [Ping timeout: 260 seconds] 20:41 -!- mxz [~mxz@user/mxz] has quit [Ping timeout: 256 seconds] 20:41 -!- mxz__ is now known as mxz 22:02 -!- Guest45 [~Guest98@1.145.248.194] has joined #hplusroadmap 22:05 -!- Guest45 [~Guest98@1.145.248.194] has quit [Client Quit] 22:48 -!- mxz_ [~mxz@user/mxz] has joined #hplusroadmap --- Log closed Wed May 01 00:00:42 2024