--- Log opened Wed Mar 20 00:00:32 2024
00:07 -!- mxz [~mxz@user/mxz] has joined #hplusroadmap
00:11 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Ping timeout: 268 seconds]
00:25 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 260 seconds]
01:09 < fenn> .t https://youtu.be/QY87rRTrJlE
01:09 < saxo> NVIDIA Omniverse Foundational Technology Montage I GTC Spring 2024 Edition - YouTube
01:16 < fenn> 7 million discrete parts on that ship model
02:16 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has joined #hplusroadmap
03:45 -!- Anachron [~Malvolio@idlerpg/player/Malvolio] has joined #hplusroadmap
04:47 -!- balrog_ [znc@user/balrog] has joined #hplusroadmap
04:47 -!- tinwhiskers_ [~tinwhiske@user/tinwhiskers] has joined #hplusroadmap
04:49 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
04:53 -!- Netsplit *.net <-> *.split quits: tinwhiskers, balrog, stipa
04:53 -!- stipa_ is now known as stipa
05:03 < kanzure> hmph
05:39 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has left #hplusroadmap []
05:41 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
05:56 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has left #hplusroadmap []
06:10 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
06:16 -!- Jenda [~jenda@coralmyn.hrach.eu] has quit [Ping timeout: 255 seconds]
06:18 -!- Jenda [~jenda@coralmyn.hrach.eu] has joined #hplusroadmap
07:16 -!- alethkit [23bd17ddc6@sourcehut/user/alethkit] has quit [Remote host closed the connection]
07:17 -!- alethkit [23bd17ddc6@sourcehut/user/alethkit] has joined #hplusroadmap
08:56 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
09:16 -!- ike8 [e8f913dbdf@irc.cheogram.com] has joined #hplusroadmap
09:16 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has left #hplusroadmap []
09:31 < hprmbridge> nmz787> https://nvidianews.nvidia.com/news/nvidia-blackwell-platform-arrives-to-power-a-new-era-of-computing
09:55 -!- cthlolo [~lorogue@77.33.24.3.dhcp.fibianet.dk] has joined #hplusroadmap
09:59 -!- boxy [~box@213.233.85.119] has joined #hplusroadmap
10:07 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
10:13 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
10:47 -!- boxy [~box@213.233.85.119] has quit [Ping timeout: 255 seconds]
11:58 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer]
12:25 -!- cthlolo [~lorogue@77.33.24.3.dhcp.fibianet.dk] has quit [Remote host closed the connection]
12:32 -!- cthlolo [~lorogue@77.33.24.3.dhcp.fibianet.dk] has joined #hplusroadmap
12:37 -!- justanot1 [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
12:40 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 260 seconds]
12:45 -!- cthlolo [~lorogue@77.33.24.3.dhcp.fibianet.dk] has quit [Quit: Leaving]
13:00 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
13:10 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer]
13:28 < geneh2> there might be private github repos in this https://huggingface.co/datasets/bigcode/the-stack-v2
13:28 < geneh2> 67 TB of code
13:30 < geneh2> or not, still 67 TB of code for training
13:38 < gwillen> it seems like you can't actually get it, though, it's "open" data
13:38 < gwillen> > Downloading the dataset in bulk requires a an agreement with SoftwareHeritage and INRIA. Contact datasets@softwareheritage.org for more information.
14:23 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
15:02 -!- andytoshi [~apoelstra@user/andytoshi] has quit [Ping timeout: 252 seconds]
15:03 -!- andytoshi [~apoelstra@user/andytoshi] has joined #hplusroadmap
15:20 -!- Guest63 [~Guest63@host-92-6-1-68.as13285.net] has joined #hplusroadmap
15:26 -!- Guest63 [~Guest63@host-92-6-1-68.as13285.net] has quit [Quit: Ping timeout (120 seconds)]
16:03 < fenn> 32TB of deduplicated code. that's pretty cool. it's good to see public institutions doing something useful
16:13 < fenn> someone please make the data available according to the original code license, without these extra restrictions that have been illegally slapped on: https://www.softwareheritage.org/2023/10/19/swh-statement-on-llm-for-code/
16:14 < fenn> i don't know what makes people think that they can re-license stuff by mere aggregation
16:21 < fenn> hmm. "To the best of our knowledge, all files contained in the dataset are licensed with one of the permissive licenses (see list in Licensing information [404's]) or no license." 
16:41 < fenn> so if you want to make your code available for humanity to use, you can't license under a copyleft license because that will prevent it from being included on grounds of lack of attribution, but if you make it available under a permissive license anyone can come along and add their own restrictions to it. is there a way out of this mess?
16:59 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has quit [Ping timeout: 260 seconds]
17:32 < gwillen> anyone who makes a database can always say "we refuse to send you our database" regardless of the underlying license, they don't owe you anything under the license terms if they haven't distributed you anything
17:33 < hprmbridge> bootstrap3141> There are better licenses for that reason than GPL or other strict reasons.  Attribution is usually less of an issue than are the source distribution requirements.  This is why many projects use Apache or MIT licenses.
17:36 < hprmbridge> bootstrap3141> It does seem absurd that they slap on an extra license… probably some CYA move
18:08 -!- Hooloovoo [~Hooloovoo@hax0rbana.org] has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
18:12 -!- Hooloovoo [~Hooloovoo@hax0rbana.org] has joined #hplusroadmap
18:20 < fenn> they still have to comply with the license that allows them to redistribute it. if that license forbids adding extra restrictions then they can't legally redistribute it with extra restrictions. softwareheritage seems to have sidedstepped this by only redistributing permissively licensed works. also, "fair use" is implicitly invoked(?) for the non-licensed code (which is copyrighted by default)
18:21 < fenn> so why can they redistribute unlicensed code but not copylefted code? it doesn't quite make sense
18:21 < fenn> thanks to the intentional ambiguity we will probably never know
18:23 < fenn> it could also be considered education or research exemptions i guess
18:27 < fenn> there is no CC-SA license, only CC-BY-SA
20:16 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Ping timeout: 272 seconds]
21:29 -!- tinwhiskers_ is now known as tinwhiskers
21:50 -!- Hooloovoo [~Hooloovoo@hax0rbana.org] has quit [Ping timeout: 252 seconds]
21:51 -!- Hoolootwo [~Hooloovoo@hax0rbana.org] has joined #hplusroadmap
22:00 -!- mxz [~mxz@user/mxz] has quit [Ping timeout: 268 seconds]
23:06 -!- alexbfi [~alexbfi@dzyhs8yyyyyyyyyyyyt8t-3.rev.dnainternet.fi] has joined #hplusroadmap
23:07 -!- alexbfi [~alexbfi@dzyhs8yyyyyyyyyyyyt8t-3.rev.dnainternet.fi] has quit [Client Quit]
23:07 -!- alexbfi [~alexbfi@dzyhs8yyyyyyyyyyyyt8t-3.rev.dnainternet.fi] has joined #hplusroadmap
23:07 -!- alexbfi_ [~alexbfi@dzyhs8yyyyyyyyyyyyt8t-3.rev.dnainternet.fi] has joined #hplusroadmap
23:55 -!- darsie [~darsie@84-112-12-36.cable.dynamic.surfer.at] has joined #hplusroadmap
--- Log closed Thu Mar 21 00:00:33 2024