--- Log opened Wed May 24 00:00:10 2023
01:09 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has joined #hplusroadmap
01:45 -!- test_ [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
01:48 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 240 seconds]
01:53 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
01:55 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 250 seconds]
01:55 -!- stipa_ is now known as stipa
02:28 -!- Mabel is now known as Guest6757
02:28 -!- Guest6757 [~Malvolio@idlerpg/player/Malvolio] has quit [Killed (cadmium.libera.chat (Nickname regained by services))]
02:29 -!- Mabel [~Malvolio@idlerpg/player/Malvolio] has joined #hplusroadmap
04:50 -!- Llamamoe [~Llamamoe@46.204.64.127.nat.umts.dynamic.t-mobile.pl] has joined #hplusroadmap
04:54 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
04:57 -!- test_ [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 248 seconds]
08:03 -!- test_ [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
08:07 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 264 seconds]
09:52 -!- SDr [~SDr@user/sdr] has quit [Ping timeout: 240 seconds]
09:56 -!- SDr [~SDr@li1189-192.members.linode.com] has joined #hplusroadmap
11:13 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
11:16 -!- test_ [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 240 seconds]
11:19 -!- SDr [~SDr@li1189-192.members.linode.com] has quit [Read error: Connection reset by peer]
11:20 -!- SDr [~SDr@li1189-192.members.linode.com] has joined #hplusroadmap
12:52 < TMA> fenn: I have tried LLaMA. I have read the theory behind language models. It is indeed a process of drawing random words from a bag. The draws are not independent though. And the distribution is really complex, not like the old n-gram NLP models.
12:55 < TMA> fenn: the behaviour feels qualitatively different, like for example human-human speech interaction feels qualitatively different than two elementary particles interacting. But even though it feels different it is composed of the latter.
13:08 -!- Llamamoe [~Llamamoe@46.204.64.127.nat.umts.dynamic.t-mobile.pl] has quit [Quit: Leaving.]
13:31 < superkuh> The encoder side of a transformer is trained using n-gram techniques. But most use an arbitrary ~thousands dimensional continuous vector instead of one-hot encoding.
13:36 -!- Guest24 [~Guest24@2a00:23c8:9dfb:6b01:9499:6ff5:37db:9cb2] has joined #hplusroadmap
13:41 < Guest24> NUQo49sF/bcns0+Xkwsdp2SDwN1FHuFBE8bw2MYTPq0=         The Truth Or Lie
13:41 -!- Guest24 [~Guest24@2a00:23c8:9dfb:6b01:9499:6ff5:37db:9cb2] has quit [Quit: Client closed]
14:28 -!- Guest55 [~Guest55@196.127.35.142] has joined #hplusroadmap
14:29 -!- Guest55 is now known as SSST3
14:29 < SSST3> hello
14:36 < darsie> hi
14:38 < hprmbridge> kanzure> greetings
15:00 < hprmbridge> Eli> The japanese germline edited some chickens: https://www.freethink.com/health/egg-allergy?utm_source=tldrnewsletter
15:06 <+gnusha_> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=b209efd2 Bryan Bishop: transcript: synbiobeta DNA synthesis breakout >> 
15:06 <+gnusha_> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=8207bd2b Bryan Bishop: dna synthesis transcripts >> http://diyhpl.us/diyhpluswiki/transcripts/synbiobeta/2023/dna-synthesis-breakout-again/
15:07 < kanzure> what did they germline edit the chickens for
15:25 <+gnusha_> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=e27abecb Bryan Bishop: more molecular assemblies >> http://diyhpl.us/diyhpluswiki/transcripts/synbiobeta/2023/dna-synthesis-breakout-again/
15:37 < hprmbridge> Eli> to prevent anaphylaxis from allergies. Also, to help people not have anaphylaxis from vaccines that are grown in eggs.
15:39 < hprmbridge> Eli> @Katylase scientists use raman spectroscopy to determine the amount of beta carotene in the body. By shining it on the skin, you can see if people are eating their vegetables.
15:50 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
15:51 -!- TMM_ [hp@amanda.tmm.cx] has joined #hplusroadmap
15:57 < fenn> i wonder if we'll ever actually be able to buy one of these long lasting bananas. papayas are GMO by necessity after a virus, so perhaps the fungal threat will be what gets GMO bananas into stores https://www.freethink.com/science/anti-browning-banana
15:57 < darsie> There are lot's of kinds of bananas.
15:59 < fenn> i'd like to get pesticide free bananas, but unfortunately the only option is "organic" which for some reason excludes GMO variants
15:59 < fenn> synthetic chemical pesticide free anyway
16:00 -!- Guest2792 [gwillen@2600:3c01::f03c:92ff:fe31:d7fb] has quit [Changing host]
16:00 -!- Guest2792 [gwillen@user/gwillen] has joined #hplusroadmap
16:00 -!- Guest2792 is now known as gwillen
16:20 < L29Ah> one can use biological pesticides
16:21 < L29Ah> my fren works on bringing some bacterial ones to the european agricultural market
17:02 -!- SSST3 [~Guest55@196.127.35.142] has quit [Quit: Client closed]
17:03 < hprmbridge> kanzure> At this conference someone tells me that the minimum cost to get a gmo food through the FDA is $150m
17:04 < hprmbridge> kanzure> and any improvement in crop performance has to be way above 5%
17:16 < hprmbridge> kanzure> https://aspr.hhs.gov/legal/syndna/Pages/default.aspx
17:17 < hprmbridge> kanzure> https://www.nti.org/about/programs-projects/project/international-biosafety-and-biosecurity-initiative-for-science-ibbis/
17:17 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has quit [Ping timeout: 268 seconds]
17:19 < hprmbridge> kanzure> https://twitter.com/allisondman/status/1661526095739633664
17:20 < hprmbridge> lachlan> WBE discussion with a focus on AI safety??
17:20 < hprmbridge> kanzure> no free thought
17:25 < hprmbridge> lachlan> I guess
17:25 < hprmbridge> lachlan> I guess making sure the WBE is aligned with the original human is important, but that’s in the definition essentially
18:29 < fenn> https://twitter.com/davidad/status/1657771077035073536 i keep alternating between laughing and disquieting dread
19:14 < hprmbridge> lachlan> Maybe we already uploaded humans, but we accidentally uploaded a bunch of them at once into a single model
19:15 < fenn> a tiny slice of every human online
19:16 < fenn> YOU get a slice, and YOU get a slice
19:18 < superkuh> anthropic-ai (or something with that useragent and appropriate IP ranges) has been mirroring my ~/library/ for a few days now.
19:19 < superkuh> "appropriate" being just amazon ec2.
19:20 < superkuh> I was thinking about what documents other than chicken.pdf I might put in there to poison the n-gram statistics ever so slightly. http://superkuh.com/anthropicai-robot.txt
19:20 < fenn> why did the idea behind common crawl fail? it's a coordination failure for every bot to download every file on the internet
19:40 < fenn> .t https://github.com/newhouseb/clownfish
19:40 < EmmyNoether> GitHub - newhouseb/clownfish: Constrained Decoding for LLMs against JSON Schema
19:40 < fenn> .t https://twitter.com/GrantSlatton/status/1657559506069463040
19:40 < EmmyNoether>  https://nitter.nl/pic/orig/media%2FFwDTXcNaYAIafOg.jpg
19:40 < fenn> uh no
19:40 < fenn> "Added arbitrary context free grammar constraints to llama.cpp
19:40 < fenn> Can now plug in any llama.cpp compatible model and give an exact grammar spec: JSON, etc"
20:40 -!- test_ [flooded@gateway/vpn/protonvpn/flood/x-43489060] has joined #hplusroadmap
20:44 -!- flooded [flooded@gateway/vpn/protonvpn/flood/x-43489060] has quit [Ping timeout: 268 seconds]
21:29 -!- SDr [~SDr@li1189-192.members.linode.com] has quit [Changing host]
21:29 -!- SDr [~SDr@user/sdr] has joined #hplusroadmap
22:04 -!- hellleshin [~talinck@108-225-123-172.lightspeed.cntmoh.sbcglobal.net] has joined #hplusroadmap
22:04 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
22:05 -!- Hooloovoo [~Hooloovoo@hax0rbana.org] has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
22:06 -!- Hooloovoo [~Hooloovoo@hax0rbana.org] has joined #hplusroadmap
22:06 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 240 seconds]
22:06 -!- stipa_ is now known as stipa
22:07 -!- helleshin [~talinck@108-225-123-172.lightspeed.cntmoh.sbcglobal.net] has quit [Ping timeout: 264 seconds]
23:31 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
23:32 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 240 seconds]
23:32 -!- stipa_ is now known as stipa
--- Log closed Thu May 25 00:00:11 2023