--- Log opened Thu Feb 08 00:00:02 2024
--- Day changed Thu Feb 08 2024
00:00 < fenn> it feels like we're at the point where we need to work on critical thinking and teaching epistemology rather than just working on data quality
00:00 < fenn> and here most humans are out of their depth
00:01 < fenn> at some point you will hit the end of the accumulated "data overhang" (i just made that up) and the AI will need to start doing real world experiments. is that 2025? sooner?
00:02 < fenn> there is a lot of dark data that could be liberated...
00:07 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
00:07 -!- TMM_ [hp@amanda.tmm.cx] has joined #hplusroadmap
00:11 -!- alethkit [23bd17ddc6@sourcehut/user/alethkit] has quit [Remote host closed the connection]
00:12 -!- alethkit [23bd17ddc6@sourcehut/user/alethkit] has joined #hplusroadmap
00:13 -!- alethkit [23bd17ddc6@sourcehut/user/alethkit] has quit [Remote host closed the connection]
00:20 -!- alethkit [23bd17ddc6@sourcehut/user/alethkit] has joined #hplusroadmap
00:25 < fenn> https://github.com/discus0434/lcm-sketch-progression/raw/main/assets/sample.gif
00:31 < hprmbridge> kanzure> https://old.reddit.com/r/LocalLLaMA/comments/1aldqx1/best_llm_with_large_context_window_not_for_coding/
00:34 < fenn> miqu is not "open" though, it's a leaked model
00:35 < fenn> possibly the llama-2 license applies, depending on how they did the training, but mistral isn't saying (yet)
00:35 < fenn> vast.ai prices are up from $0.5/hr to $3/hr for 4090s due to miqu
00:38 < fenn> it's kind of like asking "what's the best car" - it depends on what you're doing, your tastes, what you can afford, how much work you want to put in
01:15 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has joined #hplusroadmap
01:19 -!- BEXCHA [~Malvolio@idlerpg/player/Malvolio] has joined #hplusroadmap
01:37 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has left #hplusroadmap []
03:25 < hprmbridge> michaelinzo> Good morning anyone here participating in the Gitlab and AWS Devpost hackathon project? dm me let's collab
03:58 < hprmbridge> Eli> Ya, I'm not sure what he was referencing since I'm not an AI expert by any means. But it seems like llama3 and gpt5 will be incremental improvements (we were going over llama3 with him). And they've pretty much ingested all the text on the internet, or is that incorrect?
03:58 < hprmbridge> Eli>
03:58 < hprmbridge> Eli> llms can make very cool applications. I've made one myself. But, I see a lot of just strapping a UI to an llm api call. And then doing fine tuning maybe.
03:58 < hprmbridge> Eli>
03:58 < hprmbridge> Eli> I think the real issue I wonder about is how to we get to what's next? Maybe I'm somehow jaded by all the incredible things that AI/ML is already doing now. And maybe gpt4 is better than most humans at many tests. But gpt4 cannot inherently understand that a woman pushing a stroller is what causes the stroller to move. Whereas, for a child, it's trivial, and requires only 20W of brain power. The
03:58 < hprmbridge> Eli> easy things are hard. And I don't understand the causal chain of how we get to the point of where we are now to getting AI/ML to do what any child can do. Like, what are the rungs on the ladder that we know we need to develop to get us to "common sense"? More data, compute, and efficient algos? Is that really it?
04:01 < hprmbridge> Eli> One slide I saw is that llama3 is using 70B parameters because there's not really a major improvement after that. So, maybe he was referring to that as the reason he thinks gpt5/6 is it?
05:06 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
05:10 < hprmbridge> .maybeman> Whats up guzys
05:49 < kanzure> hello maybeperson
05:53 < nsh> .w sheheit
05:53 < saxo> Couldn't get any definitions for sheheit
05:53 < EmmyNoether> HTTPError: HTTP Error 404: Not Found (w:24)
05:54 < hprmbridge> kanzure>  https://cdn.discordapp.com/attachments/1064664282450628710/1205149766275833876/GFiFsIraoAAPWX2.png?ex=65d7522e&is=65c4dd2e&hm=bf99b5c1a90e88412d57a574a3886075c0680fa38ddcfde7faba625699e7677f&
06:54 -!- 076AAK90M is now known as cc0
07:49 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 255 seconds]
07:50 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
07:54 < superkuh> Oh no. It's finally happened. Google put scholar.google.com behind a computational paywall.
07:57 < kanzure> captcha? or something else?
07:57 < superkuh> javascript required captcha.
08:09 < nsh> the enclosure of the commons continues
08:10 < kanzure> do academic journals still voluntarily show the fulltext to google scholar?
08:10 < nsh> a time comes for steps towards revolutionary redress of appropriative excesses
08:14 < nsh> .wety Stosszahlansatz
08:14 < saxo> Couldn't get an etymology for Stosszahlansatz
08:14 < nsh> .wety stosszahlansatz
08:14 < saxo> Couldn't get an etymology for stosszahlansatz
08:16 < nsh> .gpt undertake an etymological analysis of the German compound word Stosszahlansatz
08:16 < gptpaste> ​The German compound word "Stosszahlansatz" is composed of three parts: "Stoss," "Zahl," and "Ansatz."1. Stoss: The word "Stoss" comes from the Middle High German word "stōz" which means a push or a blow. It can also be traced back to the Old High German word "stōzzen." This root is related to the Proto-Germanic word "*stautaną," meaning to strike or hit.2. Zahl: The  - https://bin.gy/loldracelo
08:17 < nsh> used by Maxwell in the context of statistical mechanics as the chaos of molecular interactions
08:18 < nsh> cf. [[Principle_of_maximum_entropy]]
08:19 < nsh> or rather to Ehrenfest we credit the coinage
08:21 < nsh> worth thinking about the assumptions that go into mathematical models to make them workable 
08:25 < nsh> '"The relative velocities of a pair of gas particles are uncorrelated to the relative velocities of the other gas particles."'
08:45 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
08:46 -!- TMM_ [hp@amanda.tmm.cx] has joined #hplusroadmap
09:36 -!- boxy [~box@86.120.225.32] has quit [Remote host closed the connection]
10:03 -!- cthlolo [~lorogue@77.33.24.3.dhcp.fibianet.dk] has joined #hplusroadmap
11:09 < kanzure> autocad's john walker has died https://news.ycombinator.com/item?id=39297185 https://scanalyst.fourmilab.ch/t/john-walker-1949-2024/4305
11:51 -!- cthlolo [~lorogue@77.33.24.3.dhcp.fibianet.dk] has quit [Quit: Leaving]
12:19 < hprmbridge> nmz787> Getting kid into school has been the biggest upgrade to my social life since.... College
12:20 < hprmbridge> nmz787> Found another set of parents who both have phds in biochem/bioengi and work together, publish together, and work on single cell sequencing research... One guy has experience with genome assembly (but not synthesis)
15:47 < hprmbridge> soul_syrup> game metadata encoded biomimetic neural signal generation for bio silicon synergetic learning system (for playing DOOM) https://cdn.discordapp.com/attachments/1064664282450628710/1205298803226181642/biomimetic_signals.png?ex=65d7dcfc&is=65c567fc&hm=f347c2fd496df6628541aebf79aac306385eae8b2a29cb5e93a9c05def0be61f&
16:36 < hprmbridge> kanzure> how gibson assembly got its name https://twitter.com/pdhsu/status/1755681802600210597
16:39 < hprmbridge> kanzure> "Still weird to me that Gibson assembly was actually invented by a guy called Gibson who is still like hanging out and working in science. It's as if you saw someone at a conference and everyone's like "that's Michael PCR, you know, as in the inventor of pcr""
16:39 < hprmbridge> kanzure> "it's crazy that we never talk about how Lauren Southerns dad invented such an important technique"
16:40 < hprmbridge> kanzure> "Gibson assembly was named "one-step isothermal in vitro recombination""
16:49 < hprmbridge> kanzure> "The Cambridge iGEM team of 2010 made a very serious and intentional effort (including a song) to get people calling it Gibson Assembly specifically to try and widen its use and promote its adoption as an alternative to BioBrick assembly." from tom ellis
16:57 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 255 seconds]
17:06 -!- TMM_ [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
17:06 -!- TMM_ [hp@amanda.tmm.cx] has joined #hplusroadmap
17:35 -!- darsie [~darsie@84-113-55-200.cable.dynamic.surfer.at] has quit [Ping timeout: 268 seconds]
17:41 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has left #hplusroadmap []
17:49 < kanzure> (to be clear that is southern blot)
17:58 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
19:03 < hprmbridge> Eli> Do we need to go through FDA trials to bring cryonically frozen people back to life? 🤔🤔🤔
19:44 < jrayhawk> at some point you need to treat the underlying cause of death, so, probably
19:51 < hprmbridge> Eli> Isn’t the theory that they’re brought back to life when we have a cure?
19:53 < jrayhawk> bring me back when they develop a cure for the FDA
20:47 -!- mxz [~mxz@user/mxz] has quit [Ping timeout: 252 seconds]
21:45 -!- Guest35 [~Guest35@216.80.40.254] has joined #hplusroadmap
21:45 < Guest35> kanye west
21:46 -!- Guest35 [~Guest35@216.80.40.254] has quit [Client Quit]
22:54 -!- mxz [~mxz@user/mxz] has joined #hplusroadmap
23:49 < fenn> kanye west did not invent the western blot, sadly
--- Log closed Fri Feb 09 00:00:53 2024