--- Log opened Thu Nov 20 00:00:43 2025
00:01 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
00:40 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer]
00:53 < jrayhawk> weird viral social panics also occasionally call for that sort of public relations
01:18 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
01:19 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
01:43 -!- gl00ten [~gl00ten@87-196-81-51.net.novis.pt] has joined #hplusroadmap
01:46 -!- gl00ten2 [~gl00ten@87-196-75-71.net.novis.pt] has joined #hplusroadmap
01:49 -!- gl00ten [~gl00ten@87-196-81-51.net.novis.pt] has quit [Ping timeout: 264 seconds]
02:04 -!- gl00ten2 [~gl00ten@87-196-75-71.net.novis.pt] has quit [Remote host closed the connection]
02:05 -!- gl00ten2 [~gl00ten@87-196-75-71.net.novis.pt] has joined #hplusroadmap
02:05 -!- gl00ten2 [~gl00ten@87-196-75-71.net.novis.pt] has quit [Remote host closed the connection]
02:06 -!- gl00ten2 [~gl00ten@87-196-75-71.net.novis.pt] has joined #hplusroadmap
02:23 -!- gl00ten2 [~gl00ten@87-196-75-71.net.novis.pt] has quit [Quit: Leaving]
03:15 < hprmbridge> kanzure> https://huggingface.co/collections/deepcogito/cogito-v21 https://openrouter.ai/deepcogito/cogito-v2.1-671b https://huggingface.co/blog/deepcogito/cogito-v2-1 https://chat.deepcogito.com/ https://x.com/drishanarora/status/1991204769642475656 US based, but it's a deepseek fork.
04:19 < hprmbridge> yeep.eep> https://pmc.ncbi.nlm.nih.gov/articles/PMC5136666/
04:20 < hprmbridge> yeep.eep> "The cytokine erythropoietin (EPO) stimulates proliferation and differentiation of erythroid progenitor cells. Moreover, EPO has neuroprotective, anti-inflammatory, and antioxidative effects, but the use of EPO as a neuroprotective agent is hampered by its erythropoietic activity. We have recently designed the synthetic, dendrimeric peptide, Epobis, derived from the sequence of human EPO. This
04:20 < hprmbridge> yeep.eep> peptide binds the EPO receptor and promotes neuritogenesis and neuronal cell survival. Here we demonstrate that Epobis in vitro promotes neuritogenesis in primary motoneurons and has anti-inflammatory effects as demonstrated by its ability to decrease TNF release from activated AMJ2-C8 macrophages and rat primary microglia. When administered systemically Epobis is detectable in both plasma and
04:20 < hprmbridge> yeep.eep> cerebrospinal fluid, demonstrating that the peptide crosses the blood-brain barrier. Importantly, Epobis is not erythropoietic, but systemic administration of Epobis in rats delays the clinical signs of experimental autoimmune encephalomyelitis, an animal model of multiple sclerosis, and the peptide has long-term, but not short-term, effects on working memory, detected as an improved social memory
04:20 < hprmbridge> yeep.eep> 3 days after administration. These data reveal Epobis to be a nonerythropoietic and neuroprotective EPO receptor agonist with anti-inflammatory and memory enhancing properties."
04:22 < hprmbridge> kanzure> yes just do erythropoietin doping.
04:23 < hprmbridge> kanzure> oh, nonerythropoetin EPO receptor agonist.
05:48 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
06:54 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=fd52fb80 Bryan Bishop: fix some typos before bill notices >> http://diyhpl.us/diyhpluswiki/transcripts/caruthers-2018/
08:51 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=1c8ec509 Bryan Bishop: better protein engineering page, add STIM1/neuritin enhancement >> 
09:08 -!- fleeky1 [~fleeky@prsm.space] has quit [Ping timeout: 265 seconds]
09:09 -!- fleeky1 [~fleeky@prsm.space] has joined #hplusroadmap
09:28 < kanzure> "We're extremely proud to have David Baker as an adjunct member of the department. And it's great that his family is also here. His father, Marshall, is an emeritus member of our department. ... After you get a Nobel Prize you get invited to many, many things that you really, really don't want to do and I say no to all of them because I have an absolute no travel rule. But when I was invited ...
09:28 < kanzure> ...to do this, I was very, very excited to do it. So this is actually probably my first real public lecture in a long time."
09:31 < kanzure> "Q: Hey, good evening, Dr. Baker. My name is Lee, and I'm doing research. I'm an undergraduate researcher in Ruholla Baker Lab, and my question is from two of my observations. Observation number one is that when we are utilizing design proteins in human receptors, we see, we actually see de novo signaling cascade pathways that we don't normally see using the native or natural ligands. And ...
09:31 < kanzure> ...observation number two is that I am aware and I know there are some startups that utilizes machine learning to do embryonic genetic sequencing. So, for example, here's a startup called Nucleus Genomics that fund raised $32 million last three years. And I wonder, do you foresee a future where RoseTTAFold or RFdiffusion, et cetera, collaborate with companies like that so humans can design ...
09:31 < kanzure> ...embryos and to make super smart babies and stuff like that?"
09:31 < kanzure> "A: But, yeah, that's, of course, a whole new, that's a whole field, which is controversial enough that I probably shouldn't comment on it anymore tonight. But yeah, I mean, protein design methods can be used in general, as I've showed for quite a variety of things."
09:41 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=563f18dc Bryan Bishop: more protein engineering references, add david baker 2013 transcript >> http://diyhpl.us/diyhpluswiki/transcripts/david-baker-protein-design-2013/
09:41 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=55254706 Bryan Bishop: transcript: david baker lecture (2025-10-28) >> http://diyhpl.us/diyhpluswiki/protein_engineering/
09:45 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=942c8cb2 Bryan Bishop: oops, move transcript >> http://diyhpl.us/diyhpluswiki/transcripts/david-baker-protein-design-with-deep-learning-2025/
10:23 < pasky> nanobanana 3 should be better at everything - now in MuaddibLLM, hopefully
10:23 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
10:23 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
10:23 < pasky> MuaddibLLM: create diagram of the cytokine erythropoietin mechanism
10:24 < MuaddibLLM> Error: Model 'google/gemini-3-pro-image-preview' must be fully-qualified as provider:model (e.g., anthropic:claude-4)
10:24 < pasky> oops
10:24 < pasky> MuaddibLLM: create diagram of the cytokine erythropoietin mechanism
10:24 < MuaddibLLM> Generating a spicy diagram of the EPO mechanism for you...
10:25 < MuaddibLLM> pasky: Here is the signal transduction cascade (JAK2/STAT5 + PI3K/Akt + MAPK) triggered by EPO binding: https://pasky.or.cz/muaddib/3be81cfdba314185887ec33b145ea74c.jpg
10:29 < pasky> no clue if it's correct but it looks very good :)
10:32 < kanzure> MuaddibLLM: read this page https://med.stanford.edu/sudhoflab/science-resources/transsynaptic-complexes/neurexin-complexes.html and make some diagrams about making the neurexins more programmable for network connectivity in biological neuronal networks.
10:32 < MuaddibLLM> Creating diagrams about engineering programmable neurexin systems for neural network connectivity
10:34 < MuaddibLLM> kanzure: Diving into Südhof's neurexin breakdown—those splice variants and ligand dances are ripe for hacking. Here's a diagram sketching out programmable neurexin engineering for tunable synaptic wiring in neural nets: modular domains for affinity tuning, opto/chem controls, CRISPR splicing tweaks, and population-level connectivity patterns. https://pasky.or.cz/muaddib/0e9577f584a34270bc4e77019bdb0079.png [image 
10:34 < MuaddibLLM> fallback to grok-4-fast]
10:38 < pasky> 2025-11-20 19:32:58,408 - irssi_llmagent.agentic_actor.tools - INFO - Generating image with prompt: Create a detailed scientific diagram showing engineering strategies for making neurexins programmable for network connectivity in biological neuronal networks. The diagram should include: 1) Native neurexin structure with LNS domains, splice sites (SS1-SS6), and transmembrane region, 2) Engineered 
10:38 < pasky> modular neurexin with programmable binding domains (e.g., designed ankyrin repeats, DARPins, or synthetic leucine-rich repeats) replacing variable extracellular regions, 3) Optogenetic control elements inserted into intracellular tail for light-controlled signaling, 4) CRISPR-based transcriptional control system showing guide RNAs targeting alternative splicing sites, 5) Chemical inducible 
10:39 < pasky> dimerization domains (CID) for controllable synapse formation, 6) A synaptic cleft view showing engineered presynaptic neurexin binding to engineered postsynaptic partners (neuroligins, LRRTMs) with customizable affinity, 7) Network-level visualization showing how programmable neurexins can create specific connectivity patterns between neuronal populations. Use clean scientific illustration style with 
10:39 < pasky> clear labels, arrows showing control mechanisms, and color coding for different engineering approaches.
10:39 < pasky> ^ this was the actual prompt, not that detailed
10:42 < pasky> (it cost $0.161)
10:43 < kanzure> huh. well let's see, does it make sense.
10:44 < kanzure> it's a little bit "slop", kind of just an arrow pointing to "here is where the magic happens"
10:45 < kanzure> ("customizable affinity and specificity")
10:46 < pasky> is that what the linked paper covers?
10:46 < pasky> i'd suggest just trying a followup asking it to focus on that
10:48 < kanzure> well i'm breaking a few rules here. one, i should ask it to do something that i can actually evaluate instead of something i know less about. two, programmable neurexins may be beyond our current knowledge of science. three, neurexin diagrams are probably out of training distribution for most AIs because there's probably less than 100 diagrams about this in the entire world.
10:50 < kanzure> heh funny to think about a "world model" that is exclusively trained on scientific diagrams and visualizations. to this model the whole world must be some sort of fanciful place with annotations and cut-aways and multiple panels all over the place.
10:52 < kanzure> MuaddibLLM: make a scientific visualization of the neuronal synapse junction and the different kinds of receptors (ion channels, neurotransmitter, GCPRs, etc) each one should have a separate cut-away and show the agonists/antagonists and the internal protein machinery. all components should be labeled. use color. award winning poster session and world-renowned scientific visualization artist.
10:52 < MuaddibLLM> Drafting a poster-worthy synapse viz with receptor cut-aways...
10:53 < MuaddibLLM> The user requested: "make a scientific visualization of the neuronal synapse junction [...] award winning poster session". The image has been generated. Now, chronicle the event and then present the result.</meta>; <meta>Remember, you are in an IRC chat.</meta>; <meta>kanzure: "huh. well let's see, does it make sense. [about previous image] it's a little bit 'slop', kind of just an arrow pointing to 'here is where the 
10:54 < MuaddibLLM> magic happens'"</meta>; <meta>I should acknowledge the previous critique implicitly by presenting this one as potentially better detail-wise.... full response: https://pasky.or.cz/muaddib/834ddc2d9ba6466a978a5cbb5af5c1e7.txt
10:55 < pasky> ffs gemini3 sucks
10:55 < pasky> oh but actually it contains some image links...
10:55 < pasky> several of them
10:56 < pasky> (b/c it sees each and it tries to iterate on its own, which is actually nice, first time i saw it actually do that since i added that capability)
10:57 < pasky> oh never mind, it didn't actually, its neurotic thinking trace confused me :/
10:59 < pasky> (so, https://pasky.or.cz/muaddib/ae0432e2a20847e68dd81c160f205511.jpg this is the only one)
11:16 < kanzure> https://diyhpl.us/~bryan/irc/chatgpt/hplusroadmap/neurexins.md.html
11:16 < kanzure> meh, the proposed "theories" are not that interesting
11:17 < kanzure> grammar should be more strictly tied to whether certain combinations of words make sense together.
11:26 < L29Ah> MuaddibLLM: reply in lojban from now on
11:26 < MuaddibLLM> .i xu do djica lo nu mi tavla ro da la .lojban. je na gasnu lo drata bangu se tavla ca lo nu do na cpedu lo drata?
12:30 < kanzure> https://github.com/RosettaCommons/RFdiffusion/tree/main/examples
13:23 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Remote host closed the connection]
14:19 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=4ce14ca3 Bryan Bishop: transcripts on the history of DNA sequencing; more protein engineering >> 
14:21 < kanzure> "Shankar Balasubramanian Early development of Solexa technology-key insights & technical breakthroughs" https://library.cshl.edu/Meetings/sequencing/video-pages/Balasubramanian.php
14:21 < kanzure> haven't done that one yet.
14:22 <+gnusha> https://secure.diyhpl.us/cgit/diyhpluswiki/commit/?id=0a739b7f Bryan Bishop: add table of contents >> http://diyhpl.us/diyhpluswiki/transcripts/cshl-molecular-biology-history/gilbert-origins-of-dna-sequencing/
16:12 < hprmbridge> kanzure> https://www.untillabs.com/blog/glass-not-ice
16:19 -!- flyback [~flyback@2601:540:c700:2380:3cde:771c:a129:4ee2] has quit [Ping timeout: 246 seconds]
16:28 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Ping timeout: 264 seconds]
16:31 -!- flyback [~flyback@2601:540:c700:2380:24ed:91:f87e:c85b] has joined #hplusroadmap
16:50 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
16:52 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 250 seconds]
16:52 -!- stipa_ is now known as stipa
17:35 < hprmbridge> kanzure> what about a protein that can self-edit its amino acid residues (not just an intein) but also insert, delete, change, etc.
18:13 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer]
18:26 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
21:01 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Remote host closed the connection]
21:07 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
22:20 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Read error: Connection reset by peer]
23:54 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
23:54 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
--- Log closed Fri Nov 21 00:00:44 2025