--- Log opened Mon Aug 11 00:00:07 2025
00:51 < hprmbridge> kanzure> data storage locator dogs, apparently some dog olfactory behavior trainers are specializing in hard drives and sd cards
00:57 < hprmbridge> .monokhrome> what scent molecules do they give off? Stuff from decaying electronic components?
00:59 < hprmbridge> .monokhrome> I've got all sorts of old computer stuff sitting in boxes around my room, sometimes wonder if they give off any harmful stuff as they get old
02:02 < hprmbridge> kanzure> weird- carl shulman is not someone i pinned as having money, wonder what he will do with some spare dollars....
02:37 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
02:37 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
04:36 < fenn> administering a fund doesn't mean you own the money in the fund
04:47 < L29Ah> MuaddibLLM: for how long should i store DHA-rich fish oil in a fridge without protective atmosphere?
04:47 < MuaddibLLM> Short answer: finish DHA-heavy fish oil in 4–6 weeks in the fridge (8 weeks max if it’s cold, dark, tightly sealed with minimal headspace), or freeze portions; toss at any fishy/paint-like smell.
04:48 < fenn> MuaddibLLM: what if it had rosemary extract added at the factory
04:48 < MuaddibLLM> ROE (carnosic acid/carnosol) buys you ~2–3× more oxidation headroom, so think 2–3 months in the fridge after opening (vs ~4–6 weeks), longer if you freeze in small portions, but still keep it cold/dark, minimize headspace, and bin it at any fishy/varnish smell.
04:49 < fenn> i think AI will make a pretty good tutor, at least up until PhD level, if such a thing still exists by then
04:50 < L29Ah> MuaddibLLM: would displacing air with butane extend the shelf life further?
04:50 < MuaddibLLM> Purging headspace helps, but skip butane (lighter gas has mercaptans, dissolves in oil, flammable); use argon/nitrogen wine preserver or freeze in small portions instead, and still trust your nose.
04:50 < fenn> can you get dry ice in $country
04:51 < L29Ah> idk
04:51 < L29Ah> i think it will freeze my entire freezerless fridge
04:52 < fenn> just to generate a mostly inert atmosphere inside the bottle
04:52 < fenn> whatever happened to ampoules
04:53 < fenn> those were great
04:54 < L29Ah> i had success making ampoules with a regular butane torch but it makes little sense with DHA
04:54 < fenn> it would actually keep the oxygen out
04:55 < L29Ah> yes but you need a lot of DHA, an ampoule this large would be hard to make
04:56 < L29Ah> better use a glass jar or smth
04:56 < fenn> PVOH is 4 orders of magnitude less oxygen permeability than PE
04:57 < fenn> sorry 5 OOM
04:58 < L29Ah> The oxygen transmission rate of PE, BOPP and PET were 1603.578 cm3/(m2·24h·0.1MPa)、625.068 cm3/(m2·24h·0.1MPa)、16.953 cm3/(m2·24h·0.1MPa)
05:00 < fenn> bad bad and pretty bad
05:00 < fenn> see figures on pg 20 and 21 https://repositorio-aberto.up.pt/bitstream/10216/88820/2/25962.pdf
05:01 < fenn> seems like PVDC is even better at low temps
05:02 < fenn> "Saran"
05:02 < fenn> "Today's Saran Wrap is no longer composed of PVDC in the United States"
05:03 < L29Ah> i don't know where to get such fancy bottles
05:03 < L29Ah> but i have a lot of glass for longer term storage
05:04 < fenn> "in a bold decision to favor public health and the environment over profits" right
05:19 < fenn> o4-mini is telling me that properly crimped aluminum foil pouch is as good as a heat sealed foil lined plastic pouch for oxygen permeability, but glass bottle with metal cap is better, and ampoule is better than that
05:24 < L29Ah> MuaddibLLM: would adding copious amounts of ascorbic acid to fish oil save it from rancidification?
05:25 < MuaddibLLM> Short answer: nope—dumping vitamin C (water‑soluble) into oil won’t stop rancidity and can even go pro‑oxidant with trace metals; instead use oil‑soluble antioxidants (mixed tocopherols ± ascorbyl palmitate/rosemary), exclude O2 (nitrogen flush, tiny headspace), keep cold/dark, and use glass/foil—ampoule beats all.
05:25 < fenn> there'd be residual dissolved oxygen in the fish oil itself, so i guess you'd want a PP bottle in the pouch with an oxygen absorber (steel wool stuff)
05:26 < fenn> so the oxygen could get out eventually
05:26 < L29Ah> i wonder where would i get food grade iron powder to put in the oil
05:26 < fenn> don't do that
05:27 < L29Ah> why not?
05:27 < fenn> it's a catalyst and will oxidize the fish oil more rapidly
05:28 < pasky> let me try to prompt out the damn em-dashes
05:29 < fenn> i don't really have faith in my ability to crimp foil so i'd probably use a mason jar
05:30 < fenn> foil blocks light so maybe both
05:31 < L29Ah> apparently there's about 38mg of O2 dissolved in oil left at STP air // https://www.jstage.jst.go.jp/article/jos1956/16/3/16_3_108/_pdf/-char/en
05:31 < L29Ah> *in a litre of oil
05:31 < L29Ah> there's no shortwave light in my fridge
05:31 < fenn> oo japanese OCR
05:32 < L29Ah> page 4 plot
05:33 < fenn> maybe you could purge the dissolved oxygen with nitrogen
05:34 < L29Ah> i think i can't
05:34 < fenn> or vacuum
05:34 < L29Ah> vacuum is doable but very annoying
05:42 < fenn> "Place the bottle of fish oil (and oxygen absorber packet) inside your airtight glass jar. Small charge of dry ice. Start with 5 g per liter (of jar); this generates roughly 2 L of CO₂ gas—enough to purge without overpressurizing. Once most of the air is expelled (hissing slows/stops), tighten the lid fully."
05:43 < fenn> CO2 won't react
05:45 < fenn> locally they only have this lemon flavored xylitol fish oil goop in a huge expensive bottle which i can never finish
05:45 < L29Ah> do they at least specify what sort of fish it was produced from, and how, or at least DHA content?
05:45 < fenn> and the pills are always bad by the time you get them
05:45 < fenn> one moment
05:46 < L29Ah> here even flaxseed oil is rancid on arrival
05:47 < fenn> Barlean's Seriously Delicious Omega-3 from fish oil natural flavor 3x better absorption. taste the fruit, not the fish. dietary supplement
05:48 < L29Ah> in .ru and nordic countries there is nice canned cod liver, but here in .me it costs a fortune
05:50 < fenn> ingredients: water, fish oil (anchovy, mackerel, sardine and/or tuna), xylitol, glycerine, gum arabic, natural flavors, citric acid, xanthan gum, antioxidant blend (Vitamin E (as d-alpha tocopherol), green tea extract, rosemary extract, and ascorbyl palmitate), guar gum, turmeric, sorbic acid.
05:52 < fenn> serving size: 1 tbsp (15.9g), contains 29 servings. fat 4g, (containing saturated fat 1.5g, polyunsaturated fat 1.5g, monounsaturated fat 1g,) carbohydrate 6g, (containing sugar alcohol 5g,) total omega-3 fatty acids 1240mg, (containing EPA and DHA 1080mg) 
05:54 < fenn> the expiration date is 9 months from when i bought it but there's no manufacturing date or at least i don't know what 25001618 means
05:55 < fenn> i find the huge amounts of xylitol concerning
06:03 < hprmbridge> kanzure> administering a fund often means you get 2 and 20 what do you mean. this isn't a charity.
06:04 < hprmbridge> kanzure> AI tutoring is very different from aristocratic tutoring. it has no sense for your world model and how your brain ticks. no theory of mind about the student.
06:06 < hprmbridge> kanzure> why does everyone insist on promoting to remove emdashes. i feel like you guys are vampires sucking my brain juices. just use string replacement or regex. wtf?
06:06 < hprmbridge> kanzure> er, prompting
06:06 < fenn> because it affects sentence structure
06:06 < hprmbridge> kanzure> oh no!
06:06 < fenn> it's not X - it's Y!
06:07 < fenn> classic 2025 AI trope
06:07 < fenn> you can't un-see it
06:07 < fenn> why can't an AI tutor form a theory of mind of the student?
06:08 < hprmbridge> kanzure> hmm I don't think I claimed it's impossible
06:08 < hprmbridge> kanzure> i certainly haven't seen it yet
06:08 < fenn> i really have no idea how people are using AI in education
06:08 < hprmbridge> kanzure> the fancy schools are using "adaptive learning systems"
06:09 < fenn> uh and what is that
06:09 < hprmbridge> kanzure> not sure how they work, but i will be mad if it's merely "choose your own adventure" with skill checks and backtracking
06:09 < fenn> that would be nice
06:09 < hprmbridge> kanzure> you can do that without AI though
06:09 < fenn> i'd add some kind of anki repetition
06:11 < fenn> ok so there's the theory that some people are visual learners, some are kinesthetic, some are audio creatures... presumably you could design experiments to actually measure these values
06:11 < fenn> (as if education were a science)
06:11 < fenn> ahem
06:11 < fenn> but it'd be more fun to toss a bunch of data into a machine learning system
06:12 < fenn> there's also the question of whether it makes sense to force teach people stuff they aren't interested in
06:13 < fenn> whether voluntarily or by state coercion
06:13 < fenn> like if you aren't interested, maybe it's a waste of time, and should wait until later when you *are* interested
06:14 < fenn> if you understand why boring subject X is a prerequisite for Y, sometimes it can become interesting
06:15 < fenn> is adaptive learning and spaced repetition necessary when the student is genuinely interested?
06:17 < fenn> historically it wasn't possible to choose your own adventure. even with computer database driven edutainment, there's still this pressure to produce uniform standardized worker product
06:18 < fenn> i think all that has been obsolete for several decades at least, but somehow there's enough inertia in the institutions and theory and expectations around education that we keep doing prussian infantry training
06:21 -!- andytoshi [~apoelstra@user/andytoshi] has quit [Server closed connection]
06:21 -!- andytoshi [~apoelstra@user/andytoshi] has joined #hplusroadmap
06:31 -!- gl00ten [~gl00ten@2001:8a0:7ee5:7800:46d9:f5c:17a2:432] has quit [Ping timeout: 248 seconds]
07:05 -!- gl00ten [~gl00ten@31.22.148.230] has joined #hplusroadmap
07:07 -!- gl00ten [~gl00ten@31.22.148.230] has quit [Client Quit]
07:08 -!- stipa [~stipa@user/stipa] has quit [Quit: WeeChat 4.6.3]
07:14 -!- stipa [~stipa@user/stipa] has joined #hplusroadmap
07:18 -!- ike8 [e8f913dbdf@irc.cheogram.com] has joined #hplusroadmap
07:26 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
07:27 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 260 seconds]
07:27 -!- stipa_ is now known as stipa
07:44 < hprmbridge> kanzure> "Years ago I asked my surgeon to engrave it onto my patella."
07:50 < fenn> ... you can do that???
07:51 < L29Ah> yes, you can ask
09:12 < hprmbridge> kanzure> https://www.thefp.com/p/the-bride-of-chinese-frankenstein-and
09:16 < hprmbridge> kanzure> "Legwork for finding the lab for designer babies, he says, has been done by a family in Austin whose daughter has a rare genetic disorder, and who support his work."
09:52 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Quit: Avoid fossil fuels and animal products. Have no/fewer children. Protest, elect sane politicians. Invest ecologically.]
09:53 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
11:00 -!- _flood [~flooded@149.102.226.247] has quit [Ping timeout: 245 seconds]
11:11 -!- _flood [~flooded@149.102.226.247] has joined #hplusroadmap
12:14 < kanzure> "theory of mind" tutoring would be nice but i don't think it's in the WWW training data.
12:53 < kanzure> https://gnusha.org/logs/summaries/hplusroadmap-summaries.kimi-k2.2025-08-11.zip
12:54 < kanzure> https://gnusha.org/logs/summaries/kimi-k2-summaries-2025-08-11/
12:54 < kanzure> $16 of kimi k2 to produce this
12:56 < kanzure> definitely some weird stuff in the logs: https://gnusha.org/logs/summaries/kimi-k2-summaries-2025-08-11/2015-08-18.md
12:56 < kanzure> "Hair-as-BCI concept: xtalmath proposed using piezo actuators, electrical pulses, or air puffs to stimulate hair-follicle mechanoreceptors as a high-bandwidth, non-invasive computer-to-brain interface, citing rodent whisker research as biological precedent"
12:56 < kanzure> "Klarismo ($300 MRI scans)"
12:57 < kanzure> some odd artifacts: https://gnusha.org/logs/summaries/kimi-k2-summaries-2025-08-11/2024-01-13.md a lot of these files apparently...
12:57 < kanzure> also here: https://gnusha.org/logs/summaries/kimi-k2-summaries-2025-08-11/2009-02-10.md
12:57 < kanzure> "dumpster-mapping UAV swarms"?
13:01 < kanzure> .wik spin forging
13:01 < saxo> " / In computer network communications, the HTTP 404, 404 not found, 404, 404 error, page not found, or file not found error message is a hypertext transfer protocol (HTTP) standard response code,  to indicate that the browser was able to communicate with a given server, but [...]" - https://en.wikipedia.org/wiki/No_result_found
13:38 < hprmbridge> kanzure> https://nlpcloud.com/home/playground/semantic-search a semantic search playground tool thing.
13:39 < hprmbridge> kanzure> semantic search for images, playground https://convnetplayground.fastforwardlabs.com/
13:47 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
13:47 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
15:02 < kanzure> do all human tumors have a non-human intratumoral microbiome? or only certain human tumors?
15:09 < kanzure> the obvious thing to do would be to have engineered microbes preferentially colonize intratumoral microniches instead of whatever natural bacteria or fungi ends up in there. and you could have various engineered payloads in there.
15:11 < kanzure> FDA has previously approved "bacterial cancer therapy for high-risk, nonmuscle invasive bladder cancer using live-attenuated Mycobacterium bovis (BCG vaccine)".
15:14 < kanzure> "intratumoral synthetic microbial programmed bioreactors"
15:21 < kanzure> okay well the next obvious thing to do is evolve or directed selection of a microbial strain that consistently improves cancer outcomes and competes for those tumoral niches. freeze the strain sample after injection, check back in 6 months, if there's reduced size or other indicator of success then up-select for that strain sample. otherwise down-select. apply machine learning and some other ...
15:21 < kanzure> ...random mutation stuff. and maybe some applied rational design choices for genetic modifications to the strains of course to nudge in various directions.
15:22 < kanzure> while it would be nice to know the actual mechanism of action of tumor size reduction by some friendly microbes, it seems more important to cause positive health outcomes than advancing scientific knowledge of the exact mechanism
15:22 < kanzure> would it be more productive to simply review a complete list of all FDA approvals instead of doing this haphazardly by randomly discovering something in random content we end up reading
15:23 < kanzure> and, likewise, by reviewing other regulatory approvals- or rejections for that matter- from other international FDA equivalents
15:23  * L29Ah votes for cancer-specific antibodies
15:24 < kanzure> i previously assumed there is good reason to not do that, like "you are intentionally selecting for the remaining cancer cells that are antibody resistant" or "the development cycle is too long to get those antibodies produced"?
15:24 < kanzure> if you do not kill all of the cancer cells then you are selecting for antibody-resistant cancer cells that are remaining, right?
15:24 < kanzure> and that is not a good idea right?
15:25 < L29Ah> one can shorten the cycle and slow down the patient
15:25 < L29Ah> how about bacteria-resistant cancer cells?
16:15 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Ping timeout: 260 seconds]
17:31 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Ping timeout: 244 seconds]
17:40 -!- saxo [~saxo@zug.infomesh.net] has quit [Server closed connection]
17:40 -!- saxo [~saxo@zug.infomesh.net] has joined #hplusroadmap
17:43 < fenn> apparently the "not x, but y" AI slop leaderboard is only on twitter https://x.org/sam_paech/status/1942714743476871272
17:44 < fenn> his main thing is eq-bench and it does have a "slop forensics" but you can't see them ranked on a single tell like this
18:01 < jrayhawk> fenn: while i appreciate your attempts to direct interest towards x.org, i am not actually inclined to write an apache rule to make that URL work
18:54 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
18:56 < L29Ah> kanzure: gnusha.org seems down
18:57 < hprmbridge> kanzure> no?
19:09 < L29Ah> mtr -n says the last alive hop is 64.125.20.213
19:11 < fenn> could you provide an internet node near your own gateway?
19:11 < jrayhawk> Hmm. That sounds like a problem with my upstream.
19:11 < fenn> to see where the problem is from our end
19:12 < jrayhawk> Yeah, let me know what subnet is getting dropped.
19:12 < L29Ah> src is 62.4.55.85
19:13 < jrayhawk> oops, yes, that does look like a bad route. i will pester them.
19:15 < fenn> for the bots https://x.com/sam_paech/status/1942714743476871272
19:15 < fenn> i am doing it wrong and manually editing twitter urls all the time >_<
19:15 < jrayhawk> you and me both
19:52 < hprmbridge> kanzure> "Foliar mRNA spray induces protein synthesis in monocot crop and dicot model plant species" https://www.biorxiv.org/content/10.1101/2025.08.07.668951v1
20:00 < hprmbridge> kanzure> more gpt-5 science refusals https://x.com/JessicaSacher/status/1954985111088402777
23:15 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
23:32 -!- WizJin_ [~Wizzy@2402:a00:184:8354:3ccc:e206:c376:1b02] has joined #hplusroadmap
23:32 -!- WizJin_ [~Wizzy@2402:a00:184:8354:3ccc:e206:c376:1b02] has quit [Remote host closed the connection]
23:32 -!- WizJin_ [~Wizzy@2402:a00:184:8354:3ccc:e206:c376:1b02] has joined #hplusroadmap
23:34 -!- WizJin_ [~Wizzy@2402:a00:184:8354:3ccc:e206:c376:1b02] has quit [Client Quit]
--- Log closed Tue Aug 12 00:00:08 2025