--- Log opened Thu May 01 00:00:29 2025
00:02 < hprmbridge> nmz787> As a person on the Intel physical side, I hereby challenge you to a duel
00:03 < hprmbridge> nmz787> Intel design has been shittier. They can't even profit well when having TSMC make the physical stuff.
00:03 < hprmbridge> nmz787> /me gloveslap
00:06 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
00:15 < hprmbridge> kanzure> do they let you just duel in their name like that
00:39 -!- NewtonTrendy [Newton@51.254.45.134] has quit [Quit: WeeChat 3.5]
00:44 -!- flyback [~flyback@2601:540:c700:2380:3da6:bfa4:b220:471] has quit [Ping timeout: 248 seconds]
00:45 -!- flyback [~flyback@2601:540:c700:2380:3da6:bfa4:b220:471] has joined #hplusroadmap
00:54 -!- NewtonTrendy [Newton@51.254.45.134] has joined #hplusroadmap
01:02 -!- Gooberpatrol_66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
01:02 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has quit [Read error: Connection reset by peer]
01:04 < jrayhawk> uh... what products have they released on TSMC processes so far?
02:11 < L29Ah> how bad does llama.cpp scale over multiple pcie-connected gpus?
03:02 -!- balrog [znc@user/balrog] has quit [Ping timeout: 252 seconds]
03:02 -!- balrog_ [znc@user/balrog] has joined #hplusroadmap
04:46 < hprmbridge> kanzure> https://github.com/ahujasid/blender-mcp
05:38 < hprmbridge> nmz787> jrayhawk GPT4 says:
05:38 < hprmbridge> nmz787> 1. Intel Arc GPUs (Alchemist series) – Manufactured on TSMC N6 (6nm).
05:38 < hprmbridge> nmz787>
05:38 < hprmbridge> nmz787> 2. Intel Core Ultra “Meteor Lake” (2023/2024) – The graphics tile and some SoC tiles are made on TSMC N5 and N6, respectively.
05:38 < hprmbridge> nmz787>
05:38 < hprmbridge> nmz787> 3. Intel Ponte Vecchio (Data Center GPU) – A complex chiplet-based design using TSMC N5, N7, and Intel 7 (Foveros packaging).
05:38 < hprmbridge> nmz787>
05:38 < hprmbridge> nmz787> 4. Intel Gaudi 2 AI Accelerator (from Habana Labs) – Made on TSMC 7nm.
05:39 < hprmbridge> nmz787> I'm not hip to the marketing speak, and don't really buy computers often enough to pay attention to the sales/marketing side of things
06:21 < jrayhawk> ARC was always planned as a loss leader due to lack of driver/developer inertia, Intel didn't design the Gaudi 2, amd those other two require Intel processes.
06:25 < jrayhawk> er, and
07:07 -!- 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.]
07:08 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
07:30 < hprmbridge> jsingh.1> I’ve always been curious on this, are there any things that make sf energy drinks worse than coffee?
07:41 < TMA> jsingh.1: some are sweetened with sucralose or some other artificial sweetener; the colorants used (e.g. Quinoline Yellow WS == E104) have been accused of bad things
07:43 < TMA> jsingh.1: OTOH coffee might have mycotoxins and/or other stuff (tar compounds created during the roasting) that are not screened for
07:47 < jrayhawk> https://www.cell.com/cell-reports/fulltext/S2211-1247(24)01460-8 oh, tau tangles are an immune response. that explains a few things.
07:59 < hprmbridge> jsingh.1> oh interesting, thank you
08:00 < hprmbridge> jsingh.1> yeah ive never seen any major studies comparing them
08:00 < hprmbridge> jsingh.1> well studies with fruitful results
08:41 < jrayhawk> jsingh.1: sugar-alcohol-based drinks still trigger calorie estimation in the cephalic stage of digestion and act as a fermentable substrate for the gut microbiome in highly individual ways
08:53 < jrayhawk> there are also various issues with estrogenic activity in plastic linings of cans used to limit aluminum exposure, and aluminum itself is always suspicious https://link.springer.com/article/10.1007/s00204-024-03903-2
08:56 < jrayhawk> caffeine is generally a lousy stimulant for how messy it is; i think people capable of understanding u-shaped dose/response curves are better off experimenting with cholinergics nad dopaminergics separately
08:56 < hprmbridge> jsingh.1> interesting so basically anything in aluminum can be dangerous?
08:59 < jrayhawk> that's not particularly well-established epidemiologically; if an effect exists, it is small enough to be hard to study.
09:01 < hprmbridge> jsingh.1> got it, how do you guys structure your diet around training and things
09:01 < hprmbridge> jsingh.1> or rules that you believe everyone should try to follow
09:05 < L29Ah> i strive not to eat for ~1.5h before bed or HIIT session
09:07 < hprmbridge> jsingh.1> what about amt of food/ types of food balane
09:07 < hprmbridge> jsingh.1> how long are your HIIT sessions?
09:11 < jrayhawk> As a rule, food quality should be conceptualized in terms of immunological safety, oxidative safety, micronutrient concentration, and micronutrient bioavailability. Failure to do so is the central cause of ~8 of the top 10 causes of death in western nations such as the United States.
09:20 < jrayhawk> Human RCTs on the subject: https://pubmed.ncbi.nlm.nih.gov/30050374 https://pubmed.ncbi.nlm.nih.gov/25661189 https://pubmed.ncbi.nlm.nih.gov/28440046 https://pubmed.ncbi.nlm.nih.gov/24473459 https://pubmed.ncbi.nlm.nih.gov/17522610 https://pubmed.ncbi.nlm.nih.gov/23414424 https://pubmed.ncbi.nlm.nih.gov/23890471 https://pubmed.ncbi.nlm.nih.gov/25828624 https://pubmed.ncbi.nlm.nih.gov/25304296 
09:20 < jrayhawk> https://pubmed.ncbi.nlm.nih.gov/27216013 https://pubmed.ncbi.nlm.nih.gov/27235022 https://pubmed.ncbi.nlm.nih.gov/27223304 https://pubmed.ncbi.nlm.nih.gov/21118562 https://pubmed.ncbi.nlm.nih.gov/19604407 https://pubmed.ncbi.nlm.nih.gov/17583796 https://pubmed.ncbi.nlm.nih.gov/26003334 https://pubmed.ncbi.nlm.nih.gov/26786351 https://pubmed.ncbi.nlm.nih.gov/33844651
09:25 < L29Ah> jsingh.1: i eat https://github.com/l29ah/muesli as my default food, and have some ideas on what i refuse when treated or otherwise may eat unless starving, like smoked or fried food
09:42 < hprmbridge> jsingh.1> interesting so you guys are behind the eat as little as possible type of mindset is better for longevity?
09:47 < jrayhawk> No. Long-term calorie restriction has not had a consistent demonstrable benefit for longevity in any of rodents, primates, or humans.
09:50 < jrayhawk> https://pmc.ncbi.nlm.nih.gov/articles/PMC3476836/
10:04 < jrayhawk> If you norm on pathology, you can make anything look pathological. For instance, if you norm on a society that insists upon eating CXCR3 and opioid receptor agonists that drive up cytokine expression, you should not be too surprised when calories look pathological. https://diabetesjournals.org/diabetes/article/62/8/2629/34105 https://journals.physiology.org/doi/full/10.1152/physrev.00003.2008 
10:04 < jrayhawk> https://gnusha.org/~jrayhawk/the_metabolic_syndrome/10.1111-jfbc.12629-role_of_food-derived_opioid_peptides_in_the_central_nervous_and_gastrointestinal_systems.pdf
10:07 < L29Ah> jsingh.1: i don't have a strong opinion on amounts; i'm determining the amounts instinctively and failed to gain much weight when tried to overfeed myself with junk food
10:08 < L29Ah> literal "as eat as possible" would demand me to reduce physical activity, and i find it more important for longevity
10:08 < L29Ah> *little
10:13 < L29Ah> jsingh.1: i do https://hiitscience.com/perfect-workout-high-intensity-interval-duration/ with ten 30-20 intervals
10:28 < jrayhawk> 'Dr. Jay Bhattacharya has announced that the publisher anti-paywall "NIH Public Access Policy", originally slated to go into effect on December 31, 2025, will now be effective as of July 1.'
10:29 < jrayhawk> https://www.nih.gov/about-nih/who-we-are/nih-director/statements/accelerating-access-research-results-new-implementation-date-2024-nih-public-access-policy
10:30 < jrayhawk> '[...]only about 25% of Americans have a “great deal of confidence” that scientists are working for the public good. Earlier implementation of the Public Access Policy will help increase public confidence in the research we fund[...]'
10:32 < jrayhawk> government competence is so disorienting
11:09 < L29Ah> https://juick.com/i/p/3073923.png
13:26 < hprmbridge> poppingtonic> hey all, Msep.one got a new release, so I made time to load and validate an atomistic lock based on diamond links and single-atom rotational joints, from Ralph Merkle's mechanical computing systems paper.
13:26 < hprmbridge> poppingtonic>  https://github.com/mattmoses/MechanicalComputingSystems https://cdn.discordapp.com/attachments/1064664282450628710/1367598054991204464/Screenshot_from_2025-05-01_17-06-13.png?ex=68152a71&is=6813d8f1&hm=53ebc13383299460308ae72cc4c7b4f2b7621727c8e3f81c51fb5fc1aed8ed34&
13:49 < hprmbridge> kanzure> @poppingtonic hi, is that your git repo
13:50 < hprmbridge> kanzure> oh, 7 years ago
14:18 < kanzure> https://norvig.com/chomsky.html
14:43 < hprmbridge> poppingtonic> No it's not I have just been thinking about the design since I saw this Ralph Merkle talk about it at CCC. And yeah it is ancient in msep years  😆 https://m.youtube.com/watch?v=yVX9Ob4SjGA
14:45 < L29Ah> kanzure: this looks old
14:57 < hprmbridge> nmz787> jrayhawk I'd hardly call Intel packaging some TSMC chips cause to ignore that the active transistors aren't Intel's
14:59 < hprmbridge> nmz787> my point really was that Intel designs seem to suck more than the silicon the fab is spitting out
14:59 < hprmbridge> nmz787> we shall see though, as 1278 moves into HVM
15:00 < hprmbridge> nmz787> (for external customers(
15:00 < hprmbridge> nmz787> )
15:05 < hprmbridge> nmz787> and re Habana labs: https://www.reddit.com/r/mlscaling/comments/1ilp3q0/how_intel_ruined_an_israeli_startup_it_bought_for/
15:06 < hprmbridge> nmz787> on a more interesting/relevant note:
15:06 < hprmbridge> nmz787> https://news.nus.edu.sg/advancing-semiconductor-devices-for-artificial-intelligence/
15:07 < hprmbridge> nmz787> """The NUS research team has now demonstrated that a single, standard silicon transistor, when arranged and operated in a specific way, can replicate both neural firing and synaptic weight changes — the fundamental mechanisms of biological neurons and synapses. This was achieved through adjusting the resistance of the bulk terminal to specific values, which allow controlling two physical phenomena
15:07 < hprmbridge> nmz787> taking place into the transistor: punch through impact ionisation and charge trapping. Moreover, the team built a two-transistor cell capable of operating either in neuron or synaptic regime, which the researchers have called "Neuro-Synaptic Random Access Memory", or NS-RAM."""
15:07 < hprmbridge> nmz787> https://www.nature.com/articles/s41586-025-08742-4
15:31 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
15:31 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
16:16 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
16:16 -!- Gooberpatrol_66 [~Gooberpat@user/gooberpatrol66] has quit [Ping timeout: 260 seconds]
16:27 < hprmbridge> Eli> I wonder what this looks like in practice. They've tried to force open access for science papers in the past. The publishers don't like it and fight back. I think if everything becomes open access no one will pay subscription fees and they can only get money from article processing charges.
16:27 < hprmbridge> Eli> I did see this awesome piece of news:
16:27 < hprmbridge> Eli>
16:27 < hprmbridge> Eli> An antitrust class action lawsuit has been filed against the world's six largest for-profit publishers of peer-reviewed scholarly journals. The lawsuit alleges that these publishers have colluded to unlawfully appropriate billions of dollars that should have funded scientific research. The lawsuit aims to recover damages for the class members and to obtain an injunction against the publishers'
16:27 < hprmbridge> Eli> alleged anticompetitive conduct.
16:27 < hprmbridge> Eli>
16:27 < hprmbridge> Eli> https://www.documentcloud.org/documents/25555477-doc1-complaint-uddin/
16:42 < hprmbridge> Eli> It's a complex question. AFAIK, this is the best study on CR in humans. Seems to at least have improved the biomarkers they measured. Probably eating too much is not great. https://academic.oup.com/nutritionreviews/article/79/1/98/5907056
17:00 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Ping timeout: 260 seconds]
17:31 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
17:34 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 244 seconds]
17:34 -!- stipa_ is now known as stipa
18:08 < hprmbridge> Lev> To be fair at least in the US there's been a bunch of unis etc cancelling subscriptions over the last few years because of subscription fee pricing
18:09 < hprmbridge> Lev> Doubt that there'll be a new "everything open" status quo, but maybe (hopefully) some corrections to pricing & distribution
19:23 -!- flyback [~flyback@2601:540:c700:2380:3da6:bfa4:b220:471] has quit [Quit: Leaving]
19:28 -!- flyback [~flyback@2601:540:c700:2380:c487:6e5f:b096:fe5b] has joined #hplusroadmap
21:11 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
21:12 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
23:48 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
23:51 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 252 seconds]
23:51 -!- stipa_ is now known as stipa
--- Log closed Fri May 02 00:00:30 2025