cheald

Strength is more closely correlated with muscle density than size: https://pubmed.ncbi.nlm.nih.gov/32768372/

Powerlifters who train for strength generally train low-rep, high-weight progressive overload programs and often end up with very dense musculature, while bodybuilders prioritize high-rep sets for large amounts of total volume, which produces relatively more muscle size but not necessarily superior muscle density. The two modes of training produce different hypertrophic effects.

https://pubmed.ncbi.nlm.nih.gov/3057312/

> It appears, though, that bodybuilders, relying on a high repetition training system, in contrast to Olympic weight- and power lifters, display a small increase in number of capillaries per fiber.

> Bodybuilders, as opposed to Olympic weight lifters, power lifters, or other power athletes, tend to display a relatively high percentage of slow switch fibers.

RO systems vary by efficiency, but 1:5 seems way out of whack. My tankless system claims 2:1 output:waste, and while I haven't measured it directly, I'd say that's probably pretty close to right.

"Okay, but how do I open it in Excel?"

Foods rich in good probiotic bacteria, such as strains of lactobacillus and bifidobacterium are a good place to start. Anything cultured or fermented is usually a good bet. There are also probiotic supplements, if you want a particular strain.

My guess would be that certain bacteria or strains are better than others at outcompeting the undesirable colonies, but you'd have to look for specific research to say which ones. In lieu of that, though, it's not hard to experiment with probiotics and see which, if any, have a noticeable positive impact on you.

IANAD, but I'd hypothesize that a healthy oral and gut bacterial biome heavily populated with the various good bacteria may serve the purpose of outcompeting undesirable bacteria and keeping them from growing past critical levels.

Coding on paper should ever only be done as a pseudocode or paper prototyping exercise. There is value in the ability to think abstractly without an IDE prescribing structure for you, but there is zero practical value to testing a developer on whether he can write compiling code by hand on paper. We don't use punch cards anymore.

You can dig postholes with a spoon, and you'll develop some really great forearm strength from doing so, but I'll be over here using a postholer to put up 75 posts for every 1 of yours.

SD 1.5 (using RealisticVision 5.1, 20 steps, Euler A) spit out something technically correct (but hilarious) in just a few generations.

"a woman chasing a bear, pursuit"

https://i.imgur.com/RqCXVYC.png

SD 1.5 is 983m parameters, SDXL is 3.5b, for reference.

Very interesting. I've been streching my 12GB 3060 as far as I can; it's exciting that smaller hardware is still usable even with modern improvements.

Inference should ideally be done on an instance with a GPU, and it should have enough GPU VRAM to hold the entire model. This one is only a 7b model, so it should run pretty easily in a modest amount of vram. You can likely run it locally in CPU-only mode, too, though it's likely to be rather slow.

https://instances.vantage.sh/?min_gpus=1 (wait for it to finish loading to filter)

You could try running these in Google Collab (though tweaks would have to be made to load the files), or you might try something like runpod.io, as well, which gets you GPU instances for a lot less than you'd pay with AWS - for example, a Tesla V100 (16GB) community cloud instance runs you $0.24/hr vs the g5g.xlarge which runs you $0.42/hr.

IMO, the primary target audience for good commit messages is the same target audience as good code comments: me six months from now. Being able to read why and how a particular thing was done has helped me in debugging and troubleshooting an issue on more than one occasion.

The primary perk for a more expensive unit is the ability to modulate that fan speed based on the particulate level in the room, IMO. I've got a couple of units that I paid ~$140 for which can step their fan speed up and down based on the measured air particulates. Most of the time, they're running at low speed and are effectively silent - just cycling the already-pretty-clean air in the room at a low rate - but if something causes a spike in particulates (say, cooking), they'll crank up to a significantly higher speed (and level of noise) until the air is cleared.

If you've got a box fan running all the time anyhow and don't mind the noise, though, strapping a HEPA filter to it makes plenty of sense.

First off, as much as possible, eliminate liquid calories. Those really add up fast and don't contribute to satiation.

Second, try reducing simple sugar/starch intake in favor of complex and higher fiber alternatives where possible. Increasing the ratio of protein and fat in your diet can help you reach satiety much faster, too. Fat is 9kcal/g, protein and carbohydrates are 4 kcal/g, but in terms of satiety, fat > protein >>> carbohydrate.

If you're biohacking-inclined, look into how your specific diet impacts ghrelin (hunger hormone) and leptin (satiety hormone) production and uptake. You might find that you're eating things which provoke hunger or delay satiation. And don't discount that many of us just eat because it makes us feel good - that's a much harder hill to climb, but it is possible.

Red wine in particular has a lot of flavonoids, which are particularly healthful, so it's plausible that there are some health benefits from it. The alcohol itself is probably not conferring any health benefits, though. As with most things, it's probably too simplistic to reduce it down to a single variable.

"kilogram calorie" - it's the unit that we colloquially refer to as a "calorie" in the US for nutrition purposes.

That maintenance requirement is just "keeping the cells alive when you're laying in bed in a coma" - actually using those muscles obviously requires more energy than that. Maintaining muscle mass is more than just a function of calories, though - in part because it's more expensive tissue, the body will try to make up caloric deficits from muscle before fat. Building and maintaining muscle requires that you keep your body in a state of muscle synthesis, which is itself a very metabolically-expensive process.

Numbers for energy/lb/day of muscle range from 6kcal-50kcal depending on who you're listening to, but the thing that everyone agrees on is that regardless of the number, muscle is multiple times more expensive than equivalent fat to maintain.

Congratulations! 55lb is a big chunk of weight to not be carrying around!

It's very true that for weight loss, diet is the single most important variable, but that doesn't mean exercise doesn't matter. The kind of exercise you do matters a lot, too, particularly once you start caring about lean mass vs fat, rather than just the gross scale weight. Cardio has this image as "the way you lose weight", but I think that strength training is vastly superior, overall.

Strength training doesn't have to be ultra time-intensive or strenuous, but it has huge benefits for both health and weight loss; muscle is metabolically-active tissue, taking 6-10 kcal/day/lb to maintain, vs 2-4 kcal/day/lb for fat. This has two implications: first, in a caloric deficit, the body will preferentially catabolize unused muscle tissue over fat, since it's more expensive to maintain, and second, maintaining a greater percentage of your weight as muscle rather than fat increases your daily "just keeping the lights on" caloric expenditure. A caloric deficit with a high-protein diet and even moderate weight training will result in a significantly higher level of retained muscle mass, resulting in more net calories/day burned, even on the days you don't train. Think of it as the difference between active and passive income - the more "passive caloric burn" you can rack up, the easier it is to make progress towards those goals.

Aspirin has "severe (but rare) side effects", too. Many of the OTC pharmaceuticals we take have potentially nasty side effects. The only 100% safe stuff is homeopathic, and that's because it's not actually chemically active. I'd be a lot more convinced by the argument if it weren't widely OTC in most of the rest of the world.

Interestingly, there do seem to be long-term side effects of sildenafil/tadalafil - though they seem to be positive. There is a possible link between long-term usage of ED drugs and cognitive protections in old age:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6705107/

https://content.iospress.com/articles/journal-of-alzheimers-...

New solar has the same T&D costs (plus batteries!), doesn't it?

"Could reach" is a fairly hedged phrasing, there.

The plant generated 16,477 GWh in 2021 [1]. Over 5 years, it'd generate 82,385,000,000 kWh. At the LAX average retail rate of $0.287/kWh [2], that generation would return $23,644,495,000 in revenue.

[1] https://www.eia.gov/todayinenergy/detail.php?id=53899

[2] https://www.bls.gov/regions/west/news-release/averageenergyp...

You can browse code on Github just fine (and search within a file!) - you just can't search through the tree, which is a function of whatever platform the code lives on, be it your box or Github or some other web UI.

You can still clone and grep, can't you? You can't file issues or submit PRs through Github without an account, either.

It's a little too neat to say that "turmeric cures cancer", but laboratory evidence does indeed show that curcumin may have anti-cancer properties.

https://www.cancer.gov/about-cancer/treatment/cam/hp/curcumi...

I'm still broadly convinced that the whole net neutrality outcry back in 2017 was due to a fundamental misunderstanding of how peering works and misplaced anger due to the peering dispute between Cogent and Comcast (and Verizon, and I'm sure plenty of others).

None of the promised apocalypse materialized, because the marketing materials used to make the case for it were fundamentally wrong.

Reddit may not have experts per set, but in the right subreddits it definitely has enthusiasts. In ages gone by, you'd find the same people on message boards or forums, talking up and comparing the minute details of this or that. There's obviously the same risk of cargo culting that there's always been, but there's genuinely useful information available from people who spend way more time than the common man on their area of interest.

Grepping my dpkg list for packages which start with \s0. is fun.

aspell (currently 0.60.8) has a release note for 0.11 on Sept 12, 1998

gettext began in 1994 and is currently v0.21-11

tftp is 0.17-25 and dates back to 1981

telnet is 0.17, and its origins date back to 1969!

My understanding is that for triglycerides, lower is basically better, outside of "extremely low as an indicator for disease". For example, https://academic.oup.com/clinchem/article/60/5/737/5621637 . Can you point me at reading on the lower bounds for them?

(I ask because my labs routinely show very low triglycerides - generally in the 65-70mg/dL range - which I've been told is a very good sign for heart health.)

I think you're reading something into it that isn't there. It's just not on the shelf, I'm not likely to notice it and incidentally purchase it when I'm buying ibuprofen or whatever, and I have to go out of my way to get it. I can't just have Amazon overnight some to me. As a result I've just gotten in the habit of buying it periodically without waiting for "I'm sick" to be the reason for the purchase. The 6 month period isn't "trying to not get noticed", it's "that period comfortably covers my household's needs with some breathing room".

Sure, I could buy 6 boxes at a time. I just don't. There's not really any good reason to min-max my pseuodephedrine acquisition.

Yup, it's discussed extensively, predominantly as "val/val" (GG) and "met/met" (AA).

The point is more to not be in a position to have to go out and buy some when I'm feeling cruddy. Since it's controlled, I can't just have some shipped to my door when I'm laying in bed sick. I just got in the habit of picking it up periodically to ensure that it's on hand when I need it.

"Suggestible You" by Erik Vance explores this phenomenon, and really turned me around on placebos. I'm, generally speaking, not particularly susceptible to placebo effects, and always kind of (arrogantly) chalked placebo effects up to people just being gullible and wanting to see something that wasn't there. It turns out there are actually genes that might be responsible for the physical impact of the placebo effect in the brain.

There's an enzyme - catechol-O-methyltransferase - which is coded for by the COMT gene. This enzyme catalyzes the metabolism of dopamine in the brain. COMT has three common variations - AA/AG/GG - which substantially alters how effective the enzyme is at metabolizing dopamine. AA results in significantly reduced enzyme activity, which can result in dopamine built-up in the brain, which results in increased sensitivity to stress, anxiety, and pain, but comes with the bonus of enhanced cognition, motor skills, and memory.

AA genotypes also tend to be "placebo responders", while GG genotypes tend to be non-responders. This tends to imply that placebo responsiveness isn't purely psychological, but physical - and indeed, it turns out that in AA "responders" you can turn off the placebo effect by administrating naloxone (which works by binding to opioid receptors)!

This has really interesting implications for pharmacological research, too - if there are people who are genotypically predisposed towards or against placebo effects, then a drug trial that stacks responders in the trial group and non-responders in the placebo group which would produce a drug efficacy signal that could be just the placebo effect.

Indeed. I've known that phenylepherine is worthless for a long while, and just make it a point to buy a pack of the real stuff every 6 months or so, so that I always have it on hand when I need it. In my state, it's controlled - you have to get it from a pharmacist (though you can do so without a prescription) - they take your driver's license and don't allow you to purchase it too often.