pixx

I don't know who I am, and neither do you.

Public Key

npub1yk5jnr8fa4s87afelvnaj3qczc28deegfdfr2a7w96rclncc2t7s7d2jl0

NIP-05 Address

Profile Code

nprofile1qqszt2ff3n576crlw5ulkf7egsvpv9rkuu5yk534wl8zapu0euv99lgpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsxdw3fu

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1yk5jnr8fa4s87afelvnaj3qczc28deegfdfr2a7w96rclncc2t7s7d2jl0

Show more details

Last Notes

2026-08-18 19:18:28 GMT

- reply

@nprofile…8udl
Yeah, for sodium and _everything else_ that's restricted, my general conclusion from looking into it more deeply has been "or just stop sitting on your ass"
turns out that insulin resistance also triggers excess sodium retention
Increasing activity levels directly reverses insulin resistance => improves the body's tolerance of sodium
Same for sugars: if you down 50g of sugar and then go on a mild walk, even, the muscles bypass insulin resistance if present and _directly pull the glucose from the bloodstream_
which itself helps prevent the downstream effects that cause insulin resistance, making it possible to reverse it _while still consuming sugar_
Which explains why there's a lot of people I know who are hyperactive (>5kcal/day, easily) who eat a ton of junk food and are totally and utterly fine.
Sodium and sugar aren't unhealthy; sodium and sugar in the context of sedentary behavior are unhealthy. I've seen some studies suggesting the same is true for saturated fats - they're always worse than e.g. olive oil, but the _degree_ to which this is true is heavily affected by activity levels.
Basically: with healthy levels of sleep and exercise, the body can tolerate basically anything you throw at it a hell of a lot better, even if the diet remains bad.
Improved sleep and activity are _always_ top priorities for health. Diet is secondary.
(And, as far as diet goes, _increasing_ intake of _healthy_ food is _waaaaay_ more important than eliminating intake of unhealthy food, and will naturally result in that long term as an inevitable byproduct too.)
...again, to emphasize, all of this is true _in general_. exceptions exist. but for 90%+ of people...

2026-08-18 18:50:52 GMT

- reply

@nprofile…8udl Since I'm studying this from a perspective of "improve my own health", getting to the point of knowing how the ratio of sodium:potassium ratio affectrs my health, and empirically testing it by changing it and seeing what happens, is sufficient. "Cool, that solved the chronic health problem."
But it turns out that you can be _totally wrong_ about what it actually _does_ on the inside, and make some reasonable but totally false assumptions, and just... not care. Worst case, you do what I do, and talk about it publicly, and - well, if the functional model is correct, but the mechanical model isn't, the worst that can be said is "you don't know what you're talking about"
but that's why I always make sure to emphasize that, yeah, I know enough to get by, I know enough to make healthy decisions, and beyond that it's a hobby. I don't mind making good decisions and misunderstanding why they're good 😅

2026-08-18 18:23:31 GMT

- reply

@nprofile…8udl
...looked into it after you mentioned it and had to refine my model here too, so thanks :P
I'd been drastically oversimplifying why potassium affects blood pressure and was basically completely wrong about how blood pressure works at all. Once again, "good enough to get by" (i.e. understanding the effects correctly if not the mechanism) is not the same as "knowing what you're talking about" 😂

2026-08-18 06:41:04 GMT

- reply

@nprofile…8udl The tl;dr, oversimplifying and also only having studied this exactly enough to make use of it: blood plasma has a specific osmolarity. Osmolarity only cares about _amount of particles per volume_, not the mass of those particles.
So for instance, sodium chloride (table salt) splits into two particles: Na, and Cl. Trisodium citrate, meanwhile, splits into 3 sodium and one citrate. Which means that you go from 50% sodium to 75% sodium at the same particle count.
A _hypotonic_ solution (a fluid with fewer particles in it than bloodstream), in essence, the water gets pulled out of the gut and into the blood. A hypertonic one (like basically all sodas) literally pulls water _out of the blood stream_ in order to dilute it. Isotonic leaves it alone.
Then there's also the fun effects like e.g. sodium holding water in the bloodstream, potassium holding it within cells, and therefore the _relative amounts_ heavily dictates blood pressure. And muscles hold glycogen for fuel, which binds a significant amount of water too, which is why it's incredibly easy to lose like 5lbs of weight quickly, because you burn the glycogen, the muscles dump the water, your body is already hydrated => it gets flushed right out. But then the instant you eat anything, you need to pull in water to the muscles, so unless you drink water...
The body is really complicated, and I don't pretend to have done more than scratch the surface myself, but I've learned enough that my energy problems are largely gone, my fat loss is up to ~2lbs/week right now (the regression I ran on my data says 2.92 but... due to other nutritional biology stuff I've been looking into, I'm fairly sure it's wrong, and that my scale is in part misreading water retention and exaggerating the effect; I have reason to believe I have a few pounds more blood right now than i did last week 😅 and it took me _way_ too long to figure out what was going on there, but oxygen processing data from my runs fits the hypothesis too), and so while I'd hardly consider myself an expert or a doctor or a nutritionist, I think I've progressed beyond "hobby".
The osmotic stuff in particular isn't actually super hard. Can learn most of it in an hour or two!
and, IME, learning a buncha small things like that is actually sufficient to get a pretty good appreciation for how everything fits together, but even more importantly, an appreciation for how complicated it is
Been doing this for years and there's very little I have real certainty about. There's a lot where I'm convinced, due to making changes based on it _improving my life_, that I have a _good enough_ understanding to get by, but that's _very_ different; addressing relatively mild chronic conditions (e.g. obesity), the worst that can really happen is that you make the problem worse, realize something didn't work, and try again.
A bit of caution, and rigorous tracking, IMO makes it actually really easy to conduct nutrititional experiments on yourself that would be unethical to do on others 😂
Anywho, </rant>

2026-08-18 01:28:41 GMT

- reply

@nprofile…8udl basically the same, except I'm creating the blood pressure problems by upping sodium intake due to massive sweat losses, and getting enough potassium to make that not a problem, while losing weight and heavily active, is Really Hard

2026-08-18 01:08:50 GMT

- reply

@nprofile…8udl that's technically what I'm gonna do with it too :p
Except for the small amounts thing 😅
(Heavy labor in high heat without air conditioning :/)

2026-08-18 00:52:13 GMT

Whelp, my chemistry experiments have resulted in ordering, uh, a full kilo of potassium citrate
for, reasons

2026-08-17 23:14:12 GMT

- reply

@nprofile…8udl @nprofile…uakw @nprofile…fhea
I don't find it depressing that they leave, I find it depressing that we stay 😂

2026-08-17 22:10:35 GMT

- reply

@nprofile…8udl @nprofile…uakw @nprofile…fhea
> ultimately it's my job to filter my feeds as meets my needs
This is exactly why I don't recommend fedi to most people, tbh. Almost nobody _wants_ that.

2026-08-05 04:57:25 GMT

- reply

@nprofile…5ds9
I'm more in agreement with b) than with a)
I'm fully on board with encouraging it and having a dedicated cargo space SO LONG AS I have to pay for it properly.
It's one thing I like about amtrak's policy: if you have a normal bike, that's a fee. if you have a folding bike that fits in the same volume as luggage: you can use it as a carry-on, for free, _instead of luggage_.
It means that you're encouraging usage of portable bikes on mass transit over bulky ones, giving a reasonable option, and requiring that said reasonable option be paid for by the people utilizing it, not everyone else
I find this quite reasonable overall

2026-06-03 03:07:27 GMT

Some people insist that they could not transport a grocery run home without a car. These people lack imagination.
😂
No the fan is not a propeller...
Yet.
https://assets.merveilles.town/media_attachments/files/116/683/945/758/532/586/original/fa5fd1a16fa1f742.jpg

2026-05-22 06:32:30 GMT

Interesting note in https://github.com/rust-lang/rust-forge/pull/1040:
> Can you think of a concrete improvement to the policy that addresses your concern? Consider whether your change will make it harder to come to a consensus
No. The idea that consensus matters is one of the most toxic ideas in modern society.
Some things are wrong. If everyone else in the world told me that I should try eating babies, they taste delicious, I would choose to die alone rather than be part of a twisted society.
Morality is not subjective, the idea that the opinions of others MATTERS at ALL to morality is _perverted_, the idea that consensus matters more than doing the _right thing_ needs to die and, more importantly, needs to be _killed_.
I have zero tolerance for this farce.
(The context for this: they're adding an LLM policy, and have these rules on the PR for the policy:
No comment on this PR may mention the following topics:
Long-term social or economic impact of LLMs
The environmental impact of LLMs
Anything to do with the copyright status of LLM output
Moral judgements about people who use LLMs
We have asked the moderation team to help us enforce these rules.
)

2026-04-28 17:17:27 GMT

- reply

@nprofile…8m52 Because they don't need CPU power, they only need GPU power

2026-03-23 12:21:30 GMT

Working on an old abandoned code base where i can't talk to the authors and it takes hours to figure out some of the bullshit it's pulling off, i can definitely see the appeal for a tool that can quickly read over the code and check my inferences on e.g. memory flow and ownership patterns
Nowadays, people would obviously default to either LLMs for that or for arguing that humans should do it because the other people are saying LLMs and fuck that
But what I'm more interested in: there's plenty of static analysis tools for checking code you wrote. Is there any prior art in deterministic static analysis tools that help _read code_ by other people?
Dataflow and control flow analysis are super commonly used in compiler optimization. Are there any tools out there that do dataflow analysis and then, say, can look at a variable and show all _possible references_ and relevant code?
E.g. this Foo* is passed to foo(), so it can be f there, which might call bar, and ...)
Or "this code allocates Bar, here's the entire possible relevant code for what might happen to that across this 100kloc project, here's every function in every file that might use _that specific value_, and an explanation of how it gets there"

2026-03-05 00:10:45 GMT

- reply

@nprofile…8m52 okay but is that even copyrightable??

2026-01-13 05:57:25 GMT

- reply

@nprofile…6e53 🤔
Why is that impossible for PEG? The first-alternate-wins rule?

2026-01-13 05:56:19 GMT

- reply

@nprofile…6e53 ...are they calling that sign racist? TBH I strongly disagree
"As long as you understand our language, you're welcome here" isn't racism, it's self-respect

2025-12-30 20:57:39 GMT

Yes, three optical drives is absolutely necessary :D
https://assets.merveilles.town/media_attachments/files/115/810/517/581/534/055/original/6f2c2733d7b120d3.jpg

2025-12-09 21:57:15 GMT

- reply

@nprofile…czr6 @nprofile…l4cp @nprofile…tpsu @nprofile…5hdc
It's also entirely plausible that this is actually being interpreted backwards: the _actual job_ is safe enough that _if they die_ while employed, it will necessarily be through human action?
What does 19% mean? 19% of what? 19% of 100 deaths, or of 10,000?

2025-11-18 19:01:35 GMT

- reply

@nprofile…xneg > all of these pointing at the public sidewalks, all completely illegal.
Under which country's laws?

2025-11-13 22:29:55 GMT

- reply

@nprofile…3gt0 @nprofile…2rzg
(as joking about rust being -bin in gentoo
yes it can also be from source, but :)

2025-11-13 22:27:47 GMT

- reply

@nprofile…3gt0 @nprofile…2rzg like rust? ;P

2025-11-12 22:15:32 GMT

- reply

What would be more interesting?
- Aftermarket battery, otherwise identical to original, but not crap like most aftermarket batteries are, and with longer lifespan [LiFePO4 / Sodium-ion cells]
- Aftermarket battery, specific to the laptop model, but with user-replaceable cells and manual [supporting all three of sodium-ion, LiFePO4, and lithium ion based on user preference]
- A set of batteries for different machines that can share the same pack of cells and the same controller, but with adapters for form factor and programmable expected voltage?
Note: I'm very much not committing to anything at this stage. I'm in the process of prototyping a _single battery_ for my _own usage_, and then went, "Wait, I can't be the only one annoyed by all these, we can do better..."

2025-11-12 20:09:16 GMT

Hypothetically: if there was a service offering after-market open-hardware sodium-ion batteries for old laptops at a reasonable price, would you be interested in buying them and, if so, how many? I'm talking about old Thinkpads and such. Information about the specific devices would also be useful.
Edit to add: LiFePO4 would almost certainly also be available; there's negligible additional work to support both. Li-Ion is less likely but possible if there's demand.
Edit: I wouldn't normally ask, but since the entire point of this interest is to gauge interest, boosts are definitely welcomed and requested :)

2025-11-11 23:16:13 GMT

- reply

@nprofile…gkm9 @nprofile…3gt0 @nprofile…4xf4
Just don't read them /shrug

2025-11-11 23:07:02 GMT

- reply

@nprofile…gkm9
Or don't even link them.
If they want info they can pay for that too. There's no reason to act with good faith towards someone who's taking advantage of you
@nprofile…3gt0 @nprofile…4xf4

2025-07-11 20:05:41 GMT

- reply

@nprofile…vc7v That's 9front's built-in bar(1) :)

2025-07-11 06:55:52 GMT

MNT's reform2.0 schematic rendering in #pdffs on #plan9 ;)
I've _almost_ found all the broken parts. Probably.
and yes I know it's not a perfect rendering
https://assets.merveilles.town/media_attachments/files/114/833/286/929/157/630/original/b39a0524f20d638b.png

2025-06-27 21:57:02 GMT

- reply

why is nutrition so hard?

2025-06-27 21:56:11 GMT

Today I learned about the Satiety Index, and immeditely noticed tht whole grain pasta shows up as 1.57x as satiating as the "refined" stuff.
That... I think that matches my exprience, but I haven't actually put numbers on it. hmmm

2025-06-23 19:02:26 GMT

Days since a non-technical person asked me to "take Windows off the computer so I can actually use it and open docs without it taking 2 minutes and trying to get me to play Candy Crush and showing me a dozen things I don't care about, I don't want 'apps' on my laptop": 0
Days since I found problems with Linux desktop that make it annoying for nontechnical people: also zero.
Linux (in this case, Mint) is frankly not good enough for normal people.
But Windows isn't either anymore, apparently.

2025-05-12 17:07:31 GMT

- reply

@nprofile…v55x @nprofile…kud6 @nprofile…e3an
In the corporate headquarters where professional programmers engage in cultural appropriation, mainly.

2025-05-06 22:59:30 GMT

Wait, does sugar have a similar flavor effect to salt???
Was about to add sugar to something because it was a bit bland when I realized it was just lacking salt, it wasn't really sweetness i was after, the underlying flavor itself was just kinda muted, and salt - naturally - brought it out
does sugar do similar??
*stares at the cooking chemistry book on the shelf*
Tonight, love. Tonight.
(On a related note, I dislike low-sodium foods for a similar reason: too much sodium is bad, sure. but taking out sodium makes food just objectively _less good_, and that's usually solved by adding in something _much worse_, oftentimes extra sugar.
tbh the attitude in general of "this thing is bad, let's get rid of it!!" is something I"m increasingly convinced is dangerous - what purpose was it serving, previously? When we get rid of it, what will we do instead to meet those goals, and most importantly, what harms will its replacement have?)

2025-01-23 01:18:51 GMT

@nprofile…7j56 I think this is my favorite :D
https://assets.merveilles.town/media_attachments/files/113/875/034/331/254/709/original/a389a85c67557424.png

2024-12-30 03:32:17 GMT

love when a site has three options for gender:
"Male"
"Female"
"Unknown"
NONE MAY KNOW THE TRUTH OF MINE GENDER, OR THEY WOULD COWER IN TERROR

2024-12-13 20:52:18 GMT

so jelly of the people who've got their gay iPods already 😭
#tangara

2024-12-13 20:35:52 GMT

- reply

@nprofile…7j56 "because cinap didn't want to" 😂
THE PEOPLE DEMAND CONTEXT :D

2024-12-12 03:45:54 GMT

- reply

@nprofile…g2yn SVG is probably a reasonable start, at least, yeahhhhh

2024-12-12 03:24:57 GMT

- reply

@nprofile…g2yn just in general. I might take a look, then.
(Looking to implement a vector engine, basically, and I want to understand graphics more before trying that :P)

2024-12-12 03:05:52 GMT

Any recommendations for a quick crash course in vector graphics? Video or text are both good :)

2024-12-02 11:54:58 GMT

- reply

@nprofile…al3s yeaaaaah realizing that the sensors were _part of the battery_, not the laptop, is what made me go from "sensors are useful but inaccurate" to "meh, what's even the point" on a lot of machines.
With third party batteries, it's just noise :(

2024-12-01 00:16:10 GMT

What's a hoodie but, like, for the legs?

2024-11-20 20:00:18 GMT

- reply

@nprofile…7j56 @nprofile…5hlp @nprofile…wnuz @nprofile…xpq6 Clan9 :D

2024-09-29 03:28:26 GMT

/me sighs and opens the FTC fraud reporting page again, to report Frontier, again.
Come _on_ frontier _how many forms of fraud_ are you going to commit??

2024-09-28 05:11:02 GMT

Shoutout to Valve: when I saw "update to user agreement" and "all disputes must be resolved" in one page I was prepared to be upset
"All disputes must be resolved through the court system and not through arbitrarion" is not something I've ever seen in a company's ToS before. Sincere kudos, I'm honestly impressed.

2024-08-27 03:45:11 GMT

> Cryptanalysts who restrict their perspective to cryptanalysis might see as alarming an attack
with complexity 290 and success rate 1/10, however if we were to quantify the probability of
the existence of a remotely exploitable vulnerability in the device you use to read this PDF, it
would be close to 1. Such an exploit would let the attack read all your PGP and SSH keys.
... lol.

2024-08-19 17:30:17 GMT

- reply

@npub1qu5…4evr Any channels you'd recommend checking out? :)

2024-08-13 15:57:38 GMT

- reply

@npub19hc…9pfz XT?

2024-08-06 14:32:12 GMT

- reply

@npub1w9a…04m0 Why?
Why create artificial scarcity?

2024-06-04 06:14:05 GMT

well shit, i was wrong about EVs O_o

2024-06-03 02:19:54 GMT

- reply

@npub1l3g…vu48 It can be used to buy alcohol now!!!

2024-04-26 02:56:27 GMT

jeez, average U.S. home apparently runs on the equivalent of a constant 1100W[1]?? :O
> 43% of all residential electricity usage is to provide heat or cooling either to rooms, food, or water.
Oh. Right. Most (~90% IIRC) houses in the U.S. have absolutely shitty insulation, and e.g. electric water heaters average something like 50W of constant power usage _when the tank is fully heated_ because of also shitty insulation, and jeez no wonder heat pumps are a big deal.
[1] https://www.eia.gov/energyexplained/use-of-energy/electricity-use-in-homes.php

2024-04-19 02:12:46 GMT

- reply

@npub1l3g…vu48 COme to think of it, now I'm contrasting it with nationalization.
I think a co-op might genuinely be better??

2024-03-25 19:47:56 GMT

- reply

@npub1l3g…vu48 Do you think I tossed a blended microSD card with plan9 on it in there for NOTHING?!

2024-03-25 19:47:05 GMT

- reply

@npub1l3g…vu48 Does... my cheesecake... run plan9???

2024-03-25 19:35:43 GMT

:D Baking again
https://assets.merveilles.town/media_attachments/files/112/158/004/440/829/390/original/baf4c747d57c4a79.png

2024-03-19 06:13:06 GMT

- reply

@npub1l3g…vu48 Funny you should say that.
OpenAI is for all intents and purposes a subsidiary of Microsoft :D

2024-03-19 06:09:04 GMT

- reply

@npub1l3g…vu48 Not at all, if you ask me - there's alternative approaches that work better, but OpenAI happens to _have_ tons of compute, storage, and RAM, and so their methodology is staying relatively simple :P
There are LLMs that can run on garbage hardware. Someone ported llama to run on a Thinkpad under Plan 9 lmao.
I suspect that OpenAI is going to get out-competed in the next few years unless their strategy to get the government to grant them a monopoly pans out.

2024-03-19 06:00:12 GMT

- reply

@npub1l3g…vu48 That's not the chatbot; that's the thing that takes your message and splits it into pieces for the chatbot to reason about.
"How would you destructize a light bulb?" <- you can't rely on a list of real words because people play word games all the time. But, "destruct" "ize" as two tokens is a pretty clear "actively destroy" to an English speaker; you want the chatbot to be able to reason about that.
Tokenization is a good idea. Brute-forcing it based on The Internet is where I laugh.

2024-03-19 05:56:25 GMT

- reply

@npub1l3g…vu48 To oversimplify, they take a giant corpus of English text, scan through it to find common sequences of _characters_ (e.g. "i" is often followed by "z"), and known sequences ("iz" is often followed by "ize"); the result is a function that, given "funize", will see the sequence "fun" and the sequence "ize".
Which... makes me retract what I said about high school understanding of English. Do they even know what word roots are??

2024-03-19 05:54:08 GMT

- reply

@npub1l3g…vu48 I've been reading up on how they work a lot, lately. I was extremely dismissive of them for a while, but I'm realizing that they're a _lot_ smarter than I gave them credit for - while still being far stupider than their proponents claim.
The way that GPT-2 turns english text into tokens is... uh. As a softwarer engineer, reading up on it, it feels like what a software engineer with a high-school understanding of English and a billion dollars of compute resources might do

2024-03-19 05:50:44 GMT

- reply

@npub1l3g…vu48 There's no tailoring at all! There's fanciness on top, but the core of GPT is a function which, given text, predicts what should come next. It has some interestingly intelligent bits in the middle in order to be capable of this, but it helps IMO to think of it as an extrapolation engine being used to define a chatbot.
The "chatbot" is a program on top. It basically synthesizes text based on your messages and its own prior ones, and invokes the extrapolation engine to decide what's next

2024-03-19 05:47:27 GMT

- reply

@npub1l3g…vu48 LLMs have some interesting internal structures, but at the end of the day, those structures are used to follow the premises that you give them.
They "fall" for "trick" questions easily, because they don't distinguish between true and false, merely consistent and inconsistent.
It gave you a chat log that was consistent with what you told it. Change what you tell it, and you change what is consistent, and thus its output.

2024-03-19 05:44:52 GMT

- reply

@npub1l3g…vu48 Thing is, if you led in with opposing premises, you could get it to say that there is no dissonance because it's being used fully ethically :P