The going wisdom seems to be that the EU's Galileo is the most accurate system for civilian use. GPS has undergone frequent systematic update for almost a half century.
Indeed. i have some GPS receiver modules and had wondered about this data, I had assumed it was imprecision in my device or something to do with a satellite moving around. I'll have to plug it in and go back for another look.
Two things amaze me about GPS. First, that there are still four Block IIR and seven Block IIR-M satellites operational; these had 7.5 year design lives and were launched by 2004 and 2009 respectively. Second, that L1C, L5 and L2C are all still pre-operational thanks to the OCX debacle. L1C and L2C really modernize the signal structure to improve accuracy.
fortunately that video is more gentle but the math in that youtube channel absolutely melts my brain some days, I can keep up for the first minute but then all bets are off as he dives in and I realize there are some insanely brilliant people out there
They were definitely used on Mir- in 1997 one caught fire, blocking the crew's access to their escape Soyuz, though they put it out.
It looks like NASA helped redesign it to be safer, creating the modern Solid Fuel Oxygen Generator (SFOG) system still in use on the ISS as the backup.
Candles are useful when oxygen has been consumed because of respiration or a fire. They're not useful in a leak.
Conservation of mass: if a cubic meter of air escapes, that's 1.25 kg, and you need at least that much in candles. (You actually need 2 kg because the candle isn't solid oxygen)
There's ultimately 1.2 t of atmosphere on the ISS. This will also result in a pure oxygen atmosphere, which is dangerous. You need nitrogen.
I think you're double counting; you need 1.25 kg of oxygen and nitrogen combined to replace 1 cubic metre of air.
1.2t of candles doesn't seem like an unreasonable amount of extra payload if they would really be valuable in an emergency. The ISS weighs 400 tons and a napkin estimate says it has had 1000 tons of resupply missions. The candles have a shelf life of 10+ years.
You need 1.25 kg of gas. Candles don't consist solely of oxygen, the ones the Russians use utilize lithium perchlorate (LiClO4). When they finish burning, you are left with a lithium chloride ash (LiCl), which will be 40% mass of what you started with.
It works out to be more efficient, at least in terms of mass, to send up large tanks of compressed gas instead.
NASA used pure O2 in space until the end of the Apollo program, but the Shuttle and later used the same air we breathe today, 1atm 80% N2/20% O2. Note that in space, the pure O2 was at 0.4atm, so roughly twice the oxygen partial pressure, but only slightly more dangerous than the air we are breathing now. (You need about 0.4atm to keep your lungs from collapsing, so that's the lower limit.)
Why the difference? It's a question of what risks you were most afraid of. Even today, every single spacewalk is done at 0.4atm pure O2- trying to do a spacewalk at 100kPa even the strongest man in the world would have trouble bending his arms- so before a spacewalk the astronauts need to spend several hours pre-breathing pure O2 to get all the nitrogen out of their bloodstream before they can do a spacewalk. The Apollo program thought it was safer if the astronauts could do a spacewalk at literally any point in the mission, so that's what the spacecraft was designed around.
On the other hand, for long duration spaceflight, introducing a different pressure and atmosphere is just another potential source of health problems. Even today, the largest source of information on how human bodies last under 0.4atm pure O2 is the three Skylab missions from 1973-1974. And so the Soviets- who were always more interested in space stations than the moon- and NASA during the Shuttle era went with the atmosphere that seemed like it offered less health risks for people staying on a space station.
Okay, so what about the Apollo 1 fire? To speed up testing, Apollo 1 did two tests at the same time: the Plugs-Out Test, where the astronauts were in the spacecraft with everything running and practicing their countdown, and the Overpressure test where they pressurized the spacecraft to 1.4 atm (to mimic the pressure differential in outer space). And they did it with pure O2. So you had all of these electronics running in an environment at 1.4atm pure O2. And that was incredibly dangerous, in a way that actual spaceflight, a mere 0.4atm O2, was not. But it was just a test, another in a long string of them, and no one involved ever really analyzed it as a potential hazard.
After Apollo 1 a few things were changed: one was that they did the Plugs Out test and the Overpressure test at different times, and a lot of stuff was turned off for the Overpressure test. Another was that the Apollo capsule at takeoff was 1atm 80/20 until a couple of minutes into flight, when it dumped the cabin atmosphere overboard and replaced it with pure O2 at 0.4atm. That's why the astronauts carried little packs in their arms in all the pictures of them getting into the spacecraft, that's the pure O2 tank that they were breathing off of until they could switch to the atmosphere in the cabin after it was replaced.
They breathe a normal mixture of O2 and nitrogen at 1 atmosphere of pressure. A pure oxygen environment is horrifically dangerous if fire ever breaks out.
And the 1961 fire that killed cosmonaut Valentin Bondarenko in an altitude chamber. The Soviets covered it up until the 1980's, so NASA made the same mistake.
I still don't understand how "AI" is ready for serious use beyond entertainment purposes
Every time I ask ChatGPT to make a table for a subject I know well, I will find an error in one of the results and it is very confident about it until I question it in detail
Every time I ask ChatGPT for nutritional breakdown of some dense food source and give it a quantity like 8 ounces and ask for the weight of each ingredient, the weights will be wrong and add up to more than the original weight of 8 ounces
These are variations of the old "how many Rs in strawberry" problem, it's still not solved, "AI" cannot reassemble a complex problem properly
A lot of what it tells me in detail about some subjects sounds suspiciously like Reddit posts reassembled out of order
Two things that I would recommend trying out if you're interested in exploring this further:
1. If you're not paying for a model, the results will be worse. That sucks but the free access models are just not very good for anything where you need to trust the output, even for basic queries.
2. More important than #1 is access to tool use. If the LLM is just producing a nutritional breakdown from its weights, it's almost always going to be wrong. If the LLM is allowed to break the problem down into deterministic steps, it will do a lot better. In the nutritional breakdown case, an LLM with search + tool access can pretty easily break the problem down:
- Searching the web for a recipe or ingredient breakdown for the food
- Searching the web for nutritional qualities of each ingredient per some volume of the ingredient
- Writing and running a script with e.g. Python that takes in the recipe's projected serving output, the desired serving size, the amount of each ingredient etc, and scales the ingredients to match the desired serving size, and sums the nutritional qualities of the scaled ingredients.
I've tried this specific case with Claude + Gemini for my own purposes and they both handle it very well. The challenge currently is that the models will not always arrive at this approach when provided with an ambiguous prompt; sometimes they will, but sometimes they'll just vomit up a fully autocompleted response from their weights. Being more specific in the prompt or defining a skill that details the intended approach lets you get more useful + deterministic results while still taking advantage of the fuzzy glue that LLMs can provide here between steps.
Same with the classic strawberry r-counting case. IIUC LLMs have trouble with this because of how training data is tokenized, but any LLM will have no trouble farming out to e.g.
There are basically two kinds of applications. One is where you want to correctly solve the problem at least 99 out of 100 times. LLMs generally don't (and not everybody realizes that) so there are a lot of debates and research around how useful and reliable they are or how to make them so.
The other kind of application is where you can try 100 times and you only need to be right once. Solving a mathematical research problem is like that.
https://fortune.com/2026/06/04/peptides-crypto-looksmaxxing-...
reply