The first working transistor was centimeter-scale, now billions of them fit in that space.
The first useful internal combustion engines were room-sized, now they fit on a moped.
The truck-sized hole in your argument is talking about "the smallest test". First discoveries/demonstrations of interesting phenomenons don't typically happen at the smallest scale (why would they?).
The first working transistors and engines were of the size which happened to be most convenient to work with. They could then be shrunk because fundamental physical limits to their size were far below human scale. Their inventors were neither constrained by nor interested in those fundamental physical limits. They were inventors, not scientists.
In contrast, a particle accelerator like the LHC is designed from the outset to explore physics at a given energy scale at the lowest possible cost. Shrink it any further and it will no longer work. Despite decades of attempts to come up with alternative designs, when time comes to draw up plans for a successor capable of pushing to even higher energy, it's just more of the same:
100 km/h is slow compared to passenger train (even non-high-speed ones). Depending on how packed the schedule is, it might not be possible to analyse track during the day without causing backups.
A million alternatives is peanuts. Restricting the search space to text files with 37 possible symbols (letters, numbers, space), a million different files can be generated with just 4 symbols.
A trillion is 8 symbols. You still haven't reached the end of your first import statement.
I just took a random source file on my computer. It has about 8000 characters. The number of possible files with 8000 characters has 12500 digits.
At this point, restricting the search space to syntactically valid programs (how do you even randomly generate that?) won't make a difference.
I double checked because these were not the prices I had in mind.
After looking through the options, I think that's because the designs I did quotations for had 0.2mm holes. This is standard for Eurocircuits, but high precision for JLCPCB.
Note that to get the price you quoted, you'll get lead in your PCB, and vias that are not plated, but plugged with conductive epoxy. Changing that gets you to $14 for 5 boards, which is still way cheaper than Eurocircuits.
I'll keep that in mind for the next PCB I design: keep holes bigger than 0.3mm if possible.
My laptop has one of these ethernet ports that half close when not in use. It doesn't work anymore because someone mistook it for the USB port that's right next to it when distractingly plugging their keyboard in.
I agree with the sentiment, but what stops the government designating annoying people or companies as russian (also see "foreign agent" and "matter of national security")?
Also, am I on the hook because my wife's great-grandfather was russian?
Because my idea is that it wouldn't be the government designating them, but judges. Just like a judge already today has the power to lock you up for life, but only after a lenghty rule of law process.
But the trouble is judges are human, and if the boss of your boss who could shut down your court and turf you out is loudly proclaiming "will nobody rid me of this troublesome Russian"... what lawful ruling will the judge find to appease their master?
The other thing that sprang to mind was the US interning their own citizens in camps. Would you be in favour of internment camps across Europe for all citizens with any Russian heritage, preemptively locking them up in case they might provide succor to The Enemy?
You are making an argument against any kind of rule of law in general. If somebody powerful can say to a judge "will you please sentence this competitor of mine to life in prison" and he complies, that's the same problem. But judges in Western countries are very rarely that corrupt, and thus rule of law functions.
There has been some research (IIRC by ESA) for using the upper atmosphere to feed a ion engine. That way you should be able to put satellites even lower as long as they have enough power from solar panels and are functional.
The french government has been heavily subsidizing private R&D (up to 50% of the cost, including engineer salaries). It was relatively easy to create a moonshot project worth a few millions, and have the taxpayer pay for half of it. Then you just need to find a sucker to pay for the other half, and collect the money (getting an actual result is optional).
How do I know? My company is a minority partner in one such project (wind energy, we would provide instrumentation). It's infuriating, the head company has been trying to make one of the big energy providers pay for half the R&D, with no success, and the project will be closed. Lots of taxpayer money wasted for no result, and we won't make sales.
Because of these abuses, the french government is changing the financing rules. They will only finance small proof of concepts first, then a pilot project, and only then industrialisation issues (instead of financing all in one go).
US grid carbon intensity is 0.384 gCO2/kWh (source: ourworldindata). US datacenter energy use in 2023: 176 TWh (excluding crypto, source US congress). How much of that is AI, I couldn't find.
So that's 67Mt CO2, I hope I haven't misplaced my decimal point, please double check. That would be 1.3% of the 5Gt of CO2 the US emits per year.
For global emission and future trends the IEA estimates about 500TWh/year globally today, and 1000TWh/year in 2030 (base scenario). Assuming these use the current US grid carbon intensity, that would be about 200MtCO2 today, 400 in 2030. Global CO2 emissions today are 40Gt/year, so that would be 0.5% today, and 1% in 2030 (if global emissions stay stable).
Thanks, that’s interesting. IEA definitely seems like a solid source for this kind of thing.
1% (if that’s accurate) isn’t nothing, but it’s also nowhere near what seems to be implied by the level of people’s reaction to AI buildout and the framing as an environmental catastrophe. (Of course there are other factors, such as local pollution from gas turbines.)
Interesting comparisons are blast furnaces (6% of global emissions) and aviation (2.5%). Both arguably more economically necessary than AI, for sure, but if we could make either of those meaningfully less of a contributor to climate change we’d have covered the whole AI buildout. And that’s not even getting into the possibility of a transition to solar energy for running datacenters, which China is already deep into and in which the US is far behind.
The ocean has already absorbed 30% of the CO2 humanity has emitted. It causes issues: ocean acidity rises, which reduces plankton ability to grow. Plankton being the base of the ocean food chain, all ocean life gets impacted.
You'd need to find a way to sequester carbon without it leaching in the water.
Random idea: What if we just sequestered it into elemental carbon pellets and let it sink to the bottom of the ocean? It should not react with the water.
One idea is to charcoal the wood, it makes it harder to decompose and is similar to pure carbon. I'm not sure if it's better to send it to the bottom of the ocean or just to a big hole on land.
The first useful internal combustion engines were room-sized, now they fit on a moped.
The truck-sized hole in your argument is talking about "the smallest test". First discoveries/demonstrations of interesting phenomenons don't typically happen at the smallest scale (why would they?).
reply