The sensor angle would be in a file like `/sys/bus/iio/devices/iio:device*/in_angl0_raw` (device number can vary). At least I have this in a config file and remember it working (maybe on a different computer?). I cannot get it to work anymore on my laptop.
The part about low-pass filtering using FFT ends with the conclusion that it is fundamentally a different thing from a gaussian blur ("Low Pass Filter ≠ Low Pass Filter"). However: A convolution in image space is just a multiplication in fourier space. If you multiplied with a gaussian in fourier space you should get a gaussian blur. It's just not necessarily the fastest option.
That's a fake Wikipedia screenshot! That line doesn't exist in the actual article and didn't at the time when that tweet was written, and does not even fit in the context. To me, this is at best an example for how much higher the quality is on Wikipedia than in average social media like X.
Amazing. And it's listed as having 23.7K views. I suppose far less than 1% of those people would have taken the time to see if that was real or fake.
EDIT That line did exist in the past. It was there one year ago. Can someone more skilled in Wikipedia find and link the revision where it was removed? Bonus points for finding when it was added. Thanks in advance.
The edit did exist. So the phrase "Multiple studies have found a left-wing bias at Wikipedia" can be found last in the edit (4 February 2025) [1] and was removed by the editor Aquillion with the explanation [2]
Not seeing anything in this source that supports this language; they summarize no other sources that I could see, and their own conclusions are more complex than this (as covered further down the article.)
I think that the one who previously introduced the phrase should have either not stated 'Multiple studies' or provided information about other studies. I suspect that a single research is usually not enough to be mentioned in an article
I am not surprised that the sentence existed once. But "Wikipedia once said this during the time August 2024-Februrary 2025" is not the same as "Source: Wikipedia" because of the way Wikipedia evolves. It's especially bad if you disregard the entire much more nuanced discussion in the remaining article.
Certainly not the only country. Iceland is even more extreme in this regard and unlike Finland it is powered by 100% renewables, hydro and geothermal energy. In Finland the only good renewable alternative is wood/biomass.
Seems reasonable. I'll have to dig up my source to double check. Maybe they just didn't have Iceland data in their set? It's certainly a surprising result to see other non-sunny places like the UK, Germany, Norway & Sweden have solar as their cheapest energy source.
It's hard to get really solid estimates for solar costs because they've been dropping so precipitously, and because they depend on so many ancillary factors: wiring, inspections, permitting, power electronics, storage, and so on. Getting solid estimates for solar return on investment is even harder, because it depends on the future price of energy.
I think Finland just has a large enough extent from south to north that solar might be starting to become viable in the south but not in the north. While Iceland already produces more electricity per capita than any other country, using only hydro and geothermal, so solar is pretty much non-existent.
Windmills can be surprisingly expensive. https://www.eia.gov/analysis/studies/powerplants/capitalcost... is not up-to-date, but I think the windmill prices in it have changed a lot less than the solar prices; the 200 MW onshore project they price out there comes to US$1265/kW (US$1.27/W), of which something like 61% is the windmills themselves. Low-cost photovoltaic solar modules currently cost €0.055/W (US$0.065/W), lower by more than an order of magnitude https://www.solarserver.de/photovoltaik-preis-pv-modul-preis....
So, at equal cost, the alternative to a megawatt of windmills may not be a megawatt of solar panels, but 10 megawatts of solar panels. And that can compensate for their lower capacity factor.
I don't think people are building gas-powered data centers in the US. There's a data center crunch in the US because people aren't building them because they can't get the power because of the US's anti-renewable-energy policies.
Even the EIA-commissioned study I linked doesn't include that, but it is a potentially significant cost. If we take the median price of US$4702 per Texas "acre" from https://texasfarmcredit.com/resources/texas-land-pricing-gui... it works out to US$1.16/m². At 30° latitude your panels provide about 0.86 square meters of panel per square meter of land, or more like 0.3 with trackers, so the land price is on the order of US$3/m². A square meter is nominally a kilowatt of sunlight, so that's US$0.003/W of sunlight, but mainstream panels are usually only around 21% efficient, so it's more like US$0.015/Wp. Historically this has been insignificant but may no longer be with mainstream panels costing only US$0.10/Wp.
Desert land, lakes, and harbors are cheaper, so we should expect to see more panels there instead of on potentially arable land.
Sorry, that's Texas farmland "acres". An "acre" is a medieval unit of measure defined as one "chain" by one "furlong", the area a single man can plow in a day with a team of oxen. Although people have been plowing with horses since the 12th century, the "acre" is still in use in Texas, where it is roughly equivalent to 0.405 hectares (4050m² in SI units). In Texas, latitude ≈30°, it amounts to roughly a megawatt of solar energy (3 megawatts peak) before accounting for panel inefficiency.
Also when I said "US$0.10/Wp" I was wrong. I'm in the lazy habit of rounding US$1 = €1, but that's a significant error now. The correct price of €0.100/Wp for mainstream solar modules is more accurately US$0.117/Wp.
> An "acre" is a medieval unit of measure defined as one "chain" by one "furlong", the area a single man can plow in a day with a team of oxen. Although people have been plowing with horses since the 12th century, the "acre" is still in use in Texas...
1) People plowed with oxen well into the 20th century. Most places, only fancy people could afford horses at least into the 17th-18th century. So not so totally-medieval.
2) The acre is used in all kinds of backwards (Anglophone) places, not only Texas. All of the USA for starters, probably Australia, maybe the UK... Heck, I remember my elders using the (roughly) corresponding "tunnland" in daily conversation in Sweden as late as the 1970s. (But yeah, they were really rather elderly.)
3) Aren't you the guy who should call that "the 0012th century"? (Sorry if I'm getting you mixed up with someone else.)
Actually, calling geothermal energy "renewable" is a bit of a misnomer, isn't it? At least if the heat energy in the Earth's crust, which is what "geothermal energy" harvests, comes from the inside. The Earth's core may not be cooling down very fast, but we know for sure it's not getting any warmer (not before the Sun in its death throes swells up into a red giant and swallows the inner planets, anyway).
Yeah, I know, super-nitpicky — but, hey, it's the Best Kind Of Correct™. (Unless the crust is actually heated more by the Sun than from below, but I doubt that.)
Most geothermal energy comes from the decay of radioactive elements in the Earth's crust, although heat from Earth's formation is a non-negligible fraction of it. If you check out the web site of Iceland's geothermal energy agency, I believe they do have a calculation there of the sustainable power level that could be extracted (without cooling down the crust), but I don't remember if they're currently above or below it.
If I recall correctly, however, the fossil heat trapped in the crust under Iceland is several billion years of the sustainable extraction rate.
And, on the third hand, even if you only extract energy at the rate that radioactive decay produces it, in only a few tens of billions of years, most of the radioisotopes will have decayed away if you don't replenish them.
You are correct that the crust is heated more from below than by the Sun. That's why the bottom of the crust, where it contacts the mantle, is hotter than the surface.
During play, yes, obviously you need an implementation of the game to play it. But in its planning tree, no:
> MuZero only masks legal actions at the root of the search tree where the environment can be queried, but does not perform any masking within the search tree. This is possible because the network rapidly learns not to predict actions that never occur in the trajectories
it is trained on.
It is consistent with what the commenter was saying.
In any case, for Go - with a mild amount of expert knowledge - this limitation is most likely quite irrelevant unless in very rare endgame situations, or special superko setups, where a lack of moves or solutions push some probability to moves that look like wishful thinking.
I think this is not a significant limitation of the work (not that any parent claimed otherwise). MuZero is acting in an environment with prescribed actions, it’s just “planning with a learned model” and without access to the simulation environment.
—-
What I am less convinced by was the claim that MuZero reaches higher performance than previous AlphaZero variants. What is the comparison based on? Iso-flops, Iso-search depth, iso self play games, iso wallclock time? What would make sense here?
Each AlphaGo paper was trained on some sort of embarrassingly parallel compute cluster, but all included the punchlines for general audiences that “in just 30 hours” some performance level was reached.
The more detailed clarification on what "preprogrammed rules" actually means in this case made the entire discussion significantly more clear to me. I think it was helpful.
It was essentially banned via copyright for a long time. The only reason that it is available now is that 70 years have passed since the authors death.
