I'm not positive that's true. I think the most challenging thing for MOND is that basically no one will touch it unless they already have tenure because it's a death sentence.
I saw Pavel Kroupa (big name in Milgromian gravity) present in Heidelberg (big concentration of Astro), at the time Volker Springel (author of widely used LCDM simulation code "GADGET") was there and Illustris simulation sets (LCDM major project) had just been rolled out. And Pavel basically got heckled (in a very erudite and respectable way, but constant interruptions from the LCDM majority audience).
But Pavel had one slide, I can't find it now, but it was like 72 different problems that LCDM had not solved (ok, 72 is an exaggeration; you can troll his website for mentions of a lot of them https://astro.uni-bonn.de/~pavel/kroupa_SciLogs.html).
And like, Pavel's a big boy, he has some of the most cited papers in all of astronomy, he's got tenure, and he's set, so he can take it and not care. But grad students / postdocs I imagine would constantly have their work politely ignored and get shunted into underfunded groups.
I'm just trying to say that LCDM has things it can't explain, and MOND has things it can't explain, but the amount of resources in each theory is seriously lopsided so LCDM can frequently "tweak" itself to solve problems that MOND just doesn't have time or resources to to the same (for example disk formation in LCDM models used to be impossible until they had the supercomputing resources for the resolution required, and they found that the feedback coefficient could a) not promote disk growth, b) promote disk growth, and c) destroy disk growth, depending on how much they cranked it up. That's NOT a triumph of LCDM making an amazing replication of the observation, that's some grad student in a lab with enough CPU to tweak a meta-parameter until it looks good.
Also the CMB is extremely tightly constrained... and multiple huge tightly-constraining studies, WMAP, PLANCK, Gaia, are more than 3-sigma outside of each other's results, so... perhaps too tightly constrained.
I'm not sure if "accidents" count as "cooked" statistics... modern nuclear plants are much safer than 20th century ones as well, so if we're updating numbers I think nuclear is still an extremely safe contender (and if you split nuclear deaths by country, USA is an order of magnitude again safer than all nuclear). Maybe I misinterpret what you mean by cooked though, I would be interested in reading more about it, if you have a good piece? Wind also relies on pizoelectric elements, I'm not sure what their carbon footprint is compared to nuclear fuel, probably would be less in the future if like CA/OZ stepped into the rare-earth supply chain.
And new types, like thorium salt reactors don't have classical issues like uranium waste products and much much smaller possibility of runaway reactions. And France has been "5 years away" from fusion reactors for like 30 years on a shoestring budget, with real money that could be a possibility in maybe another 30 years and also dodges all the horror-story problems.
Overall, I agree, modern solar, especially reflector-steam plants shouldn't be lumped in with silicon panels, especially when talking about total carbon footprint. But I do think nuclear carries a lot of unfair baggage that has kept us on oil for way longer than we should have been.
Not on the scale of "we need to move the bulk of the energy from Texas to New York for 3 hours, then when the sun sets shift state-level power supply routing from the cornbelt to cover the drop."
Handling a plant going offline is totally different from the logistics of shuttling variable regions of production across the entire country.
if you look up deaths by energy production method, nuclear is far and away the safest, in every single country by orders of magnitude, even wind is deadlier.
Renewables that are popular now, like solar and wind, are weather dependent, which means the grid needs to handle variable input, which it's not designed to do and would need major fixing. Also means you need to store the energy for when it's not abundant, Germany does this by pumping water into mountain lakes to produce hydro on demand, but they've run out of lakes, so you need batteries. Batteries of a grid-size magnitude need carbon-expensive materials and have their own plethora of problems.
These two problems are only feasibly conquerable in a reasonable timescale by a few countries, for others they are prohibitively expensive, far beyond the small subsidies proposed in things like the Paris Accord. Now, sure, USA, Europe could implement these, but it's not going to fix the problem if Asia, Africa don't, and additionally it's going to add significant economic and geopolitical pressure between these nation-blocs, those who are hamstringing themselves and those who aren't.
Hydro has serious geopolitical issues as fresh water supplies grow more tactically necessary, consider Egypt / Ethiopia right now with the Nile, or China's tibetan plateau snow-seed cannons to capture water before it reaches India.
Tidal has extremely short lifespans, any moving system in salt water is difficult to keep going for more than a few years.
Geothermal is good, but also one of the most expensive and a bit of a "slow burner," not going to be powering anything serious with that without sinking billions into the plant.
The only real solution I see is nuclear, but that's a naughty word for some reason.
OCD isn't double-checking your line-spacing in a document, it's being unable to leave the house in under 45 minutes because you check the stove twenty times, then the door lock twenty, then turn around 5 minutes down the street to check again, then make a deal with yourself that you'll check the stove 3 times in a row and then not allow yourself to do it again, only you do it again anyways and finally take a picture of it so you can discretely check the photo on your phone when you're out on your date that you showed up half an hour late to.
This really rubbed me the wrong way, the closing sentence sums it up: "What if a doctoral program’s prestige arose, in part, from the way that it treated its students? We should dare to dream of such a thing."
There is a huge problem in academia where the number of PhDs is growing exponentially, and the number of positions is staying basically flat. I strongly disagree with this "participation award" mentality of "you deserve it." It's... leading a lot of people to make really stupid career decisions and I think a lot of people are doing it just to get this gilded club to bludgeon people who don't have PhDs -- that's the only reason I can think of for a person to hope for an "easy" PhD experience.
My advisor being rough at the right times (and in good spirit) was absolutely essential to the defense, which is essentially gladiatorial combat against your father (well, mother in my case), when both you and your advisor hope you can defeat the master in your specific realm. I would not have been able to do it if my advisor pranced about making everything pleasant and easy and convenient.
protip: if your prof is tenured and still first author on more than 2 papers a year, they're siphoning their student's work, or they're setting you up for pipelining positions (which are not horrible, but, be aware that's what's up).
I've never had a prof insert my work into their own papers, they've hooked me up to contribute to other teams, but with us it was always "my" project.
If you get a chance to see them present at a conference (lots of these are on video online nowadays), check if they specifically mention their students in the presentation, that's a green flag.
I disagree a bit about freedom and flexibility in switching faculty, and I disagree strongly about the phrasing of "willing" to work with x faculty on y problem.
I don't care how smart a student is, they are in no way prepared to select a project that will lead to interesting findings, which is a must if you plan to continue in any capacity. A good mentor will have a project with proprietary data that has a 95% chance of successful publication. In this way the student won't get scooped (not through any fault of their own, it's just that a PhD student, even a good one, just can't compete speedwise with someone like me who already has a code library built up to do complex analyses, this is why proprietary data is essential, or an extremely good prof who has an idea that they are confident no one else is on). My projects were a mix, I had one with proprietary data, and one which was a legit eureka finding. But as you say, most profs just aren't good enough to do these Eurekas (not an insult, 90% of science is gruntwork), so they need the cloistered data playground that the child student can crash around in until they manage to build their horrible little sandcastle. That data depends on funding, which depends on a proposal, which depends on a promise to do project X on data Y, and the prof is going to get reamed next time they apply for funding if student Z decided it wasn't "fun" and went to some other prof.
Caveat: We had one turkish guy who was straight up getting academically abused by his advisor (like literally timing his lunch breaks, sending spies to make sure he attended class, super weird shit), and thank the heavens there was a judicial oversight in place for him to get a new placement.
Caveat the second: I thrive in a low-input environment, so my prof just put me in the sandbox and I built the most magnificent and beautiful sandcastle the world has ever seen and roared my way to a whopping like seven citations (lol). My prof was also the director though and the low-input environment wasn't like optional, and not everyone did well in this low-input environment.
FAST can also move it's "mirror," the shell has some huge number of triangular sections on wires, so it should be able to point even more than arecibo which has an immutable reflector. Both have adjustable focal points (sensor head) on wires like you describe.
FAST is actually having difficulty hiring (on-site specialists like operations director) the smartest minds because it's in a jungle in small-town China.
But yes, overall China is executing an excellent "brain drain" (not meant in a negative sense) by being willing to invest in these "keystone" projects when other countries are tightening budgets.
This is correct to an extent, the surface of a mirror needs to be polished to a degree such that abnormalities in the lens are small relative to the wavelength being observed (I think it's the diffraction limit equation, not positive atm). Radio telescopes (with cm wavelengths) require much less precision than optical ones (with sub-micron wavelengths). This telescope, FAST, actually is made out of a collection of triangular (I think) sheets arranged into a kind of dome, if it were observing optical it would act like a disco-ball instead of a parabolic mirror. Arecibo is literally a hole in the ground with rocks and crap on the reflector.
This thing actually had "first light" observations with several holes in it from triangular sections that were fritzing just kind of fluttering in the breeze. The first radio telescopes were literally built out of post-coldwar trash just kind of rigged up in the back-yard. There's a major array going in in south Africa where the antennas are like christmas tree wires that are just kind of hammered into the ground, one of the PIs showed video at a colloquium I attended of his 10 year old kids setting them up. If anything in modern science can be called "low specification" it's radio telescopes.
Why does this scientific publication read like an op-ed? Absolutist language strikes me as exceptionally unscientific.
"it’s time to panic."
"We are in deep trouble."
"That’s it. Forever."
"Forever. Think of what that word means."
Disregarding that, the paper seems to address the idea that stratospheric albedo modifications cannot counteract atmospheric carbon buildup since they operate on timescales of decades versus millenia. Specifically they state that:
"Deployment of albedo hacking does not in any way “buy time” to get carbon dioxide emissions under control, since once emitted, carbon dioxide cannot to any significant extent be unemitted with known economically feasible technology"
Which seems to rest on the absolutist premise that we will never know an economically feasible way to go negative carbon. Otherwise, buying time should absolutely be a reasonable thing to consider.
The author closes with a sentence along the lines of:
"To decarbonize, however, requires building a political movement that regards the climate crisis as a top priority."
Which strikes me as... not exactly impartial (Why would not an economic reason work? Or a grassroots social reason?). Which is obnoxious when the author makes specific appeals to the authority of their profession which has authority by the very virtue of being impartial:
"As a scientist, I viscerally dislike repeating myself; I like to think that once the truth is out there, it will somehow win out and it is not necessary to belabor the point."
In NLP (specifically vectorizing words, ala word2vec) there's a famous test of whether or not your training has worked properly whereby you calculate the vector of "king" and subtract the vector of "man" and add the vector of "woman," if your machine is properly tuned, you should end up with a vector close to "queen" or "princess."
I wonder if similar things can be done to address specific (i.e. racial or gender) biases in computer vision.
There's a good book on this called "Why Nations Fail," [1] they posit that extractive economic policies (corrupted governments) essentially de-incentivize entrepreneurial businesses, while inclusive policies cause them to thrive.
They use this argument to counter "Guns, Germs, and Steel" [2] to explain dramatic economic differences across country borders in similar environments.
Sure, but those "constraints" could be compared to switching from a piece of letter paper to a canvas. The modern "liberations" would be having a box of a million tiny artistic legos which are getting tinier and tinier.
In all situations you want the best piece of art, it's not entirely clear to me that a million hand-painted legos is superior to an artist's canvas painting. Sure, it's a lot easier to grab all the blue legos and have 30 hours of deep house, but it's very difficult to end up with starry night, with bits of euphoric light trance and depressing shadow dub melded together into your blue canvas, unless you go back to your 1980s flirting skills and spend 20 hours making a thoughtful mixtape.
A paranoid individual might claim that this ink could be tampered with or replaced with washable ink.
Can the same argument not be made for requiring voter IDs? youeseh I'm replying to says it's "few numbers" adding up. Another thread from sverige in this tree links vox and nyt saying that requiring ID has small numbers effect of turnout. So saying that their concern of voter fraud is a small numbers thing kind of lacks heft as an argument.
If these are both the case, then why not go with the requirement (and possibly make the requirement "easier" by issuing IDs at the same time/place) so there's no possibility of contesting the election results with regard to this particular issue?
I saw Pavel Kroupa (big name in Milgromian gravity) present in Heidelberg (big concentration of Astro), at the time Volker Springel (author of widely used LCDM simulation code "GADGET") was there and Illustris simulation sets (LCDM major project) had just been rolled out. And Pavel basically got heckled (in a very erudite and respectable way, but constant interruptions from the LCDM majority audience).
But Pavel had one slide, I can't find it now, but it was like 72 different problems that LCDM had not solved (ok, 72 is an exaggeration; you can troll his website for mentions of a lot of them https://astro.uni-bonn.de/~pavel/kroupa_SciLogs.html).
And like, Pavel's a big boy, he has some of the most cited papers in all of astronomy, he's got tenure, and he's set, so he can take it and not care. But grad students / postdocs I imagine would constantly have their work politely ignored and get shunted into underfunded groups.
I'm just trying to say that LCDM has things it can't explain, and MOND has things it can't explain, but the amount of resources in each theory is seriously lopsided so LCDM can frequently "tweak" itself to solve problems that MOND just doesn't have time or resources to to the same (for example disk formation in LCDM models used to be impossible until they had the supercomputing resources for the resolution required, and they found that the feedback coefficient could a) not promote disk growth, b) promote disk growth, and c) destroy disk growth, depending on how much they cranked it up. That's NOT a triumph of LCDM making an amazing replication of the observation, that's some grad student in a lab with enough CPU to tweak a meta-parameter until it looks good.
Also the CMB is extremely tightly constrained... and multiple huge tightly-constraining studies, WMAP, PLANCK, Gaia, are more than 3-sigma outside of each other's results, so... perhaps too tightly constrained.
Edit: Found it, here is the great astronomical bloodbath of 2014: LCDM (Springel and Rix) vs MOND (Kroupa). Great watch. https://www.youtube.com/watch?v=UPVGDXNSBZM