String theory lied to us and now science communication is hard [video]
youtube.com
youtube.com
Another example is the state of communication about fusion power right now, which is deplorable. The media has been full of claims about how they finally achieved "net energy gain." Except "net energy gain" is a narrow scientific description that does not, in fact, mean what we would normally consider net energy gain for viable fusion power.
If we look at the actual total power input to the experiment in question, and the power output, we see a 99% energy loss: the total output power was only 1% of the total input power. That means we're 100 times, two orders of magnitude, away from achieving an actual energy gain.
On HN there is probably a good amount of awareness of this, but I bet there are some people reading this comment who were not aware of it, and were misled by the reporting. Here's an article for the unfamiliar: https://whyy.org/segments/why-the-nuclear-fusion-net-energy-...
So I think the headline could be generalized: science communication has a high degree of dishonesty in general, and that makes good science communication hard.
Most people fundamentally misunderstand the issues around fission and its economic viability even though there’s hundreds of nuclear reactors. No base load power isn’t inherently a good thing, it was actually coined to describe a downside. Clarifying just that misunderstanding gets info market forces, investments, and engendering to a non trivial level but you essentially need to condense it down to an infographic or people will lose interest. I have no idea how you could pull that off in an interesting manner.
IMO, the trick is to narrowly address the common misconceptions. Flat paper airplane wings provide lift, therefore Bernoulli’s principle isn’t why wings work. However that doesn’t extend to actually explaining complex topics.
It’s very easy to tell people that the latest experiment only produced 1% of the power that was put into it. Instead, they were told that the “net energy gain” was positive, and that this was a major breakthrough that brings viable commercial fusion power significantly closer.
I suppose, when you have experiments costing tens of billions of dollars being funded by the public, that this kind of corruption of the truth is inevitable. But that quite understandably undermines public trust in science communication.
The fig leaf of nuclear power generation is why you get the seemingly ridiculous Q factors based on energy in the pellets. Otherwise the obvious suggestion to cancel the program if ITER demonstrates Q = 10 while these things are Q = 0.01 could gain steam.
then news reports should say "nuclear fusion computer models improved by 5% today" and be on page 94, instead of lies about fusion power progress on page 1.
I would be cautious at take anything they’re saying at face value.
That statement is false on its face. The press release contained more of the same. Some examples (from https://www.llnl.gov/archive/news/lawrence-livermore-nationa... ):
> “This astonishing scientific advance puts us on the precipice of a future no longer reliant on fossil fuels but instead powered by new clean fusion energy.”
> “This monumental scientific breakthrough is a milestone for the future of clean energy.”
> “This significant advancement showcases the future possibilities for the commercialization of fusion energy.”
As I said, there are plenty of ways to put out press releases that don’t claim world-changing results where none exist. The idea that this was a necessary cover for the weapons applications is laughable.
Australia has a related problem, where anti-climate-change-action forces on the Right have pivoted to supporting nuclear fission as the solution - Australia has no nuclear fuel industry and no existing nuclear power reactors, and so pushing that has the effect of delaying any actual climate action for decades.
That’s why the Fusion Energy Sciences program within the Department of Energy’s Office of Science was allocated three quarters of a billion dollars in the 2023 appropriations bill.
Here’s a list of institutions involved in fusion research in the US alone:
https://science.osti.gov/fes/Research/Fusion-Institutions
There are similar lists for all the countries involved in the $40+ billion ITER experiment.
Of course, there’s serious engineering involved in these projects as well, just as there is for the LHC, for example. But presumably you wouldn’t call the research being done at the LHC “particle engineering”.
Collier is hard-line Karl Popper - theories must make testable predictions which are later confirmed by experiment. It's good to hear that.
Around minute 34, she's covered the background and gets to the political point. "String theorists are liars". This has caused public hostility to physics. Which translates to funding cuts for particle physics.
The crux of the video is that string theory isn't really a theory since it's untestable, and that string physicists/authors have overrepresented its utility for decades. She goes through a brief history, and the broken promises (each decade: a new revolution from string theory is just a decade away).
But the arguments around the public erosion of trust in physics involve a few cherry-picked examples that and ignore the wider context around academic research and issues communicating science to the public, especially scientific sensationalism from media outlets (eg God Particle). She argues that the string theorists have played into this and should have been more media savvy in the past, but is relatively unscathing of other non-string theory physicists.
It just gets ranty at the end, "the string theorists lied to us for 30 years" gets repeated a lot.
You can run the YT subtitles an LLM to get a summary if you don't have the time.
I'm gonna say it's worth the hour, and I think you might be falling into the trap of media communication about science. If something is worth communicating to a lay-person, it's worth going into detail about to avoid the exact pitfalls that you might fall into if you were to digest a 10m version of it. If you don't want to watch the 60m version of it, I would say "stay out of the discussion on it."
Beyond that it's expanding on some history and being more detailed than the single sentence.
Anyway she's a great communicator and has the right kind of sass (this is probably her sassiest). Give the rest of her catalogue a look.
I would have cut the video to 10 minutes though.
the most important material in science
The most important material in science is glass. This is a video about the history of glass. Glass is important. Glass is complicated. Professional scientific glass blowers are the coolest people in the world. And a little discussion on art versus science.
My benchmark for a great glass video is (say) any > 30 minute full piece video of Lino T. working:
https://www.youtube.com/watch?v=DgbtaKfab7k
he's the Robert Fripp of cane work, the Eno of free form, ...
EDIT: here's a short nail biter for you: https://www.youtube.com/watch?v=7qt8-5Vx1HA
That was the most stressful video I have seen in a long time.
It’s eroding rational sciences as a whole.
Research papers are about scientists communicating with one another. It doesn't matter if it is a preprint like arxiv, Nature, CVPR, or a fucking blog, these are all scientists communicating with one another. I can't wait to be out to just take my time with research, post to OpenReview and my blog, and move on not wasting my time explaining to a reviewer why metric we reproduced is different from the one in the original paper despite already including references of other works who also could not reproduce the original number and getting rejected because I had a broken link to my appendix. It's a fucking waste of time and it is a complete embarrassment that we call this process science.
The crisis in the sciences is very much a political one. Within a capitalist system, the sciences are actively hindered from being in the position to solve humanity's big problems, because bourgeois values so often dictate how and what is researched, funded, developed. (Profit oriented)
In this book was also the fist time I heard critical analysis of the big bang theory (aside from creationists in my youth), which I actually thought had a wide consensus. It really makes no sense when thinking about it materialistically. All the laws and known processes involving matter are supposed to have come from nothing? Helped me see how the big bang theory is nothing more than naked idealism. No wonder catholic popes got on board with it so quickly.
Another interesting thesis presented is that medicine in its form today is not actually a real science - infected by positivism and pragmatism, the way it's practiced is like a constant race to cover up symptoms rather than aiming to understand the human body in its entire complexity to heal issues at the root instead.
Anyway, highly recommended read. Covers several scientific fields and how they are being plagued by regressive ideology.
I really hate these lines. Not because capitalism doesn't have loads of problems but rather that it is missing underlying principles and incentives that exist to create these structures and incentives. They are coupled with the economic systems but not strictly dependent on them. These issues aren't fundamentally broken because the privatization nor nationalization of capital resources. Many of these issues exist because of metric hacking which exist due to bureaucratization and politicization. So in that sense it doesn't matter if your society is capitalist, socialist, feudal, monarchistic, anarchistic, or whatever. It has to do with how humans fundamentally organize structures, delegate, and evaluate. It is a crisis of the 21st century because only now have we had such scale and such specialization. But as complexity has been increasing at an exponential rate we are simplifying our metrics and decreasing trust.
The idea that we need to continuously monitor people for performance (measure them), rather than (for example) accept them into a community and just let them evolve there (for better or worse), is very intimately tied to capitalism.
Also, all the KPI systems based on objective metric ignore a really simple metric, that you can simply ask every scientist an opinion (for example, who are the leading people in your field) and use that as a metric, as a recommendation what kind of research is to be financed more and what kind is to be financed less. Again, this is because capitalism cannot trust employees by default, it has to put some other authority in charge of them.
Sure, metric hacking is not a problem when the crux of the matter is actually bending truth to follow the party line or face repercussions. Repercussions that could be years of imprisonment up to death penalty.
A cursory research on "scientific research under communism" yields [0], which seems to line up with the experience from people coming from former communist countries (including Mao's China and some former Eastern European countries).
I am also reminded of Luzin [1]. The previous article does not even mention pure mathematics as falling under repression, but even this domain (which was a strong point of Soviet Russia) could send researchers into concentration camps for dubious political motives.
So yes, metric hacking does not matter when the whole thing is bogus.
[0]: https://en.wikipedia.org/wiki/Repression_of_science_in_the_S...
[1]: https://en.wikipedia.org/wiki/Nikolai_Luzin#Luzin_affair_of_...
He was never higher ranking (didn't get professorship) until the 1989 revolution, AFAICT he was more interested in research (thermodynamics) than career.
During real socialism, academia was often a refuge of smart and freethinking people, if you were able to get there (some didn't for political reasons). It wasn't a highly paid job. And it also was very inefficient. There were completely useless departments, and the department heads were political positions. But some good science got done, without having to toe the party line, and without having to bend metrics and having to publish lots of repeated nonsense.
(I am not advocating that particular system, I think it was worse, but I am pointing out that measurement is not inevitable.)
One result of liberal (capitalist) approach to science is that people get really good at marketing themselves. That's why science in the Eastern Europe often gets bad optics, because that self-promoting aspect of it wasn't as prevalent there.
Yeah no one should doubt that there wasn't good science coming out of Soviet Russia. Anyone that thinks so is ignoring a lot of history and this belief contradicts the entire premise of the Cold War. I mean how many countries have sent probes to the moon (~9 of them)? What about to other planets? What about Mars? (US/USSR/China. USSR and Russia all crashed and China is only 2020). And they produced a lot of good mathemticians, even if they discriminated against the Jews (google "Jewish Math Problems").
I agree, the system is definitely not preferred and too political, but no one should doubt that there weren't world class researchers from there. My concern is really that we're adopting such bureaucratic prioritization. It's also worth noting that this dominance of bureaucrats/politicians played a major role in the Chernobyl disaster, though neither should discount the fact that this was highly classified research at the time given how new the technology was. (Also let's not forget what happened to academic in China...)
Under tyrannical regimes some people manage to live fairly reasonably well while many suffer. I don't doubt that but it doesn't invalidate a clear general trend.
During a long time under a capitalist regime academia was a refuge of smart and freethinking people. The "real socialism" sounds like a "true scotsman" fallacy. Why wouldn't you apply the same standard and consider that the time academia was working prerry well in capitalist countries was "true capitalism" and the current state of failing academia (to a degree) is a symptom of a "false", degenerate form of capitalism?
I think your comment can be interpreted in different ways, but if I was to go with the more measured version I think you should agree with Godelski's points more.
In particular there isn't even really one example of a communist country that evolved with the political stability that democracies did. I'm not sure that if such a thing had happened, it wouldn't have devolved with the same bureaucratic tendencies --which goes back to my first point: given the level or brutal instability metric hacking could not arise.
I’m sorry but you have a wildly romantic view of your fathers time.
I'd like to make you aware of this little known thing called The 5 Year Plans.
My point is, there existed different systems of doing science, which weren't measured as much (to the extent that would cause overproduction of articles that no scientist has time to read), because they had different preference than "scientific efficacy by an objective metric". For example, preference for somebody being authority in a scientific field for life. Or preference for academia to be a place for outcasts. Or preference for democratic decision making about what to scientifically pursue. These are all non-capitalist preferences.
I was talking about metric hacking, which isn't exclusive to science. I could have been more explicit though. But I still think you unnecessarily brought in a red herring by bringing up economic systems and I could have responded more appropriately. But our discussion is off the rails now.
> My point is, there existed different systems of doing science, which weren't measured as much
Science has always been about measurements and philosophy. It is the balance. The issue is about the over metrification as to what the science means. You're definitely right to point out these issues within capitalist frameworks. You'll find many of a comment from me complaining about how we do this within corporations as well. KPIs rule everything even when they mean nothing. Quarterly profits over long term profits. Yes, these are metric hacking and incredibly short sighted things. But at the same time it would be naive to assume that there are not parallels under communistic rule[0]. There are plenty of examples as there are under all the systems that I mentioned.
The problem here is that you're willing to look at flaws in a system you disagree with (which you may notice no one is claiming the capitalism is flawless nor are many even countering your critiques. Pay close attention), the problem is that you're unwilling to look at the flaws within your preferred system. There are no globally optimal solutions to problems such as these, and thus there is no flawless system. The irony of this is that by ignoring such flaws you're actually performing the very metric hacking we are complaining about and it would not be unreasonable to call this hypocritical. Think about comedians if we need an analogy: you can't make fun of others if you aren't willing to make fun of yourself even more. With belief systems it is similarly more important that you are more critical of the ones you prescribe to than the ones you disagree with. After all, don't you want to improve upon them? We can have discussions of which is better and which is worse but that needs context and honesty about the downsides of each system. But no discussion can exist if you're unwilling to admit to flaws.
[0] It's hard to give a specific example because the Scottsman card has already been played elsewhere. A difficult part of conversations in this whole abstract "communism v capitalism" is that no one agrees on the fucking fundamental definitions and so arguments are abound. IMO the capitalists seem to have lower variance but I digress. For an example, if you are one of the people that considers modern China communist then we can point to them. They're doing the same shit as we are in the US in terms of papers, but I'd honestly say worse simply because there's more of them doing it. If you don't think China is communist, then well I'm not sure we can really make comparisons if no true communist regime has existed and it's probably not worth talking about "true communism" if it is so vulnerable that every attempt has failed within a few decades (sorry, can't just blame it on the CIA. They play a role and did fucked up shit but they aren't wizards either).
But I didn't brought economic systems into it, others did. I was drawing on my own father's personal experience (as well my observations from the transition) in academia, which is that the metrization is not an inevitability for doing (even good) science.
Anyway, I also admitted the system's flaws. Again, I have heard the stories (BTW, the communism wasn't black and white either, different countries had different problems, even in Czechoslovakia there were about 4 different periods during 40 years of communism).
I don't know what system China uses internally to measure scientific output. Perhaps they adopted the Western system based on free market competition of scientific institutes for grants and students, and that's why they are also gaming the metrics. In that sense, their current system is based on capitalism.
Addendum: I would say since free market competition for labor is the main distinguishing feature of capitalism (together with its dual - private ownership of capital), it's not surprising that such a system is most prone to gaming this competition. It also manifests in the economy at large as "bullshit jobs". There are other parallels of current organizational problems in science with current form of neoliberal capitalism - rise of large actors (for example universities that are single-mindedly focused on publishing at the expense of everything else), abusing precarious labor (for example treatment of PHD students), etc.
I hate to break it to you, but not everything is about economics.
I don't recall having heard about the greeks complaining on how they had to fine-tune their treaties for the review board or journal acceptance.
Weird take considering I'm complaining about things that happen in America and I'm not sure of anyone who would claim America isn't capitalist (well... at least not communist, but let's not get in the weeds there). Of course science is hindered by incentives driven by capitalism. But science is also hindered by incentives driven by communism. That's my entire point.
> I don't recall having heard about the greeks complaining on how they had to fine-tune their treaties for the review board or journal acceptance.
Sure. But I do recall ancient Greeks murdering people who "believed" in irrational numbers. Idk about you, but I'd call this a step up. You're acting like there's a system that has no flaws or that if a flaw doesn't exist in one area we can ignore all other flaws. Which, ironically is metric hacking
Weird take...
Like my dude quota systems were at the heart of all successful capitalist, socialist, communist, and fascist nations. I can give examples. My man your logic is deeply flawed.
If you want a hyper explicit example it would be the cybersin project in socialist Chile.
Solutions to problems involve trade offs. Trade offs come down to questions about values. This is not he purview of Science. That's politics.
I don't know much about the crisis the OP is talking about, but I do know a bit about how our Australian Government gutted environment sciences because it was telling them things it didn't want to hear.
I wonder if the CSIRO needs to be an independent body like the ABC. (Not that the ABC is immune to pressure. )
Recommend listening to the video instead, I did awhile ago and I thought it was interesting and well done.
With physics in particular there’s a perception that science should progress fast in spectacular breakthroughs. But the 1900s was a very unique time. It’s not like that anymore. We’ve figured out the “easy” things (relatively speaking), most things left to discover are probably far harder.
So we need to get used to paying scientists and researchers to just play around with whatever they feel like, over long stretches of time. There’s no guarantee that a breakthrough will come from someone strongly embedded in the science mainstream writing lots of articles with tons of citation. And there’s no guarantee that breakthroughs will happen in any reasonable time.
I think we also need a shift from documenting/publishing the results, to documenting/publishing the process. YouTube is interesting in that regard. A lot of creators there doing very interesting and unique work.. and even if they don’t get great results, people still watch it if the process is interesting, and/or if the creator teaches their techniques, what they’re learning, how they failed, etc.
The more I read, the more I went “what the heck, this theory can’t possibly be true, it’s so ad hoc with rules made up just to make it pass tests that otherwise would prove it conclusively wrong”
It made me disillusioned and I eventually ended up working as a developer instead.
I guess my instincts were right, because there has been extremely little real progress the last 25(!) years in the field…..
If not, I don't think I'll bother with the rest. I watched ~10 mins and found the game a stressful distraction that impeded her communication (was that the point?). If it's just a random shtick I'll leave it for those that enjoy it.
Don't believe me?
Okay, Einstein said gravity is caused by the curvature of space, right? Wrong! Einstein never said any such thing, he merely said you can think of gravity as curved spacetime - in fact he later pointed out to others that was a crutch, not the reality. Gravity is a force. Yet the science popularizers continue to lie and display their stupid rubber sheet model to explain why things orbit the earth. Which is not only wrong, but very misleading - earth orbit are easily explained by time dilation, you don't even need to model earth gravity with all this spacetime-as-a-rubber sheet nonsense!
Still don't believe me? Look how science popularizers explain voodoo mechanics, erm, I mean quantum mechanics. They still pretend that Copenhagen and the science of 100 years ago still reign supreme and completely ignore quantum field theory, which is the foundation of the standard model.
When it comes to string theory they never point out that string theory, brane theory, and the standard model all yield the same results and that physicists now use the model that makes their calculations easier.
No, science popularizers prefer to spout nonsense in an attempt to make science seem cool. You know what's cool? Actual, factual science!