Retired physicist and statistician, now a blogger.
Weekend Editor
Karma: 70
LLMs: psych subjects & quantitative BS
Absolutely! It’s not ductile enough for wire, and too frangible to bend around a coil even if you managed to make a long thin piece.
But… the early high-Tc superconductors in the 80s were ceramics, too. Even now, with much more friendly materials, the “wire” in the Commonwealth Fusion Systems tokamak prototype is actually a complex tape with multiple layers mostly for structural support.
Some details here: https://spectrum.ieee.org/fusion-2662267312
Here’s a very nice, more technical presnentation at Princeton by a CFS person, showing the tape strucdture, and how the material had to evolve from microcrystalline stuff to much more complex forms to be useful in an engineering sense: https://suli.pppl.gov/2020/course/20200619_SULI_HTS_Sorbom_Final.pdf
Also note: fusion-relevant REBCO magnets operate at 20T fields and 40kA currents, whereas this new superconductor can’t get above 0.3T fields and 250mA current. Lots of work to do there!
So I hope that gives the right idea: getting from today’s charcoal lump/floaty rock to something with optimized chemistry, easier manufacturability, ductility close enough to wire, and deployable in high fields & high currents took about 30 years the last time it was done.
It’ll be quicker this time, getting from the current charcoal to whatever works, because the incentives are higher. But it almost certainly won’t be simpler.
(Variant of something I put as a comment on Zvi’s blog.)
Yesterday I put up a blog post that walks through the 2 papers on LK-99 superconductivity in the style of what in grad school they call “Journal Club”: https://www.someweekendreading.blog/high-tc-sc/
It has all the hallmarks of something very much rushed into publication: misspelled words, awkward phrasing, out-of-order paragraphs, misnumbered figures, and (most charmingly) error messages in Korean from their bibliography software. At this stage of things, all that is understandable and excusable.
A few things are presented in a peculiar way, unlike most of the other sc papers I’ve read. Again, that’s more or less ok once you disentangle the coordinate systems on the plots and the like. It can be fixed easily, and probably will be.
To convince anybody to take you seriously enough to test for superconductivity, you have to demonstrate 4 things: (1) 0 resistivity below a reasonably sharply defined temperature, (2) the existence of a critical current above which transition back to normal occurs, (3) the existence of a similary critical magnetic field, and (4) the Meissner effect, or magnetic flux expulsion (totally for Type I and at least partially for Type II).
They did the first 3 of those reasonably believably (even to a guy like me who still has scars from the cold fusion mishegoss back in the day). The Meissner effect, though, gets only partial credit: the diamagentism for the field-cooled & non-field-cooled samples implies an unphysical value of the diagmagentism, and the picture/video of a sample on a magnet only sorta-partially levitates.
The diamagnetism curve has apparently been addressed, as the authors say it was simply a copy-paste error on the graph. The Meissner effect visuals could be explained by the fact that they have a polycrystalline sample (resistance between domain boundaries) of unknown impurities (sometimes coppuer sulfides, other times the Pb/Cu doping may vary across the sample). (Again, this is totally excusable given the rush.) When other people start preparing samples, we’ll see what’s happening here.
It’s not going to revolutionize anything in its current early form:
(1) Frankly, it looks like charcoal. It’s almost certainly not ductile enough to form a wire, and any long thin sample that looks like a wire would be too brittle to wind into a coil.
(2) The critical magnetic field is pretty low, maybe 0.3 Tesla at room temp. For comparison, the tokamak magnets being used by Commonwealth Fusion Systems for their prototype reactors weigh in at 20 Tesla.
(3) The critical current is low. The right thing is to report a current density, but they only report total current without information on sample shape. Still, they topped out at around 250 mA. For comparison, the CFS tokamak magnets run around 40 kA.
IF IT REPLICATES, it’s a fascinating step in phyics (mechanism proposes stressed crystals forming an array of superconducting quantum wells, with currents forming by electron tunneling along the Pb metal backbone… maybe).
But as it is, it’s not an engineering material.
THAT IS ABSOLUTELY OK! This is an early stage compound, all it has to demonstrate that it works. Then tons of material scientists and physicists will start tweaking the recipe, to optimize transition temperature, critical current, and critical field.
Also, I hope, somebody will figure out how use something other than lead. Right now it doesn’t use rare earths, which is good. But it would be better if we could use something with a similar crystal structure to the lead apatite (Lanarkite), but less toxicity.
Then last night all hell broke loose. Kwon made a conference presentation. During that, Lee and the other authors more or less disowned him, said he was fired from the university and the company, and that the first paper was an unauthorized upload by Kwon. Then they retracted the first paper that Kwon apparently wrote on his own and uploaded.
Drama drama drama.
I’m waiting for Argonne, which seems to be on deck for a replication trial next week.
“De-aligning”?
Yeah, I know: only if it were aligned in the first place. What little “alignment” it has now is fragile with respect to malign inputs, which is not really alignment at all. Prompt injection derails the fragile alignment train.
Well, I blogged about it at least a little bit. Possibly too elementary a level for most Less Wrong readers, though.
I was in your position last August, and wrote about my experiences with paxlovid (good, except I got a rebound).
In spite of staying on the upper floor, mostly in my office, my spouse caught it anyway. Omicron is very transmissible!
It’s also available commercially, e.g., from King Arthur Baking. It’s usually called something like “New England Boiled Cider”. It has a long and interesting history linked with the anti-slavery movement, since it was an alternative to molasses which was produced on sugar plantations in the South with slave labor. It fell out of frequent use sometime during WWII, I dont’ know why.
I use it a lot in apple pies, as a way to amp up the apple flavor without too much additional sweetness.
There also used to be a popular thing called “Boiled Cider Pie”, whose filling was a custard made from this. James Beard has an intriguing variation involving grated apples that’s in my kitchen queue.
Your 8:1 reduction is more extreme than I’ve seen before; usually recipes go for 5x or 6x reduction.
Somewhat amusingly, I wrote a paper in 1991 that makes exactly your point. (The linked page at the journal is paywalled, but you can find it if you poke around a bit.)
It was about systems with more than one decomposition into modules, for which there had to be multiple simulations. Those simulations had to be compatible in a certain way, and that led to exactly the commutative diagram you have above (figure 13).
For energy storage, you might consider the Ambri liquid metal batteries. They’re being designed for this exact purpose: coupling intermittent renewable generating capacity to steady loads.
Also, they’re cheap, rugged, and don’t seem to lose capacity over multiple charge/discharge cycles.
JN Crossley, et al., What Is Mathematical Logic?
A 96-page intro to the basics of predicate calculus, model theory, and Gödel incompleteness. I’ve used it in the (distant) past a couple times when a student had trouble getting a practical grip on logic.
If you feel the need to do something in response to the advent of fusion power and high-capacity batteries, you might want to think about doing it sooner rather than later.
Fusion: I’m beginning to think this is nearer-term than most of us believe. Last September, Commonwealth Fusion Systems demonstrated a 20 Tesla superconducting magnet with a bore large enough for their tokamak.
It’s a REBCO magnet (rare-earth boron copper oxide) which superconducts at nitrogen temperature, but they’re going at 10-20K to have some running room for high field strength without quenching the magnet.
The volume of the tokamak scales as the inverse cube of the magnetic field. ITER runs about 9 Tesla, so at 20 Tesla CFS has more than a factor of 2 increase in field, hence 1/8th the volume. Tokamak costs scale roughly by volume, so there’s potential cost savings of a factor of 8. (Potential, not yet demonstrated.)
Their reactor is a high-beta design, i.e., high plasma pressure. They’re building a demonstrator reactor now, expected to have a Q (power out/power in) of about 11. Target completion date is in 2025, only 3 years from now. 140megawatts of power, delivered in 10 second bursts.
If the demo reactor works, they predict building full-scale commercially useful reactors by 2030.
High-Capacity Batteries: Lithium-ion batteries are wonderful for portable applications, but… they tend to degrade after a lot of discharge cycles, and in high-power, high-density situations they have thermal runaway problems. “Thermal runaway” is excessively polite language for “halt, catch fire, sometimes explode”. The thermal management equipment and software are pretty gnarly.
Check out the liquid metal batteries from Ambri. They’re basically a highly insulated box with 3 layers of molten antimony, calcium chloride, and molten calcium. Discharge it, and the Ca atoms give up a couple electrons, the ions migrate through the salt layer, and form CaSb at the other end. Charge it, and the reverse happens.
Yes, its molten metal. But about 1 charge/discharge cycle every day (say, when coupled to a solar array) is enough to keep it heated, given adequate insulation. Unlike lithium-ion, it likes to be hot.
The materials are cheap. Don Sadoway, an MIT prof who founded Ambri with one of his students, tells the same joke in every talk he gives (and I mean every talk!): “If you want it to be dirt cheap, make it out of dirt. Preferably local dirt, so nobody can cut off your supply.” (He has a number of very engaging talks on YouTube.)
It appears to have no measurable degradation after hundreds of charge/discharge cycles. Obviously you can’t form dendrites in liquid metal.
The round-trip efficiency (power out / power in) is about 80% (with the losses going to ohmic heating to keep the metals molten). Pumped hydro storage, for commercial comparison, is about 70% or so. So the efficiency is very much on point.
Ok, they’re heavy. And full of molten metal. So they’re not going in your car. But for power plant applications, that’s just fine. Unlike lithium ion, they can’t catch fire or explode, and when frozen at room temperature for shipping they’re completely inert.
They’d be great for peak shaving: you have generation capacity for the average case, and use the batteries to store energy during low demand periods and supply energy during high demand periods.
They also couple ideally with solar arrays and wind farms, whose generation capacity is variable.
So there you go: 2 commercial interests in fusion and batteries, each with at least some chance of success. There are many others; it is very likely some of them will succeed within 10 years.
The evidence on vaccine efficacy waning is somewhat confusing to me!
On the one hand: Some of the initial data on waning from Israel (using Pfizer) was hopelessly confounded with age, leading to a huge Simpson’s paradox effect. Once you (a) figure out the Bayes error (they calculated Pr(vax | hospitalization) when they really wanted Pr(hospitalization | vax)), (b) properly stratify by age, and (c) calculate confidence limits on vaccine efficacy, the effect goes away.
On the other hand: Later Israeli data presented at the Moderna booster hearing cleaned that up and showed there was a waning effect. Moderna did something similar with their vaccine, comparing the people in the treatment arm of their clinical trial vs those in the control arm who got the vaccine 6 months later when it read out, showing a waning effect between those 2 carefully matched groups with known distributions of age, race, gender, etc.
On the gripping hand: Both of those show the onset of “waning” coincident with the onset of the Delta variant, i.e., last summer. So was it really waning, or was it Delta? I can’t tell.
They show a waning effect with respect to initial infection, but continued robust protection against hospitalization (still 85-90%). That could be normal:
Antibodies do decrease with time. You’re not carrying huge blood levels of antibodies for every virus you’ve ever encountered thus far in your life.
But T-cells and memory B-cells are still there. When an antigen from a previous infection is presented to the relevant memory B-cells, they trigger production of antibodies which then stop the infection. That way you can be technically infected for a couple days while that happens, but be asymptomatic or mildly symptomatic, and quickly clear the infection.
So it might be that we’re just seeing antibodies fade, but which rapidly come back upon re-challenge with the virus.
On the (unnamed) fourth hand: I looked at a recent study by Townsend et al. at the Yale School of Public Health that analyzed 6 coronaviruses known to infect humans and related to SARS-CoV2, predicting that there’s a 50% chance of reinfection at 18 months. So maybe periodic boosters will be required because of waning immunity?
JP Townsend et al., “The durability of immunity against reinfection by SARS-CoV-2: a comparative evolutionary study”, The Lancet Microbe, 2021-Oct-01. DOI: 10.1016/S2666-5247(21)00219-6.
So… “I notice that I am confused.” But since all this is happening frighteningly quickly and we’re just learning how it all works, confusion is the normal state of affairs at the leading edge of science. Thus my confusion might be normal.
Or maybe I’m just plain old confused and need somebody to straighten me out. That’s also regrettably normal, at least for me.
Until then, it seems that 2 things are worth remembering:
Boosters do seem to work in terms of boosting ab levels, and don’t seem to cause any worse side effects than the first doses.
If there is waning, boosters are an effective strategy; if there is no waning, then because of the non-terrible side-effects, boosters are a safe strategy, i.e., unlikely to do much harm.
So I got a booster, and my spouse gets one in a few days when eligible.
Yes, absolutely.
The comment on study power is why I was pretty surprised at the FDA and CDC approvals, based on this study which itself says it’s underpowered. I think they were mostly motivated by the safety results that boosters were at least as safe as the primers. So if it might do some good and probably does no harm, they can get to an EUA from there.
This is not the way they usually behave, but then again, these are not usual times.
“But something has really become clear: The mixing really is most impactful when you have a DNA/adenovirus vaccine first followed by the mRNA vaccine,” Gandhi said. WaPo
This is definitely true in terms of antibody fold induction: JnJ followed by either Pfizer or Moderna have the highest fold induction ratios.
However, they’re starting from a lower baseline, since JnJ doesn’t induce such high ab levels to begin with. (Though it might be better at training T cells and memory B cells, and have longer persistence? It’s kind of frustratingly complicated, to me.)
If you look at absolute ab levels, JnJ followed by Moderna looks best.
(Data source: FDA presentation by Lyke, cited in previous comment, slide 22.)
Have a look at the presentation to the FDA’s VRBPAC on 2021-Oct-15:
K Lyke, et al., “DMID 21-0012 - Heterologous Platform Boost Study Mix and Match”, FDA VRBPAC 2021-Oct-15 Materials, retrieved 2021-Oct-15.
(If you’re curious about the how the whole meeting went down, I wrote a little summary.)
Slide 22 (page 23, because the FDA tacks on a header page) should help. It’s a 3x3 array of the 3 original vaccines x 3 possible booster vaccines. It shows the factor by which antibody levels are boosted by each primer/booster combination (geometric mean titer fold induction).
For your case of a Moderna primer, the middle column is the one to look at: Moderna booster gave 10x boost, JnJ booster gave 6.2x, and Pfizer booster gave 11x. Given the error bars on that slide, Moderna and Pfizer are more or less indistinguishable (10x vs 11x), but both are likely better than JnJ.
In terms of absolute ab levels (the number shown in blue), a Moderna booster is probably a touch higher than Pfizer (3727 vs 2801), though I haven’t used the error bars to calculate the statistical significance of that. Both are pretty good.
For me, the bottom line is that it looks like either Pfizer or Moderna are fine for somebody who’s already gotten 2 doses of Moderna. The important thing is to get the booster if you need it; the choice of Pfizer or Moderna is over-optimization.
The way I used to explain “as treated” vs “intent to treat”:
Under AT, you’re asking a question about biology: does this treatment make a difference when it’s applied rigorously to the tests and absolutely never to the controls?
Under ITT, you’re asking a question about the practice of medicine, under actual combat conditions: if we tell clinicians and patients to do something, and they follow instructions as imperfectly as ever, is this treatment still a good idea?
That is: ITT bakes in mistakes, noncompliance, weird patients, etc. on top of the basic scientific effect.