Hacker News Reader

Comment by carabiner | Hacker News Reader

People were saying the initial papers were trash in the first day of their release over a week ago, so that's ancient history and immaterial now. Same with the "replications," they're obviously insufficient when they're generating pieces the size of a pin head that can't be attached to probes. All of us are in a holding pattern waiting for some big lab like Argonne to confirm or deny its SC status.

View on HN

How would you *deny* the SC status?

Replications keep failing until everyone is pretty sure it's not it. Proving the absence of something is rare to see.

> How would you deny the SC status?

Maybe we should adopt the stance that it’s not a SC by default and have the people asserting that it is prove that it’s a SC.

I think this is functionally equivalent to the status quo?

Not on HN where it’s definitely a SC we just need more time for the miracle to happen, failed replication attempts be damned.

Or even better adopt no stance on truth value and rather adopt a stance on whether there is sufficient evidence to expend further effort.

No need to guess at the truth. Just remain agnostic until you know the truth.

Exactly. That's the only reasonable stance. Everything else is just jumping to conclusions, positive or negative.

Sometimes a decision has to be made though - e.g. is a particular person a murderer.

In most countries, the burden of proof is on the accuser and the accused doesn’t need to proven his or her innocence and is treated as innocent until proven otherwise.

Sure, but science isn't a murder trial, it is entirely evidence based. Show the evidence, if there is none we'll move on and if there is you get a Nobel (in the case of superconductivity at room temperature and ambient pressure).

We (the rest of the world) can't prove a negative, so it is up to the claimant to supply the proof. Given proof it will be believed, no proof then it's a no-op.

With the murder trial you can assume a-priori that there is a murderer, the question then is 'who', not if a murder is possible or not. These are entirely different problems for a logic perspective.

Checking the original samples, for one.

I hope that ultimately it will come to that because I don't think anything else will put this to rest one way or the other (unless it is 100% confirmed/reproduced by another lab). It also seems to be the shortest path to a resolution.

That also adds little value. It would add no information on whether or not the sample is a fluke. We need to understand the material properties and its composition.

If it is a fluke or not doesn't matter. The only thing that matters is whether or not it is real, not how it was made. If it is real then we can expend a fortune on the 'how it was made' question, if it isn't real we can all descend from this holding pattern.

What benefit does that bring?

Appropriate allocation of resources is intrinsically beneficial.

Exactly right. If we concede that our present level of societal R&D is appropriate then it's easy to argue that resources are presently wasted pursuing less promising leads.

Until we have definitive evidence that LK99 is a SC, there's no reason to believe that it is a promising lead.

If it's real, "we" can brute force a reproduction. Just make each graduate student that works in a superconductivity lab spend two random months trying variants of LK99.

The attempts will fail. The idea is that the rest of the time the graduate student will be making standard superconductor to collect enough boring papers to escape the rat lab. But if the LK99 clone is a lucky one, everyone in the lab gets a groundbreaking paper.

To stop the waste of resources in case the whole thing turns out to be a dud. By providing proof that it is not many more resources would become unlocked.

Grinding against worse leads is not a better use of resources.

If the original sample is superconducting, you can perform XRD / other analysis on it to establish a proximate structure and then simulate things / find better methods for synthesis.

Couldn't there be solid evidence that the observed behavior was due to a different effect?

Maybe the original LK-99 sample was contaminated with actual high temperature superconducting impurities.

There is, a combination of strong (but not amazingly strong) diamagnetism at low field strengths and soft ferromagnetism (in floating samples) at high field strengths. Which is exactly what the paper in the link observed while getting the same results everyone else has. The only unambiguous sign of superconductivity was the original resistance graph which can apparently be explained by contact issues with four point probes (you can argue that they must have done many experiments and surely didn't cherrypick just the one where they found "superconductivity", but consider the quality of the rest of the paper...)

Doesn't diamagnetic levitation require a specific arrangement of the permamagnets that is not recreated in these videos?

That's not clear, but either way--as I've told you several times--there has been no credible demonstration of stationary levitation for even very weak uses of the word "credible."

This is explained in Fig. S2 in TFA.

What you'd have to do is 1) create bulk LK-99, 2) prove that that's what you'd created, 3) possibly also show that it's monocrystalline or something, and then 4) show that there's electrical resistance.

The problem is that the stuff is really quite difficult to synthesize reliably. Until that can happen, good luck finding out in either direction!

It's definitely difficult to synthesize reliably when people dismiss any replication attempt that doesn't show superconductivity as a priori not being the right compound.

Heh. But it does sound like it's honestly pretty difficult to make, too.

Cool it down until you DO see a SC transition far below room temperature.

Reply on news.ycombinator.com