If I encounter something world-changing I'd consider it an ethical obligation to share it with the world. Show your work, be transparent about what you know and what you've found, and try to help others in replicating. Maintain humility.
Something I dislike about modern research is that there's not a lot of transparency or rapid updates. It wouldn't take someone much time to record a few additional videos and try to address concerns regarding the magnetic properties. If you really think the findings are important you should try to have public communications with other scientists where you share as much information as possible.
Too many people get caught up with playing politics and lose sight of the real goal, which is pushing humanity forward.
In theory you should privately invite someone to replicate your work, then publish it if the replication works. You'd probably get scooped though, so you have no choice but to arxiv it immediately.
> In theory you should privately invite someone to replicate your work, then publish it if the replication works. You'd probably get scooped though, so you have no choice but to arxiv it immediately.
What are we optimising for here? I'm not even convinced that that's better. What's the harm in the way they went instead, assuming they're not frauding.
I think point 3b depends on how you publish your findings. It can be either:
"I found the first superconductor material at room-temperature", or
"I found a material which seems to be superconductor at room-temperature, please help me to test if it is true or otherwise to find the error in my observation data"
The latter doesn't seem to me to be at risk of embarrassment, on the contrary it is how the scientific method should work.
4) Go to your government to get the biggest cash out immediately, for the lowest risk (they may be unhappy if you didn't disclose it to them first), also putting your nation potentially massively ahead of others.
They supposedly started playing with this in 1999. If they discovered this stuff nearly two decades ago and SK was let in on it, maybe they have a whole array of military projects built around it already?
I'm currently skeptical, as these are extraordinary claims.
But it doesn't matter what my current guess is. This claim either replicates or it doesn't. I really hope it replicates. If it does, then even if this material can't be directly used to change industry, it's quite likely to lead to materials that do. So I'm crossing my fingers!
it really seems like a "can we get people to believe in/falsely claim replication of a miraculous scientific breakthrough" study. Red Mercury, if you will.
I realize this is probably a geopolitical can of worms but if it exists and they've produced samples of it, send it to switzerland (CERN?) or someone and let them look at it. That's significant even in the absence of reproducibility of the manufacturing process itself, and if they have something, that's an easy way to prove it. The scientific world would love to look at a sample of a very unusual material.
but yes, it would always be nice if the world was reproducibly quite different than our current models imply, and if there's something novel here then this may just be the first in a family of materials that may behave similarly, once the effect is understood. 97C is a wildly high superconducting temperature, having hot superconductors would be world-changing, there's all kinds of places that would be applicable - you don't even need cryocooling at that point.
Most of the algorithms we use in machine learning are chosen specifically because they parallelize well. If we could increase clock speeds arbitrarily high, a huge selection of algorithms we passed on would suddenly become viable.
you'd be surprised how much of what you think is CPU limited is actually constrainted by memory bus bandwidth, locks, high cache evictions / bad cache use, or I/O.
something as seemingly innocuous as checking and incrementing an atomic counter across a bunch of threads can induce a 10x performance drop
it's totally batshit insane how much faster things run when they're able to work with stuff in L1 cache where they're not waiting on slower memory.
modern CPUs are insanely fast and crazy concurrent, but the rest of the machine can't keep up.
I think for tasks like parsing javascript the parser is usually unable to saturate memory bandwidth. Whether that's because it's parsing to an in-memory representation that is a tangle of pointers all over the heap or because it is full of unpredictable branches seems unimportant as a user.
The story is similar for LLVM. The fact that a CPU-bound task can be pessimized into stalling on cache and TLB misses all the time doesn't mean that it isn't CPU-bound.
A commonly deployed compression format which is lauded as revolutionary has a decoder that contains 0 SIMD intrinsics and uses significantly-fewer-than-optimal number of interleaved streams (according to papers published many years before the format was standardized!), which drives up the length of the dependency chains, and if you try to decode many chunks at the same time then you also get into trouble because you can't fit many tables into cache. The decoder does not saturate memory bandwidth and toy decoders of similar codecs from papers published years before are predictably faster.
The more ubiquitous compression formats are worse.
It's common to encounter hash tables that underperform robin hood tables that support bulk operations by a factor of 100, applications that use red-black trees with individually allocated nodes as priority queues (as opposed to ternary heaps or b-trees or whatever), and sorts that underperform by a factor of 10 compared to radix sort of multi-key quick sort because they are written in terms of a comparator function that must run in its entirety, so that the subset of the items whose first-sorted-by field are the same cannot skip this comparison. This comparison is commonly a comparison of data behind a pointer rather than data present in the struct.
I still submit that if you cranked CPU speed 10x or 100x higher than now that you'd just hit the wall so hard that actual quality of life improvements would not be 10x or 100x at all, unless you could simultaneously make bus speeds and memory speeds jump with it (presumably magic superconductor would do that, too, I dunno.)
In any case it's a pretty silly contrivance, how about the oil age? or the silicon age? Both of those are likely to remain more impactful than the LK-99 age, in the event that it does turn out to be an ambient temp/pressure superconductor.
While we're in this phase of Schrodinger's superconductor of 'waiting on replication', one thing is clear...
Science is broken.
If they really saw this as being possible in 1999 why has it taken so long to progress? Well, I guess it's because they had day jobs... but think of what could have changed if it worked 24 years ago... think of the carbon we've spewed out in a quarter century :(
We really need a multi-billion dollar x-prize for tech like this... and make it open-source so it's not paywalled behind a patent.
Edit: He's been breathlessly tweeting his hopes for the compound, and has now done complete 180. This is just meant to be a data point showing that one hypester has lost faith. His physics academic background (which is non-expert but still better than the average joe) shows that he's got a tiny bit of credibility, but his being of a failing coffee startup shows that he's probably got other motives.
Yeah, I think he knew his initial breathless positive tweeting would go viral and he was really trying to attract new investment for his struggling startup. Which he did.
Yeah, we should give the author the proper chance to replicate and full peer review before making a final call. I sympathize withe the frustration of work being seen prematurely, it's hard to undo the initial bias.
Also this guy was the one shilling the paper in the first place and he’s now in damage control mode as he knows that it’s BS and doesn’t want to loose more credibility in the fallback.
1) Get paranoid, figure out where you've gone wrong
2) Kicked it as hard as we can, can't find a problem. So...
3a) Keep this potentially important result under our hats for as long as it takes
3b) publish it with with some risk of embarassmment, and - hopefully - get others to either verify the result or point out where we went wrong.
I'd always go for b) - after strenuous effort to falsify my own theory.
Of course, some people - scientists are not an exception, just less susceptibl to it - can't bear to abandon a lovely result.