Hacker News Reader

Comment by duchenne | Hacker News Reader

I see many comments wondering why the original authors do not reveal their synthesis process.

The reason is simple. They work for a private company. Not a university. Not a public lab.

They do not reveal the process for the same reason than openAI does not reveal its process. It is because they are sitting on a quadrillion dollars opportunity and they want to grab it.

Similarly as OpenAI, many initiatives are trying to make an open source alternative so they have a risk of not profiting from their extraordinary invention.

View on HN

I understood that they have disclosed the process, and are actively advising some of the groups to support replication.

However, they acknowledge that the process is difficult to get right and claim that they have had to run it many times to get the results that they are claiming.

I don't have any clue about materials science, but I can imagine why this would be the case.

The lab has a patent though which a lows for disclosure while also profiting from it. Thus I am not totally convinced of your claims.

AFAIU, the patent is limited to South Korea.

That is insane and would be a major error if true.

As far as I know, there is no such thing as an international patent.

What is called an international patent is just a standardized application so that only one national patent office has to examine it. If it is accepted, other national offices will grant the patent much more easily since the hard work has already been done.

And here, there is an international application, so what I think is that there is already a patent in Korea, but it is still pending for other countries.

Wait, correct me if I'm wrong, but aren't patents only valid in the country they are filed in? So to be able to enforce a patent, you'd need to have filed in the US, Europe and China(?). Also patents take a couple of years to get accepted...

Filing in juridictions worldwide is a routine process when you apply for an important patent. Any respectable patent lawyer should be able to deal with it.

But you have to choose to do it and there is an additional cost, if I am not mistaken. When our startup filed our patents in the US market, we chose not to do the same in Europe...

Sure, and most large corporations with worldwide influence do it automatically.

The last action one of my previous employers took was to send me payments for a patent I'm named on when they filed it in the EU. I had already quit, but it was easier to just send me the check than to cancel it :-)

Good point about that! Obviously there are alot of big players like that. However I heard this was a small lab, so I guess we don't know yet what their patenting power is.

But they did reveal their synthesis process, they just didn't explain it very well and there seems to be some random chance involved.

They revealed quite a lot, but the process seems to be unreliable and frankly unready for publishing.

If I understand correctly, one of their colleagues forced their hand by dumping the first preprint on Arxiv without consent of the rest of the team.

If that'd be the case, why publish the paper at all? Just fill as many patents as possible (assuming this would be helpful) and send the technology

Rumour is that the first paper was from a disgruntled ex employee.

The rumours of a disgruntled ex employee where, in fact the rumours of a disgruntled ex employee /s

The publication of the paper on LK99 was an accident.

> I see many comments wondering why the original authors do not reveal their synthesis process. The reason is simple.

Yes, but the reason is a different one: It's that they're busy writing / publishing a journal-grade paper and so they have postponed helping out other labs with reproduction efforts to the time after that. (Source: Hyun-Tak Kim said this in some interview.)

Then why to publish a paper at all?

Reply on news.ycombinator.com