back
15 comments
150+ short clips make up the full interview. Playlist: https://www.youtube.com/watch?v=rs1jGsn61p8&list=PLVV0r6CmEs...

Edited to remove link to video that was in the original article anyway.

Laura Fermi has written a nice memoir about life with Fermi: https://www.google.com/books/edition/Atoms_in_the_Family/zpP...
If you watch the video attached in the footnotes, Dyson says, w.r.t the strong force's explanation, that "quarks were invented", rather than "quarks were discovered". This isn't a slip of the tongue, as later in the interview he restates the same claim in paraphrase. In that use of the word "invent" there is a deep understanding of how physicists view reality - perhaps as it ought to be. I believe that his use of the word "invention" is a mark of humility, and not hubris. The quark was invented and is a useful approximation to reality, but also that maybe someone else will come along, at a later time, and "invent" something a little bit closer to reality.
What an amazing life. Prescient choice to GTFO of Italy while he and his wife still could. https://en.wikipedia.org/wiki/Enrico_Fermi
Watching that video, it's astonishing to see him recalling that moment and having to pause and search for how to describe it, only to end up with saying "That was his genius". In two sentences Fermi demolished his theory and changed the course of this man's life.
What I'm wondering is where was my Enrico Fermi to warn me off of some of the bad paths I went down :-(

It would be heartbreaking to be told that what you are working on could never succeed--but not anywhere's near the pain you feel when, at long last, a death march comes to an end.

One can learn to think like Fermi. Perhaps not at his level of intelligence (this is to some extent genetically fixed, after all) but with the same qualities.

A good start is asking what the fundamental uncertainties are and which ways there are to resolve them. Try to formulate these uncertainties in terms of physical, mathematical, or otherwise experimentally established laws.

From what I understand, this is one thing Elon Musk is decent at, answering "is there any world in which this would even be physically possible, given no constraints on the innovation we can do"?

In a more fluffy example, I started out on a project recently and realised that if I improve on the state of the art along only one dimension, I don't think the product will be good enough to gain traction. So I need to outperform the state of the art along at least two dimensions -- this is significantly harder and I now doubt my ability to do this with the resourcew I'm willing to spend on it. That reasoning requires no technical knowledge, just thinking about the fundamentals of market adoption and what the end product could look like under no constraints on innovation.

   Because it was important for the students to have their names on a published paper, we did not abandon our calculations immediately. We finished them and wrote a long paper that was duly published in the Physical Review with all our names on it.
Isn't that shameful for the advance of science? Publishing something that you know is most certainly wrong.
It depends entirely on what the paper says. If it says "we found the perfect answer to this problem!", then yes, that is shameful. But if it says something more like "here are some potentially useful calculations, but the results are ultimately unsatsifying because of the cut-off procedures requiring arbitrarily chosen parameters", then it's still a useful result.
If they hadn't published it, comments here on HN would probably complain how negative results are never published. You can't have it both ways, though.
To be clear, I was meaning that it looks that they did publish it like they intended to. Not something like "this is what we tried but we are probably wrong".

But also, even in the other way, this is not a scientific experiment that failed. It is more like a non sense mathematical demonstration with some magic variable that were made up to get the expected result. So it does not really have any value except the content of the current article.

For example if I do a paper like: "Getting a glass of water from starting a TV"

We were able to demonstrate that we can get a glass of water by starting a tv with the following steps: - take the tv remote - power on the tv - assume we have XX - result: we have a glass of water - note: we considered XX to be "Michel brings us a glass of water" for this experiment. - note2: famous mathematician told us that our study is gargabe

I believe Dyson was a man of considerable integrity. I doubt he would have published anything misleading or plain wrong. But I haven't read the paper in question.
There are ways of presenting these results for what they are "we have explored the possibility of using cut-off functions to explain strong atomic interactions. It turns out that this solution is unsatisfactory because it requires four arbitrary coefiences."
Is there value in publishing the wrong even knowing that is wrong so that others might not have to spend the cycles doing the same wrong? Of course, that's assuming people are publishing wrong and not calling it right
Don't hate the player, hate the game.

Also, probably anything written by Freemon Dyson would be worth reading, just to see how a top-flight might engages with the subject matter, if nothing else.