Feynman Algorithm | Hacker News Reader

Feynman Algorithm

c2.com

91 pointsalexitosrv36 comments

Somehow, I'll take that "I have an unusual toolkit" theory over the "he's just super-smart" theory, especially since the later theory isn't really a theory but a "throwing-up of the hands" - and also, Feynman might know Feynman a little better someone else (though that's never certain either).

One thing I noticed in the Feynman book I read was that he seemed good at being willing to going for an outcome without engaging in the usual activity people think of as leading to the outcome - his safe-cracking involved a lot of social engineering rather than insight into the safe's mechanism. He got the safe open without "cracking" it.

Once you see the problem the right way, the solution will present itself. Instead of trying to "solve" the problem, work on trying to see the problem from different perspectives.

Like in chess -- "see the whole board."

Feynman said, "You have to keep a dozen of your favorite problems constantly present in your mind, although by and large they will lay in a dormant state. Every time you hear or read a new trick or a new result, test it against each of your twelve problems to see whether it helps. Every once in a while there will be a hit, and people will say, “How did he do it? He must be a genius!"

This was one of his "tricks" for finding associations among things others had yet to see.

The more associations you can find, the better you will see the problem and draw from different domains of knowledge. This helped him see things in way where solutions presented themselves.

"A change in perspective is worth 80 IQ points." --Alan Kay

I was telling someone about the safe cracking "algorithm" he had. Sometimes a problem seems hard until you recognize special pieces of information that dramatically change the constraints. In the case of the safe, he realized that the precision of a safe was to the closest multiple of 5, which reduced the set of combinations from 216,000 to 1,728) and then later that certain combinations are far more common than others (e.g. the default combo). It makes you realize that real brilliance is generally in understanding the problem and its constraints, not in solving it.

Another insight into his process was all the talk about envisioning math problems as "ok you have a sphere, and now the sphere has hair on it", and things like that. I think Feynman's brain was wired in a way that he could use strange pathways to come up with an intuitive solution.

Note however that the "unusual toolkit" and "being super-smart" are not mutually exclusive. He probably had both.

This is how I hear very talented physicists describe themselves and others in that category. Stephen Hawking did not achieve in spite of his condition, but rather because of it. It forced him to develop a very specific set of visualization skills likely unique to him. Feynman had similar skills. There is a story in which he was once able to calculate something that took a presenter 6 months in a single night due to his better methods. Similarly he was once able to derive a strictly better model of some physical system (liquid hydrogren iirc) simply by going off of the obvious interacting forces.

I once had this taped to the wall of my cube when I worked for a small internet company. One of the C-level employees saw it and loved it.

In company meetings when we would be talking about a genuinely hard problem, this exec would say "well what I think we need to do is write this problem on the white board, think real hard, and come up with a solution". Silence + confused looks. This happened multiple times.

He ended up being fired within 3 months.

I think a huge reason Feynman seems so brilliant is because he didn't learn physics or math in a traditional way. He explored. He had passion for it.

Not unlike the manner in which many successful hackers learn.

Nothing kills your interest in a subject like taking it as a class and getting grades.

Nothing kills your interest in a subject like taking it as a class and getting grades.

Maybe if you have a terrible teacher. The good teachers did nothing but enhance my interest in a particular subject.

For me, the teachers will enhance my interest, but the grades aspect of it will muffle it. I'm not exactly sure why this is.

I really like the variation offered by IraCooper on that page,

   1. Write down the problem
   2. Become convinced it's very important, then think about it
   3. Write down the answer

My own problem solving attempts are often thwarted by me working for 20 minutes and then getting discouraged and convincing myself that what I'm working on isn't really that important. After which I, of course, waste the next hour watching hour tv or something.

If I've heard anything about Feinman it's that he was always deeply interested in everything he did, and his accomplishments reflect it.

I think that many such great thinkers have, due to circumstances and early on, have developed

(1) a unique, uninhibited way of looking at problems, i.e. not different from what's generally prescribed to entrepreneurs: Consider every possible lead, even if people say it's stupid/has been done, etc. Einstein, famously had difficulty at school as a kid, so he actually thought about many concepts that we take for granted or find silly ("what would happen if I shine a flashlight while riding a bike").

(2) power to concentrate on a problem. When asked how he came up with his solutions Newton said that he thought on the problem until the solution came to him. This sounds simple but embedding yourself in a problem like that is beyond most people's focusing skills (I can't go by 10 minutes without checking HN, for example).

I can't help but feel like this page does a huge disservice to what he was all about. He wasn't some mystical super brilliant genius. He was a guy who had some particular skills in thinking about problems and communicating them effectively. I have enormous respect for him.

Maybe I'm being too sensitive, but it seems like WAY too many things on that page miss the point completely.

"mystical" -- no. But "super brilliant genius" -- yes. That's just the best quick way of summarizing it.

Alas, this is a "method" for achieving greatness that you or I can't replicate.

Feynmann was not "just a guy who had some particular skills in thinking about problems". This is just not accurate.

At age 24, his Princeton advisor suggested he join the Manhattan Project. He routinely dazzled the other physicists there, probably the greatest collection of physics talent (Bohr, Oppenheimer, Bethe, Fermi) gathered in one place since the Cavendish Lab in the early 1900s. And that was only the start of his long career.

Yes, I'm not denying that he was unconventionally an incredibly talented physicist. But in the process of figuring out his process (and the processes of lots of 'brilliant' people), we discover that we can't pinpoint exactly WHY or HOW he worked. That's the point. Nobody can. He was a human being and he certainly had flaws, which I imagine he'd be happy to acknowledge.

I just dislike putting people onto pedestals like this. If you had hung out with him over a beer, I'm willing to bet you'd be surprised how normal of a person he was.

Of course he had lots of flaws. But he does deserve a pedestal, actually. (All those folks on pedestals have flaws; some of them don't deserve their pedestal at all; Feynmann does.)

Also, I don't think he was a very "normal person". Being that brilliant tends to make you not very normal -- although he could impersonate a particularly sharp "normal person" if it seemed advantageous.

Read some anecdotes about his career and life -- for instance, the time he had a concussion and acted weird for about three weeks before anyone thought something was really the matter. That's not a "normal person".

It seems to me that you're letting your desire to not idolize him, get in the way of recognizing the reality of his genius and what he accomplished. That's not fair to reality.

Feynman and other physicists built a mass destruction weapon that they knew it was a mass destruction weapon. Feynman is a murderer http://science1.wordpress.com/2010/09/23/ten-most-wanted-phy...

Things to consider.

- Feynman got involved in the belief that the Americans needed to build this before Germany did. Which Germany certainly had the talent, motivation, and resources to do. In short if the bomb was going to be used, he didn't want it used against his country.

- Once the nuclear bomb was exploded, and Feynman realized what he had helped do, he fell into a significant period of depression.

- In later years he publicly criticized himself for getting caught up in the project and continuing it after Germany surrendered and it was clear that there was no race.

- Feynman died believing that his work on the nuclear bomb is what caused the leukemia that killed him.

It is easy to paint a picture of black and white and make people out to be pure villains. But reality is seldom so simple.

Nuclear weapons are perhaps mankinds most terrible invention - but it is fairly clear that political leaders on both sides of the Cold War were utterly terrified of using them. If they hadn't been invented then I suspect there probably would have been direct armed conflict between the Soviets and/or China and/or the West at some point after WW2, so we were probably saved from another horrific conflict by scaring ourselves rigid with these weapons.

Rational, perhaps not, effective - arguably yes.

Do you remember the source where you read he died of leukemia and he believed it was caused by work at Los Alamos?

His ex-wife died of leukemia. On the other hand he died of stomach cancer. (he had 2 cancers).

Sorry I don't remember the source on that. I'm going off of memories of what I read some 20 years ago. As my misremembering the the type of cancer shows, my memory is imperfect.

I only wish it were so easy.

Deleted comment

Big Physics.

That's the funniest thing I've heard all day.

How to Solve It, by G. Polya

http://www.math.utah.edu/~pa/math/polya.html

I think I remember reading Feynman saying that you should always have 15 or 20 problems that you have percolating in the back of your mind at any given time. Then, any time you learn a new trick or technique, apply it to all of those problems. Sometimes the new technique will work really well on one of those problems and you'll look really smart, but all you did was apply a technique from one person to one of your old problems.

I can't find the reference now...

The story that best illustrates Feynman for me is in Surely You're Joking Mr F (I think - it might be the sequel). On a trip to Japan he insists on going to an area where there is 'nothing to see' rather than the standard tour, and of course discovers a unique and unspoiled part of the country. But even if he hadn't, he would have considered it a win - in another of his stories, he talks about going to bars and waiting for something interesting to happen, knowing that much of the time nothing of note would occur. That drive to keep experiencing something new, and the understanding that the drive would often lead to nothing, is inspirational to me, and I think is what led Feyman to so many interesting things in so many fields.

Reminds me of another classic Feynman/Gell-Mann quote, from Surely You're Joking:

Telegdi also sent us a letter, which wasn't exactly scathing, but nevertheless showed he was convinced that our theory was wrong. At the end he wrote, "The F-G (Feynman-Gell-Mann) theory of beta decay is no F-G."

This works, sometimes. The other one that works even better is when explain the problem to your dog.

Here is what I would prefer to do:

  1. Write down the problem. 
  2. Think real hard. 
  3. Program the solution. 
  4. Understand changes caused by the solution. 
  5. Iterate.

I think between 1 and 2 it also helps to talk to other people in your network who might have knowledge / wisdom to share about the problem at hand.

Would totally add that! Real problems never get solved in isolation.

Hmm, I tried it in LISP and did fine until step 4. There were no side effects and it didn't iterate...

This must be later version — while a kid he skipped steps 1 and 3 (I am reffering to "He fixes radios by thinking" in Surely You're Joking).

For people not familiar with Feynman, here is a recent review of one of his story books:

http://goo.gl/5QLaG

Reply on news.ycombinator.com