Interview with Freeman Dyson
52-insights.com
52-insights.com
Freeman Dyson is the person responsible for legitimization and adoption of Feynmann diagrams, via proof of equivalence with Schwinger's methods. Feynmann didn't care about why his methods worked, Dyson did. Think a brilliant hacker (Feynmann) who basically points to a working system in a new programming language and says "Look it works. What more do you want? Clearly this is because of my awesome new language." and another engineer (Dyson) trying to get him to document the damn thing, explain how it works in terms of previous programming paradigms, and then discovers failure modes in the process. As usual, the engineer who did the documentation gets a bit of short shrift of the glory.
Now, this of course does not mean that he's right about this issue; but please do do not disregard his honesty or truly tremendous contributions to human knowledge.
On evolution, he says that Darwin got it right "up to a point", and asserts that selection is only an effective mechanism if you have large populations, but that small populations are often key turning points, and for small populations, "random drift" is more important. It's said to have some explanatory power in addressing why there is a variety of e.g. beetle species inhabiting a given ecological niche, instead of a single super-optimized beetle species. He's got a couple of authors he likes on this point.
On climate change, his position here is that since we don't understand how the ice ages started or stopped, then we don't really understand climate. Further, he says if we planted "all the wasteland over the globe" with trees, it'd be enough of a carbon sink to absorb the increased atmospheric content.
I am not really equipped to defend either the scientific mainstream or Dyson's views, but hopefully this is a useful summary.
Are there really two or more beetle species occupying the same ecological niche, though? It would violate the competitive exclusion principle [0], and a quick google search doesn't return any results. (Also, the articles doesn't specifically mention niches.)
I thought the large variety of niches, each of which requires different kinds of specialization, is why there's such a diversity of beetles. (Different niches requiring different types of camouflage, different jaw-shapes, different digestive enzymes, different temperature tolerances, etc.)
It doesn't really make sense to expect a single "super-optimized" beetle, if, for examples, being able to eat food A and food B requires fundamentally different jaw-shapes.
Of course, random mutations in the genome is what makes different phenotypes possible in the first place, but (according to my layman's understanding of biology) this is already acknowledged.
And Dyson's example about a Peacock's feathers is famously explained by sexual selection.
I really like Dyson's talk about living through four revolutions, though. [1] (Space Technology, Nuclear Energy, Genomics, and the Computer Revolution.) He's a super interesting guy.
[0] https://en.wikipedia.org/wiki/Competitive_exclusion_principl...
I work in the intersection of AI and another discipline. Unlike some people (e.g. Elon Musk), he actually knows what he is talking about.
I skipped to where I thought I see something about something "sciency", and I found the section when you skip to
> Just to clarify here for our readers, obviously, you’re poking holes in Darwin’s Theory of Evolution but you’re saying it only tells the story up to a certain point...
I don't understand what he is talking about. I mean, where is that "controversy"? I can only see what's in that article, I don't know what else he may have said or written elsewhere. I only took biology classes on edX the last few years (the most important one was [0]), and nothing he says in that section is "controversial". Genetic drift and small populations are part of the teaching, if you go to edX's course "Nature, in Code: Biology in JavaScript" [1] you even get to simulate just that numerically, population bottlenecks, genetic drift, and more. I recognize everything he said from the courses I took, or from lecture videos on Youtube (by biologists and geneticists, specifically, e.g. this one: [2] - very funny guy too).
Going only by what I see in that article, in that section, I can only assume he is making a mistake in declaring some imagined and made-up "controversy" where none exists.
What am I missing?
To me his statements are like somebody trying to defend their "controversial" point of view that the earth is on a course around the sun, making very defensive statements because they think, for whatever reason, they are saying something incredibly controversial.
[0] https://www.edx.org/course/introduction-biology-secret-life-...
[1] https://www.edx.org/course/nature-code-biology-javascript-ep...
[2] https://www.gresham.ac.uk/professors-and-speakers/professor-...
When Freeman talks I always listen, even when he's wrong he still has a point: don't underestimate him, even at 94.
(*We're still using quite a bit of his QFT formalism... laid down almost 70 years ago now.)
> My comment is not about the person... only about one specific section in the linked article.
I don't see why anything you say should influence my opinion about what I wrote, when that is all I comment on? Do all his accomplishments mean he can say whatever he wants and I can't say anything? I admit I don't understand your response, given that I already limited the scope of my comment myself right from the start.
Yeah, in my book Freeman Dyson can say whatever he wants, he's earned it. Of course I'm not saying he's right, but I'll double check.
But he's a physicist!... well, at some point Biology has to be Physics, what else could it be?
Would you please read? I did not say anything about the guy, only about the article in front of me! And I DO know my biology well enough to know that what was written in that article did not make sense! (Because it's actually taught at beginner level already so how can he claim to have said something "controversial")
He's arguing that maybe genetic drift is more dominant than selection for small populations, and that small populations are not a bug, but a feature. A 30 second google search would have gotten you a Wikipedia page for the Kimura theory: https://en.wikipedia.org/wiki/Neutral_theory_of_molecular_ev...
And yes, it's strange to me that some people dismiss very important scientists, don't have a problem with not even being familiar with their name, on top of being unable to find any relevant information by themselves, at the same time they feel entitled to make objections because they've watched some lectures on the internet. And can't even read themselves because that interview wasn't about no one considering drift before, just what's dominant in evolution.
This uncalled redditization of HN is a waste of everybody's time. BTW, you should think twice before accusing people of using another accounts for I don't know what, because you're just making it up and being extremely unpleasant for no reason. Don't do this here.
I'm not an evolutionary biologist, so this is just a layman's pondering...
Theory of evolution is like theory of how brain works: incomplete.
Consider for example the game of Go board game (https://en.wikipedia.org/wiki/Go_(game)). It has an extremely small set of rules, yet it produces astounding patterns in the way games develop. The Go board game obviously has a creator; humans invented it. But that doesn't mean that whover thought up the rules of Go knew upfront the intricate complexity that we explored throughout the centuries (by playing games).
We gain no further insight in the higher order that emerges by following the rules of Go by knowing the precise circumstances that led to its creation.
When thinking about the universe and life itself, I think this aspect is more important than the unknowable answer about whether there was a creator or not: would knowing the creator necessarily explain most of the complex behaviour we see?
if a random mutation produces a protein that cannot act as a catalyst for a given chemical reaction that reaction won't happen and that can make a hell of a difference to the the organism that had everything else setup depending on that thing to happen, decreasing the odds of survival (which often simply means not even survive for long enough to develop).
There are so many constrains at so many levels, many we know of, many are in plain sight, many are hidden, some we know since long time, some we have yet to discover. There are so many things we don't know, but pretending we are completely oblivious about the rules of the game is a stretch.
The sheer complexity of it all is mind-boggling, but that doesn't mean the only reasonable approach is just to give up. So far every time humanity has thought that we've reached our limit in understanding and that only the existence of a deity can explain the next step, we unveiled the next layer of the puzzle.
Even if a deity does exists, it doesn't appear reasonable to expect that we've finally exhausted all there was to learn and we finally found the truly unexplainable. We no longer have to posit the existence of a deity to explain our diurnal cycle and other everyday miracles, because we now know; we convinced ourselves that what seemed utterly incomprehensible to our ancestors (who we have no reason to assume less intelligent individually than us) is to be attributed to simple celestial mechanics governed that rules that no longer seem that incomprehensible. We moved our questions to the next layer, to the why those rules exist: what did set planets in motion etc; if you want, we never really "answered" the question, we just pushed it one layer further, and in the process we uncovered so many new beautiful layers of our reality.
What does "failed" mean? Evolution is a universal concept with infinite potential implementations. Trying to measure "success" means nothing unless you have some particular goal in mind, and as soon as you add a goal you severely restrict the pool of potential "successful" implementations, which makes success harder to find. Even then, you can look to processes like stochastic gradient descent as a "successful" use. It goes a lot deeper than that but as someone who has spent many years studying evolutionary biology, computer science, information theory, cognitive science, socioeconomics, and complex adaptive systems I have no idea where your comment is coming from so I won't elaborate here.
The only ever work for very specific things. On a larger scale there again is no design anywhere, see for example the economic principles, the "invisible hand". As soon as humans tried to "design" such larger systems it all fell apart (communism - in which I grew up).
No single entity can manage the complexity, unless that entity is the universe ("it models itself") - at which point we have a useless tautology.
Also, it's funny you mention communism, because no "design" was planned there originally, and the economy was supposed to just "emerge" from fundamental principles and real life conditions.
My understanding of biological evolution is that seemingly random mutations take place. This is like having a bunch of monkeys with typewriters, and hoping they produce a Shakespeare's sonnet. Yes, it's possible, but highly unlikely.
I see several explanations of why biological evolution is successful:
1. trillions of trillions of planets launched evolutionary processes during the last 14B years, and ours just happened to be extremely lucky one, where intelligent life has appeared (through random mutations) against all odds.
2. intelligent life has emerged through random mutations on some other lucky planet, and those intelligent beings planted some kinds of seeds capable of developing into intelligent life eventually on other planets.
3. life on Earth was not planted like in the second scenario, but the random mutations are not really random, and are guided by something we don't understand (yet). Living in a simulation, all-mighty God, or some not yet developed branch of complexity theory, would belong here.
Note that the second scenario is extremely more likely than the first one, and the third one imo is more likely than the second one. In any case, it's not just "controversy". It's a serious lack of understanding of biological evolution (either on my part, or in general), similar to how there's a serious lack of understanding of how a brain works.
That's a sharp observation. Sharper yet what it leaves unsaid about the present.
You have a long history of breaking the HN guidelines and we've warned you and asked you to stop repeatedly. If you can't or won't stop, we'll eventually ban you, so could you please re-read https://news.ycombinator.com/newsguidelines.html and use this site as intended?
doesn't make for an adequately clickbaity headline though
But, more back on topic, it seems to me that despite the headline picked suggesting something very controversial, nothing in Dyson's intereview really was.
re: evolution and natural selection vs. random drift, it's not like this are opposing concepts. They're mutually complementary.
re: climate, he's not wrong about humans not understanding how big climate changes (ie: ice ages) worked or what they were caused by. It's not a huge step to suggest that that inability to understand or create predictive models of past climate records suggests current extrapolations may not be very predictive.
No reasons to get everyone riled up like this HN thread has been. Read the article and you'll see it's all very nuanced and not him just saying "EVOLUTION IS WRONG." or "CLIMATE CHANGE ISN'T HAPPENING.".