[lend000]

The article wasn't presented as any sort of proof of mechanism or even authoritative description of what's happening.

An aside: I'll venture a guess that the majority of discoveries in human history started by throwing out Occam's razor. I don't understand why it's such a religion on HN.

[anm89]

The point of occams razor isn't to prove anything or to be correct, it's to point out when speculation is low value.

And this case pushes the extreme of the delta between two options.

[luckyscs]

Speculation is of high value to human discourse and entertainment though.

[tshaddox]

> I'll venture a guess that the majority of discoveries in human history started by throwing out Occam's razor.

What’s an example of this? At first glance it looks very false to me. All the examples I can think of “simpler” theories being replaced by less simple ones occur in the face of new reasons to believe the simpler theory was false. You’re not really throwing out Occam’s razor unless you have two competing theories for the same phenomena and you choose the less simple one, right?

[lend000]

Replacing Bohr's model with Schrodinger's equation is one example. Or Newtonian physics with relativistic physics. These were proven to be better models, of course, but if great physicists applied Occam's razor to everything like some on HN, Einstein/Schrodinger would have never bothered investigating further.

Perhaps the Catholic church invoked Occam's razor when Galileo suggested the orbits of celestial bodies are more complex. Occam's razor is a good way of advocating that the existence of one single god is the default answer we should all accept, because you and I have no evidence for or against anything else and any other explanation for existential questions is going to be more complex.

The point is that bringing up Occam's Razor on every thread adds no value, does not need to be verbalized on a forum of smart people like HN, and frankly seems to be a way to justify intellectual laziness/closed-mindedness (which might be useful in high stakes/low time situations, but not in an 'intellectual' internet forum where people have plenty of free time to make better arguments).

[tshaddox]

Newtonian physics is the easiest example to criticize, because it’s precisely what I said: special relativity was a new model to fix blatant flaws with the previous model. Occam’s razor only concerns comparing two theories which account for the same phenomena. Special relativity was a new model that accounted for more phenomena, so it’s really not surprising (and certainly not a violation of Occam’s razor) if it was less simple.

[devenvdev]

Before we ditched the gods, Occam's razor would bring you to a conclusion that Earth is the center of the universe as 'world created by god' was the explanation with fewer details than 'world with many planets created by god.'

You can apply the same logic here; who said we see the objective truth now? Zen and the Art of Motorcycle Maintenance covers the issue well - church of reason AKA science is the current faith that makes us blind.

[kjellsbells]

That's actually a great point. To create, you need to throw out preconceptions like the razor. But to achieve persistence, simplifying ruthlessly... ie bringing the razor back in, helps. It's a bit like brainstorming: if you shut the idea stream down with the cold water of rationality, you lose what could have been really good ideas.

[twofornone]

People seem to forget that it's a heuristic, not an actual logical inference.

[naasking]

Solomonoff induction provably converges on reproducing any input function by only observing that function's outputs, and it formalizes Occam's razor as a key principle. I don't think it's fair to call it a heuristic anymore.

[logologology]

Occam's razor is formalized in Solomonoff induction but that doesn't mean that it is not still a useful heuristic in Science more generally.

[naasking]

Sure you can use the simpler version heuristically, but the OP's claim is that it is only a heuristic and not a valid logical inference. I'm saying we now have proof that it is logically valid to claim that one should prefer theories with fewer assumptions, all else being equal.