@gershbrain Almost there. Just as the word 'create' glosses over too many things, the word 'generalize' does that too. What is need is an idea that straddles between the two.
@gershbrain Your argument against create employs a degenerate notion of create (i.e. brute force copying)(degenerate in the math and physics sense). The notion of creation is that there is some competence (however developed) that leads to the replication of the features of the target system.
@gershbrain But in this word create, is an entire logic of constructivism. An intuitionistic logic to more precise. The cognitive dissonance in the quote is that Feynman was a theoretical physicist.
@gershbrain So what Feynman means by create, he means something that is closer to proof (proof in the physics sense). Feynman of course is the first to tell you that you are the easiest person to fool.
@gershbrain But what is being proved here? In Feynman's case, a hypothesis of how the world works. The notion of generalization banks of the creation of a hypothesis that is tested to be true in different reference frames.
@gershbrain A flat-earth theory is one that is true in a few reference frames but false in many other reference frames. We can say that this generalization is useful for a narrow context. Also being aware of the scope of this generalization.
@gershbrain The notion of generalization therefore is dependent on context. Unfortunately, the notion of understanding transcends context. To understand gravity, we may use Newton's laws which is good enough but know that in certain contexts (i.e. GPS) we need General Relativity.
@gershbrain Generalization is to hypothesis making as Creation is to hypothesis testing. To Understand is to know the scope of a hypothesis.
The interview with CNBC of @chamath is a much watch because there is so much insight hidden underneath.
There are many things that caught my attention. The one thing was the abstraction that hedge fund strategies are all 'momentum' plays. What it seems to imply is that the marshaling of resources at an opportune time drives the future behavior of a stock.
From basic physics, we know that momentum is mass times velocity. So any 'momentum' tactic employs the variation of mass, velocity or both. Wrt stocks, mass is money and velocity is speed of trade.
Constructivism is perhaps the most important idea that will shape the future of humanity. The ills of society are a consequence of the willful ignorance of constructivism.
There are two important definitions of constructivism, one comes from mathematics and the other from psychology. The mathematical definition leads rejects the law of excluded middle. It's relevant in understanding causation.
The psychological definition: "Humans actively construct their own knowledge, and that reality is determined by our experiences as a learner."
Lazy Twitter: Is the phrase 'symmetry breaking' unintuitive? Just asking because I like using the phrase.
Yes, I agree that it's a difficult phrase to parse. That's because you have to understand what symmetry is meant for physicists and how it is used to analyze a system.
Then once you've understand what symmetry means, you introduce a notion that is not the opposite of symmetry (i.e. antisymmetry) but rather a verb that says the original symmetry transitions to a state of non-symmetry. The process of breaking is what interests physicists!
Human intuition sees reality through a deceptive lens. A lens that is biased to seek objects rather than noticing processes. Modern language shares this noun centric bias.
The notion that species were fixed and never changing was ubiquitous in Darwin's time. Darwin broke this deceptive symmetry in arguing that all life was connected through a process we now know as evolution.
During Einstein's time matter was thought to also be fixed and unchanging. Einstein broke this deceptive symmetry that matter itself had a continual exchange with energy. Hence he formulated his famous relativistic equation E=mc^2.
Does artificial anthropocentric intelligence lead to superintelligence?
When I use the term Artificial General Intelligence, my meaning of 'General' comes from the psychology definition of the G-factor that is tested in general intelligence tests.
It is an anthropocentric measure. The question that hasn't been explored in depth is whether a "human-complete" synthetic intelligence leads to a superintelligence. The prevailing assumption is that this expected.
I prefer it to mean, human-complete intelligence. This of course doesn't mean super-intelligence. It just means it's a kind of artificial intelligence that has the same capability as humans.
But does an AI with human-level cognitive capabilities very quickly evolve into superhuman AI that can solve everything? I will be wary about statements about 'solving everything'. The universe is open-ended and thus, it is impossible for something to solve every problem.