A "computer" is just a function from {0,1}^N -> {0,1}^M
A computer is an abstract mathematical description (eg., like "prime") of a certain mathematical object, a function.
A computer is a way of specifying a discrete function (ie., one which maps a finite number of bits to a finite number of bits), in terms of a sequence of mathematical transitions.
It's an "algorithmic" way of specifying the domain and codomain of a discrete function.
Electrical digital computers aren't actually computers in this sense, and are extremely aproximately described by them. Inasmuch as the shape of the earth is aproximately "spherical".
In any case, pretty much all of physics does not use discrete functions (indeed, I can't think of a single case). In every way physics describes reality, ie., parameterised on space and time, functions are continuous.
They map an infinite amount of spatio-temporal information to an infinite amount of spatio-temporal information.
And there is yet no reason whatsoever, other than the AI PR machine, to suppose that all of physics is wrong in this regard, and the universe is describable by anything else.
This is relevant here, since the problem that cannot be represented to the machine uses ordinary equations of physics, none of which are computable.
"{0,1}": the set containing the values 0 and 1
"{0,1}^N": a discrete n-dimensional space, where the possible values in each dimension are 0 or 1
So they're saying a computer takes a length N binary sequence input and produces a length M binary output.
(As for "what level is this", I didn't cover any of this in my double A-level[0] in maths/further maths, but I am covering it in brilliant.org and some popular maths books, so my best guess is it's first or second year degree level?)
Here, for example, the mathematician has to imagine a scenario to describe with mathematics (two couter-flow fluids, etc.). The notation gains its meaning from this imagined scenario.
Rules for manipulating symbols are therefore insufficient. The proof has to follow from the scenario, which the machine is unable to represent.
If you believe that a computer will eventually be able to accurately simulate a human brain, you might as well give up right now.
Since if a computer with all its constraints is able to simulate a human (brain), but cannot do this, then a human can't do it either.
Conversely if a human can do this but a computer can't, then a computer can never simulate a human.
Don't tell this to a software developer working on AI. They might quit their job and become a baker instead.
I don't think you would have any effect on a mathematician, since they would already be acutely aware some things provably cannot be done.
Computer, as defined abstractly, is just any abstract function from {0,1}^N->{0,1}^M.
Any realisation of that, eg., by providing each {0,1} as an electrical switch, realises physical properties associated with electrical switches only.
The reason that electrical computers are useful has vastly more to do with the electrical part than the computer part. The "computational properties" of the electrical devices we call computers are relatively trivial.
But in any case, no system in virute of being an implementation of a discete function thereby acquires physical properties. A woodern "comptuer" is useless precisely because you can't play video games on it.
Likewise, even if the brain can be described by a discrete function -- (which is so implausible as to be a bit mad and certainly purely an act of faith) -- then it still requires the relevant physical properties to implement. These properties are extremely unlikely to be those of electrical switches.
The "computational work" done by biochemical signalling alone should probably be regarded as "infinite", saying much about the limitations of discrete conceptions of information.
I would say that it's impossible in a finite universe to calculate infinite values. But it's quite possible to define and manipulate relationships that involve infinity.
You can't calculate the sum of all integers - but you can relate it to the sum of all real numbers. Or compare the sum of all integers lager than one hundred, to the sum of all integers larger than ninety.
They don't explain this very well, I think.
Take the ordinary differential equation x'(t) = x^2(t), with initial condition x(0)=1. It has the solution 1/(1-t) which blows up to infinity when t tends to 1.
If you try to solve it numerically, using, let's say Euler's method, then this is how you go about it. You pick a step size, let's say 0.1. And iterate this way: you know x(0) = 1, and you also know its derivative x'(0) = x^2(0) = 1. You assume x follows a straight line, so you get x(0.1) = 1 + 0.1 = 1.1. At the next step you add more because x'(0.1) is now 1.21, so x(0.2) = 1.1 + 0.121 = 1.221. You keep going like that.
The numbers will go bigger and bigger, but they will never be infinite. Of course, floating numbers with double precision overflow around 10^308, but if you use a multiple precision library you'll be able to keep going forever and ever.
If you make the time step smaller, the solution will be closer to the actual solution, but still, the algorithm will produce finite values at all times (until it hits overflow).
I believe this article is talking about numerical methods, which are always bound to finite values, because of finite memory.
I think the author didn't understand the concepts of precision, range, or of arbitrary precision calculations, and so mades bullshit proclamations without sufficient understanding of the fundamentals.
PS: I notice many people I work with (in MAANG) lack an understanding of Computer Science, aren't formally educated, and lack curiosity. It's discouraging because they also lack concern for code craftsmanship and don't know how to use a debugger. This and they're paid $350-450k / year. If you had to recommend a computer science postgraduate program, which would you choose?
Why not?
Is it impossible to calculate infinite values in general? I suspect not, My understanding is that a lot of calculus is in fact on how to calculate infinite values.
And a computer is a universal machine, this means that while it can not calculate everything, it can calculate anything that is calculable.