Let's consider a protein of 100 amino acids; although most proteins are in the 200-500 range (IIRC) it makes the maths simpler. Even for such a short sequence, the probability of coming up with one specific sequence at random is incredibly small - 20 to the power of 100 or 1.267651e+130.
The simple answer to this apparent puzzle ('how did proteins evolve from random mutations') is of course that it did not happen like this. Not at all like monkeys typing out scripts and a human checking if they have produced a Shakespeare play yet.
In fact, there are multiple solutions to the same problem - in the analogy, this would be like multiple different versions of the same play being just as good as the other. Even better for evolution, you can - in a sense - put on a version of a play that is pretty bad, but just good enough to put in front of an audience. In other words, some enzymatic functionality can be better than none.
In this way, an evolutionary system can build up complexity by making parts that are _good enough_ and are roughly equivalent to other parts that can be replaced gradually. Combine this with the massive parallelism and the scale of the molecular world, and a billion years can accomplish a lot.
Abiogenesis really only need to explain how the first life arose. It could have been really simple since it got out competed by more sophisticated descendant that took over planet and left no trace of original life. We don't know how it happened, if required lots of random chance or if there was a replication step that made if go faster. We know that evolution can make complicated stuff over time.
Where would you get an intuition about how much time a billion years is? Or for what is "mind boggingly complex" and what not? How many eons have you lived? People who claim it is possible have better arguments than "to me it seems". Do you have a number for "this would seem to me as enough time"?
Do you also struggle with the idea of a poll being close to the truth when done on a sample of 1000 persons, out of tens or hundreds of millions? Math and science are often not intuitive. But they work.
I don't really see biology as being a "little" pattern. To me, looking at the jump from replicating molecules in chemical soup to single cell organisms is like looking at the jump from an abacus to an Intel i9 processor.
But in reality, the universe is full of patterns. The laws of physics are not random. Water behaves in very specific ways. Our planet revolves around the Sun in 365 day intervals. The planet rotates every 24 hours. Even out of these two interlocking patterns, you can see how energy delivery to the surface of the planet could result in mechanical and chemical cycles that may result in non-random mineral formation, concentration gradients, and thermal dynamics.
Scientists argue that abiogenesis is a result of inherently chaotic but non-random processes, not a hypothetical probabilistic scenario like monkeys with typewriters.
Call it 75 years (1942 - before electronic computers to 2017 - first i9) which is about 39M minutes. Assume human innovation can work on an minutely cycle[0] and you've got 39M iterations from "no electronic computers" to "Intel i9".
I'd put decent money on there bring several orders of magnitude more chemical reactions than that per second in the primordial soup (good conditions for fast reactions and there was A LOT of it.) And it had about half a billion years (fifteen quadrillion, seven hundred seventy-eight trillion, eight hundred billion seconds) to cook before things really started happening.
Statistically, it was likely.
[0] Obviously optimistic but giving parallelism a chance.
Of course, it's a few billion years scaled across >1 life-friendly planets (since it only has to happen on 1 of them), but still just seems like not enough monkeys and typewriters in the given timeframe to produce something as astonishingly complex as life, let alone Shakespeare.