You are right. 2^n would suffice

I read the problem statement incorrectly. I thought we need to open all doors as well. That’s why 3^n

Ah good point. On the other hand, making an intcode reader would be useful, I’d guess. Intcode->Racket Expander is pretty much what I’d write anyway.

I got the impression from comments here and on Reddit that intcode is a big part of everythng.

One issue with intcode programs are that they are potentially self modifying. I haven’t checked any of the test cases I got, but that is what prevented me from doing anything quite that elaborate.

Making an assembler could be useful though. I ended up coding a couple test cases using my gray matter assembler and comments.

Yep. 2^n * n is enough to represent the search space.

I burnt myself out trying to finish in the first couple hours the first few days.

Next year we’ll all be doing it in proto-rhombus :stuck_out_tongue:

I’m still hoping to finish … just, later

Every other day, we have a problem that involves using intcode. So, it is important but I don’t see it as a bottleneck to solve the problems (i.e simulating it is just fine).

And SamPhillips is right, they can modify them selves what makes difficult to compile.

I am not sure whether my intcode “interpreter” works correctly, because I have not finished day 13 part 2.:disappointed:

I got points for the first time today!