Please no more comments to the extent of "i can define a much larger number in only 1 bit". What makes my blog post (hopefully) interesting is that I consider tiny programs for computing huge numbers in non-cheating languages, that are not specifically equipped for doing so.
> that headline doesn't even include "program" (or "compute").
Neither does Scott's article titled "Who Can Name the Bigger Number?" [1]
The title is just a way to invite the reader to find out why the answer isn't simply 2^64-1.
BLC can output any literal 60 bit string x as the 64-bit (delimited) program 0010 x, so in that sense it would be some 61 bit number. But if ask about just lambda calculus terms without the binary input, then I think it would be some small number of at most 10 bits. BBλ looks at the normal form size so it cannot even reach numbers 0,1,2,3, and 5.
like the rest of simple solutions to complex problems suggested by otherwise (seemingly) intelligent people on HN, it simply wouldn't work. do you really not see why?
Are you talking about that people could just click on the link then not actually read it? The thing is that clicking on the link then closing is serves as both a slight barrier to entry targeted at the people who comment without reading and also a reminder that there is an actual article to comment on. It's not going to fix discourse but the theory behind it is to be a slight nudge to get people who don't click to at least consider reading the article (or just not comment) while being an invisible change to people who actually read the article. It's not meant to be a magic fix, it's just some something I would want (though I'm biased since I click on links so there's no downside to me).
Question: but how many different numbers can you fit in 64 bits using your encoding (sorry I understand the general approach but I have no idea how that fast hierarchy works). I guess it's still 2^64 different numbers?
So basically you have a very low density of representable numbers (2^64 / w218), I wonder how quickly it grows as you use more and more 1-bits, and is there even a correlation between the bit pattern and the corresponding number value?
Surely you've been on HN long enough to know people just read the headline. Not that it would stop all sniping, but that headline doesn't even include "program" (or "compute").