@Connor natural rights in the sugar beets?
Sparse is a great way to extend the C type system; sure, it's no ML type calculus, but it gives us a lot of the things we need (e.g. __user pointers, endian annotations).
@WomanCorn Some of us still hack on kernels, where your program *is* the runtime.
As https://www.usenix.org/system/files/1311_05-08_mickens.pdf puts it,
"if you find yourself drinking a martini and writing programs in garbage-collected, object-oriented Esperanto, be aware that the only reason that the Esperanto runtime works is because there are systems people who have exchanged any hope of losing their virginity for the exciting opportunity to think about hex numbers"
@amir see this is why I was only ever rat-adj. I have (and loudly voice) *many* object-level opinions, and can't derive *any* of them directly from theoretical arguments about the Solomonoff prior & what kind of aliens control it, so I was always Insufficiently Rational to consider myself a rat. Sure I could correctly calibrate my confidence levels, and never be more than 58% sure of anything _interesting_, but where's the fun in that?
(Don't get me wrong, I love the ratsphere. But still.)
@amir "Eat my shorts." As usual, the Simpsons predicted the future 😛
@WomanCorn schools are clearly projecting. (Different kind of abuse, but still.)
@PstafarianPrice "R17 is not a fixed velocity, but it is clearly far too fast."
@lolitsnotreal_ Go insane faster and save your sanity for later? (The Ford Prefect strategy.)
@amir although I do wish that these internet weirdos would stop referring to themselves as "autists". They're like 5% away from baseline NTs at best, as an Actually Autistic guy I feel like they're diluting my brand :P
@amir it's almost like, reducing transaction costs through information technology makes it harder for closed institutions to capture value from wide, chaotic, yeasty free markets. I'm not in on GME myself (just don't have that kind of risk appetite) but *loving* watching it.
@amir I tried to Fermi estimate around that here https://twitter.com/ecree429/status/1348608399999709185
Conclusion I came to was that lockdowns are a terrible idea. Granted, I already thought that before doing the calculation, but still...
@ayegill The beginning of wisdom about entropy is understanding that it's defined only in relation to the set of macrovariables under consideration.
@ayegill The beginning of wisdom about programming is that code is primarily a communication to other* programmers, rather than to the compiler.
* including your future self, who will _not_ remember how that 'clever trick' you just came up with works
@iamnoah when you're a really good engineer, subtle functional judgements present themselves to you aesthetically (code that looks ugly to you turns out to harbour bugs). Linus Torvalds calls it "taste".
The problem, of course, comes when you *think* you're that good but aren't. Only reliable way I know for determining if someone has taste is to ask someone else whom you know has it & is familiar with their output.
@WomanCorn Unfortunately AIUI the only fully robust way to handle this is Bayes on a Solomonoff prior, which is so uncomputable it's not even funny.
But I think you can get somewhere by not fixating on a _single_ model; instead, have an ensemble of models, one of which is the maxentropic "Something I have not thought of" model, and weighting them according to their predictive performance over time. (Make sure to *actually* predict and not just retrodict.)
@WomanCorn My understanding of it is that identity isn't transitive. I.e. all my past selves are me, and all my fork's past selves are him, but even though those sets overlap my fork isn't me.
But quantum suicide does still freak me out. Especially when invoked to explain nuclear near-misses like the one Stanislav Petrov averted.
@cuttlefish So, uh, are you trying to communicate, or are your toots purely performative? I want to assume good faith, but you're making that difficult.