(1) is equivalent to "typedef long* $(void*, ...)". '$' is the name of the defined function type.
it's obfuscated as follows:
- typedef is a specifier and can go on either side of the type, same as int const/const int
- [[ ]] is an empty list of attributes
- "void*($)" is an argument named '$' of type void*, but parameter names are ignored in declarations.
- you can put parens in declarations since sometimes you need to group prefixes and postfixes in a different order
- dollar signs in names are allowed in many c compilers (and some other c-like languages)
- "..." are variadic arguments. puts doesn't have those, but the first argument uses the same register either way so it happens to work out in this case
(4) declares a function named '$' of type '$', shadowing the type, and links it to the assembly label "puts". calling $ calls puts.
normally it looks like `int call_foo(void) asm("foo");`
(5) is parsed as: "label $: address-of-label $ and function $ called on str. the address of a label is truthy, so the function also gets called.
unary && looks for the argument in the label namespace, but other operators don't, so the two dollars refer to different things.
that's about it. if you're familiar with the c standard and some common gnu extentions, you can deconstruct the examples into their basic elements easily enough
What do you mean? It's just simple beginner's examples in C. Should be easy enough to understand, C is famously very simple ('portable assembler' as they say!).
> In our experience, C has proven to be a pleasant, expressive and versatile language for a wide variety of programs. It is easy to learn, and it wears well as one's experience with it grows.
Absolutely. And I've done it to quite a few places.
Like it's no secret it's expensive as shit to live lately. If you're just giving away your software/knowledge/ideas, and I feel I gain value from them, I'll happily throw some cash in your hat.
If Dave Arnold says he really doubts it, it's a pretty safe bet he's read like a dozen papers related to the prediction and is basing it on something. He's like the Bunnie Huang of cooking.
In the article the only explanation he gives is that it doesn't make sense to him, doesn't mention any papers or anything at all. But I'm pretty certain that he's wrong and it works. The difference in gas if I've been eating beans recently vs if I haven't eaten them in a couple of months is not just "I feel like maybe I get a little bit less gassy maybe" it's going from "two dozen farts at least, guaranteed" vs "one or two, if any at all" it's a night and day difference and if there even was a paper that says the contrary rather than change my mind I'd just assume that there must be something wrong with how the study was made. Of course, I'm just a sample of one and I haven't done any study either, so I don't mean to imply that there might not be other factors or that it may not work to the same degree for everybody, only that I'm pretty sure that dismissing it is wrong because I know at least one counter example.
More anecdata here. For various health reasons, about six months ago I started eating beans every day for lunch. At first it was… intense. But within a few weeks, the gas entirely disappeared. I now have no issues whatsoever.
FWIW, some beans are easier on my system than others. Kidney beans (sadly), my favorite, can be murder if you don’t cook them thoroughly enough. Lentils, on the other hand, seem to be pretty gentle. They also soak up whatever flavor they are cooked in so they are a great “starter bean” if you want to eat a more vegetarian diet.
Yeah, given the number of vegans / vegetarians I know who have gone through this I find it completely self obvious. If I went and ate meat today, I would almost certainly barf… But I wouldn’t spend ages trying to find a way to cook meat to make me not barf, I would just slowly introduce it back into my system little bits at a time until the adverse effects go away.
> Are there leaderboards that you follow or trust?
Not for OCR.
Regardless of how much some people complain about them, I really do appreciate the effort Artificial Analysis puts into consistently running standardized benchmarks for LLMs, rather than just aggregating unverified claims from the AI labs.
I don't think LMArena is that amazing at this point in time, but at least they provide error bars on the ELO and give models the same rank number when they're overlapping.
> Also, do you have preferred OCR models in your experience?
It's a subject I'm interested in, but I don't have enough experience to really put out strong opinions on specific models.
Having ripcorded out after realizing the author was trying to prove that water was wet, I'll assume that it's "normalized entropy", in a range of 0-1, indicative of the distribution across the space.
