> It makes two characters that look identical to the eye look as two completely different tokens internally in the network. A smiling emoji looks like a weird token, not an... actual smiling face, pixels and all
This goes against my limited understanding of how LLMs work — and computers generally for that matter. Isn’t that rendering of a smiling emoji still just a series of bits that need to be interpreted as a smiley face? The similar looking characters point makes more sense to me though assuming it’s something along the lines of recognizing that “S” and “$” are roughly the same thing except for the line down the middle. Still that seems like something that doesn’t come up much and is probably covered by observations made in the training corpus.
All that said, Karpathy knows way more than I will ever know on the subject, and I’m only posting my uninformed take here in hopes somebody will correct me in a way I understand.
You’re reading it backwards. He is not praising that behavior, he is complaining about it. He is saying that bots _should_ parse smiling face emoji’s as smiling face emoji’s, but they don’t do that currently because as text they get passed as gross unicode that has a lot of ambiguity and just happens to ultimately get rendered as a face to end users.
And a FOSS re-implementation of which is still alive and in active development. The project recently released beta 5 -- it's very nearly ready for version 1.0 and now runs GNOME Epiphany, Firefox, LibreOffice and more.
That you felt the need to explain these makes me feel really, really old. (And you did need to explain it; it's practically Elder Lore at this point.)
It's like explaining to young people that we used to have to pay per message for SMS, or that only one person could be on the Internet at a time at your house.
Now if you'll excuse me, I see some suspicious looking clouds I need to go yell at.
It's worth to remember that C "happened" to stick around because of UNIX, but it's been just another iteration of what started as BCPL; with Go coming from the same people who made Plan 9 (the spiritual successor to Research UNIX) and Alef/Limbo. These guys are equally interested in pushing both PL/OS research and practice.
(I also have no doubt that just like Go fixed C's type declarations, in another 20-30 years, Go's successor will finally fix error handling.)
I think a proper type system that doesn't special-cases arrays are better, e.g. Array<String>. Pointers may also be part of this uniform structure, e.g. Array<Ref<String>>, and there is zero question on how to parse it even if you are not familiar.
that's the beauty of type after name, it's all left to right, even in special-cases of array, references, and the most confusing-contributor to cdecl: functions `func(int)[5]*string`: perfectly clear, just read left to right.
But you can't know how many parameters do a user-defined "node" have. I don't see how it can work for generic data structures without a specific syntax for "the children of this node". Go also uses [] for generics, doesn't it?
At that point I find a uniform system (without special-casing arrays and pointers - they are also just types) would be simpler.
true; but at the same time, any sufficiently complex type is begging to be type-aliased into something more legible/sensible in the first place ... and this is also true for "right-typed" systems.
I've seen julia types unnecessarily fill a whole terminal for otherwise simple operations before, and I have to say I wasn't too impressed...
> America/Nuuk does daylight savings at -01:00 (yes, with a negative)
Somewhat related: Europe/Dublin has a negative DST offset. Irish DST runs through the European winter (i.e. the opposite of the other European timezones).
I think you misread that. America/Nuuk doesn't have reverse DST (which is easily solved by just switching DST and non-DST around). It starts DST at a negative offset because the offset is defined as relative to the previous day.
The music at many concerts and nightclubs is unfortunately loud enough to cause permanent hearing loss.
Ear plugs reduce the volume to a level where you can still hear the music, but the risk of long term damage is reduced. (You can get "musician's ear plugs" which attenuate all frequencies equally, so they don't make the music sound weird.)
There's a big difference in seriousness between being able to show whether they were likely doing something they shouldn't be doing for the sake of making more money on the fishing, vs. if they were doing what they were doing with the intent of causing damage, though. But of course, that difficulty is also exactly why it'd also be a great way for an adversary to damage your cables.
given the use cases re: satellites on the island, and the continuing cold-yet-increasing tensions with Russia, it's absolutely reasonable to make that assumption
maybe they'll try to blame this on Ukraine like they tried with the undersea pipelines that got blown up, too
It's reasonable to assume. But it's wildly insufficient for either a prosecution or diplomatic consequences, especially given that all 3 also would apply to a trawler just intent on profiting from trawling.
I agree the odds are reasonably high Russia arranged this on purpose. And that's fine for the purposes of a discussion that doesn't have consequences, but not for much more.
Fees such as trading costs are a percentage of trading volume.
Therefore, the more money you are managing, the higher your trading costs. (i.e those costs are "fixed" but its a "fixed percentage" rather than a static number.)
In that case, how can the rate be fixed or does that mean that the hedge fund limits its trade turn-over. In other words, if your trading fees are 0.2% and your trading volume is 10 times the capital raised, you already burned through your management fee.
Endlessh periodically sends data so the read timeout won't trigger. Specifically, it draws out the crypto negotiation stage indefinitely by exploiting a feature of the SSH protocol.
(Of course, the bot author could detect that behaviour too.)
For those wondering how they transfer the generated energy from the tidal kite to the shore:
> The turbine shaft turns the [onboard] generator which outputs electricity to the grid via a power cable in the tether and a seabed umbilical to the shore.
