MarkusQ

31 ms·

751.
by MarkusQ 2y ago
One trick that works surprisingly often when you don't understand the root cause: scale the number of batches, rather than the batch size (think merge sort or map/reduce). If 10% of your batches flake out, you rerun them and expec
752.
by MarkusQ 2y ago
It would be even harder to read if the author didn't do that. We are being invited to consider a premise (i.e. presume that it is true) and explore the consequences. At any point, we can choose to disagree with this premise, but at le
753.
by MarkusQ 2y ago
Ah, the old My Dear Aunt Sally vs. My Dear Aunt Sally debate. :)
754.
by MarkusQ 2y ago
Source? I get Starship 34,000,000 kg + 12,000,000 kg vs 747 ~200,000 liters ≅ 150,000 kg, or about 1/300 th of what Starship holds.
755.
by MarkusQ 2y ago
The key word here being hypothetical. Getting to a significantly more advanced model will require significantly more work, and yet again we don't really know which direction to go in. Every time we crest a hill and see the ocean in th
756.
by MarkusQ 2y ago
> it shouldn’t be possible to derive the formulas for chemical or nuclear weapons. This information is widely available in textbooks, so...what exactly is the problem? If it's just a lossy-compression parrot "AI" like Cha
757.
by MarkusQ 2y ago
Ships _spread out_ over 100,000km will literally cover interstellar distances as a swarm by _traveling to another star_. That's the whole point. A bunch of insects can swarm a thousand km from the time it forms to when it breaks up,
758.
by MarkusQ 2y ago
It doesn't appear to be heavily dependent on the distribution. It mostly depends on how large your buffer is compared to the number of unique items in the list. If the buffer is the same size or larger, it would be exact. If it&#x27
759.
by MarkusQ 2y ago
You're the one who brought up the "crack wars" narrative; the article doesn't mention crack at all. It does, however, state (contrary to your claim) that murders are _down_ (by 34%). If you are taking a narrower view of
760.
by MarkusQ 3y ago
> > I just got into the habit of saying “yes” to pretty much anything. > Isn't this also just a form of survivorship bias? You wouldn't be posting this from prison for example. Note that "pretty much everything"
761.
by MarkusQ 3y ago
Not really relevant though. The airship to orbit concept aims to get both high and fast, whereas the XKCD/What-if you linked hinges on the fact that just getting high is not sufficient. (This still leaves open the question of whether
762.
by MarkusQ 3y ago
> These models are obviously capable of acts of intelligence. Except they aren't. They are capable of language and pattern manipulation, and really good at it. But if you concoct a novel problem that isn't found anywhere on
763.
by MarkusQ 3y ago
Two words for you: Jon Skeet. https://stackoverflow.blog/2018/01/15/thanks-million-jon-ske...
764.
by MarkusQ 3y ago
I disagree. It clearly tells us _what_ is different, it just doesn't tell us _how_ it differs from case to case. So it's not a full explanation (which would, as you note, require a model of how it differs and, optimally, _why_)
765.
by MarkusQ 3y ago
You sure like to make assumptions, don't you? :)
766.
by MarkusQ 3y ago
> Just saying it's so doesn't make it so. Does that apply to you as well?
767.
by MarkusQ 3y ago
This is just wrong. First of all, you're rejecting teleological anthropomorphizing (saying it's weird to act as if "the weights have some goal") but then in the very next line you talk about the optimizer making "a
768.
by MarkusQ 3y ago
Well, IIRC, he said there was no economic or technical advantage to reusing rockets, and predicted the demise of SpaceX. That Tesla would never produce a semi-truck, then when it did he significantly misrepresented the details so he could
769.
by MarkusQ 3y ago
Nah, it's the fish oil that does it.
770.
by MarkusQ 3y ago
Then what's the point? It would take effort to develop these "features", the only benefit of which would be predatory. For example: > Bambu’s cloud slicer is a massive lost opportunity for Prusa. Bambu's cloud slice
771.
by MarkusQ 3y ago
So on the one hand you're objecting to "extrapolation" and on the other you're claiming that hypothesizing feedback mechanisms we don't know how to model, that have not been observed in humans, and so far as I know
772.
by MarkusQ 3y ago
Not really. They are basing it on extrapolation of our understanding of humans, not some abstract "non-living systems"; specifically, our present understanding is that as people age (past say forty or so) their chance of dying st
773.
by MarkusQ 3y ago
This looks like a really good use case for LLMs. When you're small and starting out in person, record everything the dev team does -- meetings, whiteboads, pairing sessions, etc. (yeah, creepy, but I'm in brainstorming mode) and
774.
by MarkusQ 3y ago
Why? There are lots of examples of people who knowingly do things to aid and abet a primary criminal offense getting charged as accessories. Fences, bookies, pimps, pushers, shills, etc. This isn't a new pattern, and certainly would
775.
by MarkusQ 3y ago
> I remember reading once the idea that cats can remember the future. Uh, that won't be discovered until 2057. You aren't by any chance a cat, are you?
776.
by MarkusQ 3y ago
Mars's atmosphere has nothing to do with its color. It's red because the dirt/rocks are red. The atmosphere is so thin that, except when it kicks up dust, it has little to no effect on the planet appearance (in visible ligh
777.
by MarkusQ 3y ago
Why? Do they seem shiftless?
778.
by MarkusQ 3y ago
I too have an answer, that isn't one that the author of the blog would like. If you can't explain how you know something, the simplest hypothesis is that you don't, in fact, know it after all. If you can't explain why y
779.
by MarkusQ 3y ago
Seems clear to me: "change the system so that publications in big-name journals are not important for career advancement." There are many fields that don't have this requirement, or anything like it, and the suggestion is th
780.
by MarkusQ 3y ago
> I began to think that the reason I found them > endearing and not evil was that while many had > made quite a bit of money, almost none had amassed > significant power. Unlike the Elon Musks and Jeff > Bezoses of the world,

More ›