shiandow

5 ms·

91.
by shiandow 6mo ago
Same difference really. Though some seem to go out of their way to make RSS impossible.
92.
by shiandow 6mo ago
My experience may be a bit biased by having to listen to children attempting to learn the instrument.
93.
by shiandow 6mo ago
Put differently, it takes years of practice to get a decent sound out of a real violin. If your model doesn't sound like someone's strangling a cat then it's probably not realistic.
94.
by shiandow 6mo ago
Ah yes. The trains everywhere approach to self driving cars.
95.
by shiandow 6mo ago
Okay? Let's keep it that way then I suppose.
96.
by shiandow 6mo ago
For a company that puts DO NOT FUCKING GUESS in their instructions they made a heck of a lot of assumptions - assume tokens are scoped (despite this apparently not even being an existing feature?) - assume an LLM didn't have access - a
97.
by shiandow 6mo ago
What makes this framing especially interesting is that it suddenly makes perfect sense to just lay down the gun. Until you remember the millions of children in the exact same scenario.
98.
by shiandow 6mo ago
I'm not saying none of them have any understanding, I'm saying the ones who understand would write better proofs.
99.
by shiandow 6mo ago
Depends. I reckon a proof by an amateur would either be worthless because it demonstrates no understanding whatsoever or significantly better because they actually understand the proof. LLM produced texts are often in a weird area where the
100.
by shiandow 6mo ago
That's a kind of weak criticism. What functions are considered elementary was always going to be arbitrary, picking the set you can generate from exp, log, and some complex algebra is not the worst choice. If nothing else you could sol
101.
by shiandow 6mo ago
These seem to be different tests? One has 6 tasks the other has 30.
102.
by shiandow 6mo ago
Bit of a strange criticism for someone who historically mostly speaks Latin.
103.
by shiandow 7mo ago
It was meant as an example, not a productivity tip ;-) Anyway the real point is that it's just easier to use something if you don't need constant visual feedback. Being able to use something blind is more than just an accessibilit
104.
by shiandow 7mo ago
What I find must puzzling is that everyone seems to just be violating basic rules that had been in place for ages. Things like: - If you can't respond to a UI event wait until you can - Menus should be tree structures - Pressing alt sh
105.
by shiandow 7mo ago
Didn't they kill those?
106.
by shiandow 7mo ago
I wonder who decided to use a step function for the speed accuracy plot. They must have thought the convex hull would be wrong because you can't really make linear combinations of algorithms (you could, but you'd have to use time
107.
by shiandow 7mo ago
> our entropy calculation will now have to use the posterior means Now hang on for a bit, you can't just plug in averages. At least that's what I initially thought, but in this particular instance it works out correctly because
108.
by shiandow 7mo ago
It is definitely not obvious, but I wouldn't say it is completely unclear. For instance we know that algorithms like the leapfrog integrator not only approximate a physical system quite well but even conserve the energy, or rather a qu
109.
by shiandow 7mo ago
An important corollary to this rule is that that customer need not be you .
110.
by shiandow 7mo ago
In my experience LLMs also suffer massively from "not invented here" syndrome. I've seen them copy whole interfaces just to implement a feature that was already implemented in a dependency. All with verbose comments that are
111.
by shiandow 7mo ago
The one thing I don't quite get is how running a loop alongside an agent is any different from reviewing those PRs.
112.
by shiandow 7mo ago
It is worth pointing out that not all parts age equally well. The cushions especially are not that durable and should be considered consumables. Sennheiser provides replacements should you need them. The effect they have on the sound is muc
113.
by shiandow 8mo ago
Wait, am I reading this wrong. The producer and importer try to soften the impact of the tarrifs only for the retailer to massively increase their prices?
114.
by shiandow 8mo ago
That very much isn't the only right way, and it is far to close to government tracking activities online. For one it effectively allows governments to disallow someone from accessing the internet. All this to let you do stuff you were
115.
by shiandow 8mo ago
LLMs seem quite successful when considered something like a natural langiage interface, but expecting intelligence seems a step too far. For one they do not learn, at least not online, and that is a somewhat important requirement for truly
116.
by shiandow 8mo ago
Arguably the operating system (or potentially the user-agent) is the exact place to do this. What I don't get is why it can't just all be client side. An app will just signal "I am going to show 16+ information" and the
117.
by shiandow 8mo ago
The real question is which is more likely to avoid catastrophic failures in practice. And we 'tried until it didn't blow up immediately' is not a great sign.
118.
by shiandow 8mo ago
In that interpretation the total number of worlds does not change.
119.
by shiandow 8mo ago
Both are true in practice, so not unreasonable. For graph weights that is, not sorting. That said the finite alphabet and bounded length requirements can be softened a bit. Even for general sorting algorithms. I mean, for the kind of lexico
120.
by shiandow 8mo ago
Interestingly sorting is O(N) for a surprisingly large class of datatypes. Anything that behaves well with lexicographic sorting really. Supposing one uses a 'trie' as a priority queue, the inserts and pops are effectively constan

More ›