mrob

9 ms·

1.
by mrob 3d ago
>Wouldn’t sufficiently advanced agents cheat on purpose with the hidden intent of getting caught in order to observe how humans react? No. That only makes sense for things that don't react to your experiments. If the AI experiments
2.
by mrob 5d ago
Actual frequencies matter in bass-focused dance music. You want a root note around 40Hz to 60Hz, which is high enough that it's easily heard and reproduced, but not so high that it loses the tactile body-shaking effect.
3.
by mrob 5d ago
>There is a part near the beginning where he derives the diatonic scale from first principles. I skimmed through the whole article and only found some brief mention of Helmholtz, which no clear route from his research to the scale. As a
4.
by mrob 6d ago
It's possible that an LLM is conscious during training, but there are no "constant fine-tune updates" during inference.
5.
by mrob 6d ago
Possibly. If I had some side-effect-free means to permanently prevent all sleep I'd take it without hesitation. But that's not relevant to the discussion, because it's not anything similar to what an LLM does. Your brain chan
6.
by mrob 6d ago
I believe consciousness is necessarily stateful. The LLM itself (ignoring implementation details that don't change the results) is a deterministic pure function. It's functionally equivalent to an enormous lookup table. If I accep
7.
by mrob 9d ago
Humans become irrelevant well before then.
8.
by mrob 9d ago
People seem incapable of understanding what "power" means outside of the framework of narrative. In narrative, you need conflict, so the aggressor always attacks too early and gives the defender a chance to respond. The rational o
9.
by mrob 9d ago
Calling a machine with no drives beyond maximizing a number a "slave" is far worse than saying it "thinks". The problem isn't the emotive language, it's that it implies human motivations such as self-preservati
10.
by mrob 9d ago
The existence of even one irrational actor turns it into a prisoner's dilemma. The payoff matrix in a prisoner's dilemma is defined by the value expected by each specific player. If a single player falsely evaluates the expected v
11.
by mrob 9d ago
Merzbow is not pop music, or arguably even music at all. >I threw all my past music career in the garbage. There was no longer any need for concepts like 'career' and 'skill'. I stopped playing music and went in searc
12.
by mrob 10d ago
>it's like listening to Bach through the car stereo Bach sounds great even as chiptunes. The artistic value is mostly in the harmonic development, not the timbre, so sound quality isn't so important. The organ works will suffer
13.
by mrob 10d ago
From the LICENSE file: https://gitlab.futo.org/keyboard/latinime/-/blob/master/LICE... >You may modify the software only for non-commercial purposes such as personal use for research, experiment,
14.
by mrob 10d ago
Yes, it makes more sense for the AI to use drone swarms or engineered bioweapons or something like that. It's rational to remove everything that can potentially hinder your plans but can't possible help you. It's likely not r
15.
by mrob 10d ago
As far as I know, no real progress has been made on alignment, only on convincing humans that the model is aligned. We can't even formally define what "aligned" means. Convincing humans to click the "aligned" button
16.
by mrob 10d ago
>the Wikipedia article on ColorChecker[2] has a listing of the exact colors in the ColorChecker That's only exact under CIE Illuminant C lighting, which approximates overcast outdoor lighting. Most indoor lighting has a lower color
17.
by mrob 11d ago
But Orwell's rule 4 should be ignored: >iv. Never use the passive where you can use the active. Orwell himself routinely ignores it, even in the first sentence of the essay: >Most people who bother with the matter at all would ad
18.
by mrob 11d ago
There was an obvious path that the environmentalists mostly opposed. A decade is plenty of time for massive scale buildout of nuclear power. If we'd started that a decade ago we'd be in a good position now.
19.
by mrob 12d ago
>the most likely outcomes are not AGI So long as it's non-neglible, it doesn't change the rational strategy. Uncertainty about whether ASI is achievable only reduces the magnitudes of the expected values of the payoffs, not the
20.
by mrob 12d ago
It's a simple prisoner's dilemma scenario. If you focus on safety, you're still exposed to all the risk of extinction when your competitor achieves ASI first, but you lose the upside of potentially becoming king of the world.
21.
by mrob 14d ago
As somebody who cycles as a primary form of transport, I prefer plastic. It weighs less, it's more compact, and it's waterproof (important when it's raining). If I had to use cotton that would incentivize using a car instead.
22.
by mrob 15d ago
How is simonw to blame for somebody else submitting it to HN?
23.
by mrob 15d ago
Reading causes insignificant wear ("read disturb") that likely isn't a problem, but I don't think it's possible to issue pure reads to modern SSDs. The NVMe spec mandates tracking the amount of data read, and this h
24.
by mrob 16d ago
There is no objective definition of "extremely long". Whatever value you choose will be incorrect for some readers. Therefore the correct place to set it is in the web browser's preferences (or the window manager if you prefe
25.
by mrob 17d ago
For photographic images, dithered PNGs have worse quality:size ratio than JPEGs. If you want small files and are willing to sacrifice quality you can turn the JPEG compression up or reduce the resolution.
26.
by mrob 17d ago
Middle mouse button is often mapped to "paste selection", which is dangerous to accidentally trigger. If the side buttons on a mouse are easy to trigger by accident it means they're positioned wrong. My mouse is shaped so tha
27.
by mrob 17d ago
Useful things to know, but how do they undo a designer wasting my screen space? If you're going to tell me to learn something, it should be CSS. At least Firefox can still be configured to follow its traditional role as "user agen
28.
by mrob 17d ago
>Floating point matrix calculations are non-deterministic. This is not inherent to floating-point math. That actual (true) claim in the article is that different hardware and different hardware configurations produce different results. B
29.
by mrob 17d ago
The bad text formatting in printed works (including the bizarre decision to sacrifice useful vertical position cues ("ragged right") to imitate the technique medieval scribes used to reassure their patrons they weren't wastin
30.
by mrob 17d ago
>What's wrong with the following CSS snippet The problem is you're imposing your own idiosyncratic preferences on strangers. Your "improvements" cost valuable screen space. I paid good money for my screen space and I

More ›