photonthug

6 ms·

1.
by photonthug 11mo ago
Adding complexity is just one aspect. Everywhere there is someone whose job is to ensure the bottom line never changes and status quo for the powerful is preserved. Insurance, taxes, rents.. in the absence of effective regulation, the av
2.
by photonthug 11mo ago
Yes, unfortunately a phrase that's used in an attempt to lend gravitas and/or intimidate people. It sort of vaguely indicates "a complex process you wouldn't be interested in and couldn't possibly understand".
3.
by photonthug 11mo ago
As something of a technophile myself.. I see a lot more value in arguments that highlight totally ridiculous core assumptions rather than focusing on some kind of "humans first and only!" perspectives. Work isn't necessarily
4.
by photonthug 11mo ago
> fully AI generated and fully AI review This reminds me of an awesome bit by Žižek where he describes an ultra-modern approach to dating. She brings the vibrator, he brings the synthetic sleeve, and after all the buzzing begins and the
5.
by photonthug 11mo ago
In philosophy 101 the usual foil for Rousseau vs.. would be Hobbes, but that framing with a realist/pessimist would not be popular with the intended audience, where the goal is to lionize the nationalist, the inventors/owners, t
6.
by photonthug 11mo ago
Funny but the eyebrow-raising phrase 'recursive self-improvement' is mentioned in TFA in an example about "style adherence" that's completely unrelated to the concept. Pretty clearly a scam where authors are trying
7.
by photonthug 11mo ago
Yeah, I mean I hope there are not many people that still think it's a super meaningful test in the sense originally proposed. And yet it is testing something . Even supposing it were completely solved and further supposing the sol
8.
by photonthug 11mo ago
Outcome would depend on the rest of the test, but I'd say the "human" version of this answer adds zero or negative value to chances of being human, on grounds of strict compliance, sycophancy, and/or omniscience. "
9.
by photonthug 11mo ago
I think the TT has to be understood as explicitly adversarial, and increasingly related to security topics, like interactive proof and side channels. (Looking for guard-rails is just one kind of information leakage, but there's lots o
10.
by photonthug 11mo ago
Hah, tools-or-no does make things interesting, since this opens up the robot tactic of "use this discord API to poll some humans about appropriate response". And yet if you're suspiciously good at cube roots, then you might
11.
by photonthug 11mo ago
Turing test isn't actually a good test of much, but even so, we're not there yet. Anyone that thinks we've passed it already should experiment a bit a with counter-factuals. Ask your favorite SOTA model to assume something a
12.
by photonthug 11mo ago
Agree with all this; see my other comment in thread for more color, more math. I don't want to come across as embracing pseudo-science, misinformation, etc. I do like to think about the distinctions and boundaries for hard/soft&#
13.
by photonthug 11mo ago
Not OP, but I think the plot twist is, maybe we need to be able to entertain "obviously absurd" ideas to be able to land on a correct position if the culture we're inside of is not ready for those ideas yet. (No idea if the
14.
by photonthug 11mo ago
There's some interesting stuff in here if you can tolerate the meandering and the way-back-when. Like you'd expect from po-mo wonks, everything's gotta be infinitely subtle and infinitely contextualized. So no big mea-culpa
15.
by photonthug 11mo ago
Most people do this so that they can eat the rocks afterward. They are shiny and very nutritious, and it strengthens the teeth. It's normal for some teeth to break off during this phase, but a) you already have colorful rocks to repl
16.
by photonthug 11mo ago
You might prefer this sort of thing: A Definition of AGI https://arxiv.org/abs/2510.18212
17.
by photonthug 11mo ago
The capital was better fortified and the French wanted food and loot. So softer target sounds nice, especially if you think this crushes morale immediately and don't believe the opposition will go scorched earth (but they did). > I
18.
by photonthug 11mo ago
> And temperature 0 makes outputs deterministic, not magically correct. For reasons I don't claim to really understand, I don't think it even makes them deterministic. Floating point something something? I'm not sure tem
19.
by photonthug 11mo ago
> It DOES fail more when the numbers are longer (because it results with more text in the context), I tried to raise this question yesterday. https://news.ycombinator.com/item?id=45683113#45687769 Declaring victory on &q
20.
by photonthug 11mo ago
Glad to see the pentagonal multiplication prism is just as weird as the addition helix https://arxiv.org/abs/2502.00873
21.
by photonthug 11mo ago
Thanks for doing this. OpenAI is not in fact open, so referencing their claims as obviously true on anything else is just a non-starter. Counterpoint though, it's been a while since I've run this kind of experiment locally, so I
22.
by photonthug 11mo ago
Agree, this stuff was trending up very fast before AI. Could be my own changing perspective, but what I think is interesting is how the signal it sends keeps changing. At first, emoji-heavy was actually kind of positive: maybe the project
23.
by photonthug 11mo ago
> “Settling Defendants have agreed not to provide nonpublic data to RealPage for use in competitor pricing recommendations and to refrain from using RealPage’s RMS that relies on non-public competitor data to make pricing recommendations
24.
by photonthug 11mo ago
Seems like an optimistic read on things. This is the kind of common-sense approach you would expect in a world without lawyers, just observing that collusion is bad because the effects are bad, and digging into the details of the causes ar
25.
by photonthug 11mo ago
> Modern price collusion is more apt to happen with A/B testing if prices at locations to see what the local market will bear. One of my first thoughts as well. If you're big enough, you collect so much data and run so many ex
26.
by photonthug 11mo ago
> feels like the wrong question to me I agree but had different questions. TFA mentions the consideration of whether failure cases are correlated, but of course if OpenAI wins big, there's a good chance this directly or indirectly
27.
by photonthug 11mo ago
> What's to be gained from talking like this in public as a corporate figure? Diving into the game theory of a 4-player setup with executives/investors/customers/workers is tempting here but I'll take a different
28.
by photonthug 11mo ago
If we fill those abandoned buildings with people, air-conditioning the inside of the building for them will obviously add even more heat to the outside? Parking lots that are full of cars aren't going to be that much cooler than empty
29.
by photonthug 11mo ago
It's actually kinda noteworthy that corporations don't talk like this (yet). Masks are off lately in political discourse, where we're all in on crass flexing on the powerless, the othering, cruelty, humiliation. How long be
30.
by photonthug 11mo ago
Obviously platforms get advertiser dollars, but the question is what the business paying for that gets. The answer is.. almost nothing? Dedicated marketing/advertising resources at your business is probably just a fifth column, where

More ›