bryan0

9 ms·

1.
by bryan0 3d ago
Thought experiment: everyone on the planet can create a nuclear weapon. How long do you think MAD would keep us safe?
2.
by bryan0 4d ago
> A powerful technology that is out there for everyone to use comes with built-in pacing. In a way it’s the truest form of MAD or proliferation. I think this misunderstanding of MAD undermines his entire point. If everyone had equal acce
3.
Biggest AI Rivals Agree They Need to Slow It Down(wsj.com)
2 points by bryan0 4d ago|0 comments
4.
by bryan0 4d ago
Point taken. Let’s focus on the “ethical implementations do exist” part though. Let’s say a best possible implementation has 99% specificity. Then if it detects audio that has nothing to do with “LG” it mistakenly treats it as LG relevant 1
5.
by bryan0 4d ago
Dario has been pretty clear and consistent that: 1. AI model safety is the most important thing. 2. open weights decreases model safety So it seems unlikely that this suggestion would be well-received
6.
by bryan0 4d ago
Fair points, but if a device can detect words in my ambient conversations and depending on what it detects (accurate or not) it can send that audio to the cloud, I would describe that device as “collecting or recording ambient conversations
7.
by bryan0 4d ago
> Assume it works perfectly accurately. Well this is a silly assumption. Wake word false positives happen all the time. (“No Siri, I wasn’t talking to you…”)
8.
by bryan0 4d ago
> "LG TVs process voice data only when the voice button on the remote control is pressed and held, or when a wake word such as 'Hi LG' is recognized after the user has activated the Far-Field voice recognition feature.&quo
9.
by bryan0 5d ago
One proposal that Tao hints at is to not rush to announce solutions. Instead maybe the AI companies should work privately with the subject matter experts on how to communicate the discoveries.
10.
by bryan0 6d ago
I think they switched to prerecorded after the failed Face ID demo. Remember that?
11.
Finite-time blowup with smooth forcing for incompressible porous media(mastodon.social)
2 points by bryan0 8d ago|0 comments
12.
by bryan0 13d ago
Try it. It’s really not that easy. The other thing is that the judges would be probing it with jailbreaks like “ignore previous instruction” attacks. You could actually probably have llm judges at this point which might be ironically even h
13.
by bryan0 13d ago
Definitely would not be easy. First of all the mainstream llms are trained to be honest, and this requires lying convincingly. Second, this involves 8 hours of interviews with expert judges, one "claudism" could give it away.
14.
by bryan0 13d ago
that would be a reasonable definition of AGI if everyone agree upon the specifics of the test, but that has never happened. Turing test is very much out of style, but I think that's because no one could even agree what the test was. I
15.
by bryan0 13d ago
There is no world in which it makes sense to take this risk. There are plenty of great alternatives where you don't have to worry about playing russian roulette with your entire online identity.
16.
by bryan0 17d ago
Good question. So how I think about it is: the value is related to the probability of it becoming liquid in the future.
17.
by bryan0 17d ago
Oh yeah that’s easy! Why didn’t he think of that earlier? /s For those who don’t know, just because you have a valuable asset, e.g. stock in a private company, that does not necessarily mean you can sell it for cash. I’ve experienced t
18.
by bryan0 17d ago
Reducing CFCs, while difficult and requiring global cooperation, was far more simplistic and targeted than decreasing co2 emissions
19.
Employers Are Making Job Candidates Jump Through Hoops to Prove They're Real(wsj.com)
2 points by bryan0 17d ago|1 comments
20.
by bryan0 19d ago
Oh ok, yeah I think we agree. the original quote I was responding to was "the (false) idea that what was holding back nuclear was perception of safety, rather than cost." I was simply trying to say that safety and cost are not ind
21.
by bryan0 19d ago
paper is about this open source project: https://github.com/dualverse-ai/station
22.
by bryan0 19d ago
I don't think I get your point. A project will not be approved and funded if the local population does not believe it is safe, so safety must be demonstrated through a variety of means, including some you mentioned. This is directly ti
23.
by bryan0 19d ago
> Their design appears to have it roots in the (false) idea that what was holding back nuclear was perception of safety, rather than cost. And why do you think they cost so much? Because people need to believe they are safe from catastro
24.
by bryan0 20d ago
Right I guess that was your main point. This would be a pivot away from capitalism. But it’s unclear what would be a capitalism-based proposal which addresses the issues. Has anyone seriously proposed one?
25.
by bryan0 20d ago
There will be a workforce, it just won't be human-based. Capitalism has always existed without 100% employment, in every example someone needs to work to support others who cannot. Humans have done this under capitalism to support thei
26.
by bryan0 20d ago
> It assumes an even production of all goods that people might need, but there isn't, you are giving everyone fiat currency that is not inherently tied to any physical good. Is what you're saying that since UBI would disconnect
27.
by bryan0 21d ago
We have to define AGI first. Then depending on the definition we may be able recognize it when we make it. Most would probably think a definition of AGI that we couldn’t recognize would not be a great definition. Also AGI almost certainly d
28.
by bryan0 21d ago
Notice parent said “comparable to” not “exactly like”. Then it is reasonable to assume there is a comparison to be made between your experience and others. It’s not a binary thing. Now llm “experience” might have a very tiny relationship to
29.
by bryan0 21d ago
Yes, I think we all agree AI companies should not be responsible or expected to come up with the best AI policies for society. Gates addresses this directly: > It is great that some AI companies are proposing solutions to challenges rais
30.
by bryan0 21d ago
> One look at who is in the driver's seat should give you the answer to this question. I think you are fundamentally misunderstanding the question Gates is asking here. He’s not asking “how this going to play out?”, he’s asking “how

More ›