ben_w

42 ms·

811.
by ben_w 4mo ago
> On that front, I think we're maybe 8 years out from mostly-competent robots. I think this depends on what is meant by "mostly-competent". As you say: > I'd peg robots at years after perfect self-driving, and we&#
812.
by ben_w 4mo ago
> FABs are another issue entirely. Re terafab. It's very difficult to audit silicon. Just cutting the top off a chip layer by layer and using an electron microscope, isn't going to fully give you trust. People have been worried
813.
by ben_w 4mo ago
> It very frequently feels like management is making strategic decisions after snorting a long line of social-media-psychosis and TED talks. It is remarkable that investors have any faith in such founders/entrepreneurs at all. My gu
814.
by ben_w 4mo ago
Aside from all the stories where AI does exactly what it was told to instead of what the creators meant? Apart form them, never underestimate British humour's ability to contradict narratives of competence*: https://www.yout
815.
by ben_w 4mo ago
That % of tokens sounds about right from the things I did with free trials of the $€20 plans earlier this year. However, the LLMs on the projects I did back then really sucked at project planning and task breakdown, they couldn't relia
816.
by ben_w 4mo ago
> Their interest is irrelevant in the face of technical impossibility. I'm responding to "If the data was separable in this way, you would equally be able to train an AI to mask those signs.": yes, if you wanted to you cou
817.
by ben_w 4mo ago
Really just one shot?
818.
by ben_w 4mo ago
IIRC, the big names in LLMs have no real interest in cloaking the LLM-nature of the text, Google adds deliberate watermarks to text, OpenAI developed a watermark for text but reportedly arent't actually using it.
819.
by ben_w 4mo ago
> There's no real shortage of steel around the world that I know of. We could just stockpile it instead, for example. And in the hypothetical tail-risk scenario this is all supposed to insure against... how do we even get the raw ma
820.
by ben_w 4mo ago
> On the other hand I don't think many governments are that concerned about another pandemic, sometimes the discourse regarding it very much sounds like "what are the odds it'll happen again" Gambler's fallacy wi
821.
by ben_w 4mo ago
As I recall, the economic damage of the pandemic was ~ "it's worth spending ten billion dollars to make the vaccinations arrive one day sooner": https://www.newsweek.com/operation-warp-speed-what-deal-opin...
822.
by ben_w 4mo ago
Both things, and both reasons, can be true at the same time. Not everyone's going to care about Anthropic requiring data collection (a similar debate plays out with regards to "pay or consent" on website tracking), just as no
823.
by ben_w 4mo ago
> A single Dorito would be so flavorful you wouldn’t be able to bear it. Did this by accident once, using a flavour pack meant for curry to serve 6 people in a single home made naan bread. (No I will not elaborate why). It was indeed so
824.
by ben_w 4mo ago
> You've mistaken me for the guy you originally replied to. It happens, no worries, but I'm not that guy. I'm just replying to the list I saw you state. Oops. Yup. blfr vs b112. > And of course, gmail endlessly scanned
825.
by ben_w 4mo ago
Nah, that part already began a while back.
826.
by ben_w 4mo ago
A nail, from what I hear about that side of things, but on the Apple side I could substitute the dual replacement of UIKit and AppKit with SwiftUI that somehow still isn't good, plus separate pressure coming from all the cross-platform
827.
by ben_w 4mo ago
Careful, looks like you're shifting definitions. Remember, I was saying "by this definition", where "this" isn't about capabilities, but about effort needed to get there, which in the case of your link is &quot
828.
by ben_w 4mo ago
> I there's anyone I don't trust with AI, it's the worlds #1 company in spying on people, in collection of Pii, in tracking, and many many many times caught literally lying about it. If you gave me this description in isol
829.
by ben_w 4mo ago
In fairness, there is no contradiction between "most trustworthy" and it being naïve to then actually trust them. Our ability to debug or decompile artificial neural networks is only better than doing so for living neural networks
830.
by ben_w 4mo ago
They put it behind a paywall and didn't fix it, according to more recent lawsuits than that article. Also, failed to correctly notify authorities even when they eventually notified them at all. https://storage.courtlistener.
831.
by ben_w 4mo ago
Rank ordered by reputation / caring about having a trustworthy corporate identity: [Google, Anthropic] in either order depending who you ask, OpenAI, most of the Chinese AI corporations, then Grok. This is unfortunate situation to find
832.
by ben_w 4mo ago
It doesn't matter what I think in this case (I have already decided to not give Musk any money), what matters here is what Musk thinks and does, and how the actual real market reacts to that.
833.
by ben_w 4mo ago
> people who can make a vastly superior product time and time again, and they all like to work with other people who can do the same These parts are likely pushing the best minds away from Musk these days. Not sure when that turned, perh
834.
by ben_w 4mo ago
Also, the lifetime of the panels on the ground is longer by about the same multiple as the power advantage of going to space, so if you're limited by the size of your photovoltaic factory, there's no reason to go to space.
835.
by ben_w 4mo ago
A few days ago people were telling me the Starlink was brilliant because it gave small African villages with no mains electricity an internet connection to share. This kind of behaviour does rather dampen the value proposition for such peop
836.
by ben_w 4mo ago
Yes. Have you met most people? Most people won't bother, find this kind of thought process to be somewhere between nerdy and paranoid.
837.
by ben_w 4mo ago
As a European, I'm hoping this plays out like the Russian attempts to do the same to pre-invasion Ukraine. They took the money, spent it on themselves instead of an influence campaign, reported positive results upwards. I couldn't
838.
by ben_w 4mo ago
I'd say this explains why all the code is bad now, but I remember what it used to look like too.
839.
by ben_w 4mo ago
> For it to be worth a trillion dollars it has to take a trillion dollars from somewhere. Musk's SpaceX TAM is roughly as you say, it's "subsume all currently existing desk jobs", which makes it simultaneously absurdl
840.
by ben_w 4mo ago
> But you still have AI founders going out in public and saying they want to get rid of jobs instead of some pitch about how easy work will get or you'll be happier or etc. Why? Even if they did, would you believe them? Here's

More ›