ben_w

12 ms·

181.
by ben_w 1mo ago
> Humanity will be wiped out by 2030. Sure. So far as I'm concerned, 2030 is much too short a timeline. It's not physically impossible, it's just that atoms are harder to get right than bits are, so an artificial (as oppos
182.
by ben_w 1mo ago
Mmm. This was published at around the same time the OpenAI models were hacking HuggingFace: https://metr.org/blog/2026-05-19-frontier-risk-report/#pilot... The research in the publication ended about 2 months befo
183.
by ben_w 1mo ago
> Even if they aren't, an AI with enough resources can continue doing that research on its own. Unclear how true this is. Atoms are harder to get right than bits are, simulations need grounding against measurements. > Ultimately,
184.
by ben_w 1mo ago
I have good news and bad news. The bad news is that the world's richest man (on paper) is currently building something he himself described as a "robot army". The good news is that his timelines have historically been wildly
185.
by ben_w 1mo ago
You think China's incapable of faking performances at least as well as Tesla? Or just lacking the motivation to do so?
186.
by ben_w 1mo ago
I'm enjoying myself, but then again I'm not currently employed, so not in your target audience. Related: Anyone looking for a nerd in Berlin, mit B1 Deutsch?
187.
by ben_w 1mo ago
> I talked about progress, which is indisputable, whatever its pace or current limitations might be It's extremely disputable, and indeed disputed. Just as with my example of Tesla's video, it's easy to fake progress. All
188.
by ben_w 1mo ago
At the present time, that's like being handed a random human and asking if there's any proof/evidence of their brains including signs of being a secret double agent. What we can monitor for is output, and very crude probes in
189.
by ben_w 1mo ago
It might. One potential failure mode is the owners of the robots doing something that echoes words Musk himself has used: "robot army". Who is sovereign in this future? The government? The people? The bosses of the AI/robotic
190.
by ben_w 1mo ago
I'd temper my views of the apparent recent Chinese robotics progress. It's more than nothing, but there can still be a big gap between what is presented and the real timeline to follow; see e.g. the US example with Tesla's &q
191.
by ben_w 1mo ago
Given the pension ages in much of the world, and the USA in particular (I'm guessing which side he was on in Vietnam, but Vietnam's state pension age is lower), I'd hope this was in the past tense rather than the future.
192.
by ben_w 1mo ago
The robot part exists, the hard part is the intelligence for that robot to be autonomous. Stuff like this is already available as a standard industrial product: https://www.hilti.com/content/hilti/W1/US/e
193.
by ben_w 1mo ago
> Mostaque cites an analysis of a Chinese open-weight model: “If you say you’re Uyghur or from certain parts of China, it will write backdoors and hacks in the code.” Every model deserves the same scrutiny, he argues: “You have to get th
194.
by ben_w 1mo ago
> The immediate reaction of the other parties was roughly the same as the USA's immediate reaction. Other Western parties, I should have written. Wikipedia list of parties to the Budapest Memorandum: Belarus, Kazakhstan, Ukraine,
195.
by ben_w 1mo ago
> Second, yes the Budapest Memorandum was violated. Why didn’t the other parties do anything or immediately declare war on Russia besides just the United States? The immediate reaction of the other parties was roughly the same as the USA
196.
by ben_w 1mo ago
> And if he has any levers, he certainly doing a terrible job one distraction after another. I think you may have misunderstood, my point about levers was there's may ways for him to *start* a war with everyone else in NATO, potenti
197.
by ben_w 1mo ago
> In the last couple of weeks the director of the CIA met with Russian counterparts in an attempt, at least as alleged, to provide information to Russia to get them to come to the negotiating table. Giving away intel is the exact opposit
198.
by ben_w 1mo ago
> Didn’t suggest it was, but at the end of the day it’s the US who is the partner at the table who can make the biggest difference and bringing Ukraine in to the room with Russia will just result in conflict and no hope of figuring thing
199.
by ben_w 1mo ago
I can't tell if you're being sarcastic (because you want a $5k car and they don't exist in your market) or serious (because you want a $5k car from China which they do actually sell at that price).
200.
by ben_w 1mo ago
> Was it the land that Russia currently controls and Ukraine is very unlikely to retake? There's a spectrum between ideological behavior and pragmatic behavior. If Ukraine gives up this land and achieves peace, maybe it's worth
201.
by ben_w 1mo ago
> The real luxury, for any age, in general, is in not having dependencies and needs. The need to impress someone, the need to know every ugly news from around globe, the need to read everything and be knowledgeable, need to stay connecte
202.
by ben_w 1mo ago
Probably algae. Land only slightly beats the ocean, like a 53-47 split or something, and trees are only part of the primary production on land: https://en.wikipedia.org/wiki/Primary_production This is also why you some
203.
by ben_w 1mo ago
I've managed to avoid driving one for all but 1-2 years of my life, and I'm 42. Your mileage, ironic phrase noted, may vary.
204.
by ben_w 1mo ago
Tells you how seriously they take the idea of putting a significant wattage of compute in space any time soon. Whatever happened to SolarCity and solar roofing shingles? Rhetorical question, of course: https://en.wikipedia.org&#x
205.
by ben_w 1mo ago
Yup. If "near term" is "about half a lifetime". Doomers think I'm a ridiculous optimist, people who think "doomer" is a slur think I'm a ridiculous pessimist. For comparison, if you take the general f
206.
by ben_w 1mo ago
Depends what you mean by "spiral catastrophically". My P(doom in my lifetime) is about 0.1, because I think minor catastrophes will break something hard enough that it doesn't get to escalate to actual-doom. A recent-ish link
207.
by ben_w 1mo ago
I miss this kind of skeuomorphism. It was fun: https://github.com/1etu/nos4/blob/main/nosmarketingassets/no...
208.
by ben_w 1mo ago
• SwiftUI: it's just that bad. • Analytics. Or at least 99.9% of it, no need for a literal thousand companies: crash logs and anonymous API use stats will get you 90% of what really matters. 1. Analyse everything 2. ??? 3. Prof
209.
by ben_w 1mo ago
In addition to what superb_dev's comment says: > If the point is to get your gold away from the US and its president, why leave ~19% of if there? Not all of the risk is "the USA may literally steal it", some of the risk is
210.
by ben_w 1mo ago
Same. I'm in Berlin, I may say "I'm going to the city/town/village" for different parts of the city-state.

More ›