Anthropic made Skynet? I think not.
As it happens, the Chinese are capable of making their own near-frontier models, many of which they release publicly as open weights.
It's not hype, Anthropic is literally saying there's no actual risk they're aware of.
What actually happened is they refused to let US Intelligence agencies use their models to perform illegal surveillance. So now the US government is illegally punishing them.
Congestion pricing is only an option in places that have good alternatives to driving, something that a freeway in California does not have.
Working from home is an alternative, one that we should use more.
(Of course, I WFH full time and have for 20 of the last 30 years, so I have a bit of a bias.)
While I'm all for the American dream there needs to be a hard limit on how much money a single person is allowed to accumulate.
You do know that Musk doesn't actually have a trillion dollars in *money*, right? He doesn't have anywhere remotely close to that much money. His total liquid assets are extensive, sure, maybe as much as a few billion, but nearly all of his incredible net worth isn't money. You could probably call it "potential money".
There's no doubt that Musk has near limitless funds at this point. But, "trillionaire" is just paper games.
And everyone should keep in mind that this is true for basically all of the billionaires. Not that there isn't real wealth there, but it's a lot fuzzier than the numbers appear. Basically everyone with astronomical wealth mostly owns shares in companies, and how much of that value is real in any near-term sense depends on a lot of factors.
Musk's wealth is more speculative and fuzzy than most because his companies' valuation is based not on the revenues the companies generate now but theories about what they might generate in the future. Tesla's high valuation is all about the promise of self-driving cars restructuring transportation. SpaceX's is a little bit about cheap access to space changing a lot of stuff and more about AI. In all cases the high valuations are bets on world-changing technology being becoming real, and on Musk's companies being able to capture a good chunk of the resulting revenues.
The whole point of stock markets and such is that you have hard core rational investors ensuring valuations are accurate.
But Musk figured out that you don't need solid fundamentals, all you need is a hyper-loyal core following of retail investors who support you regardless of fundamentals. So you just bullshit just enough to keep them happy, without crossing too far into fraud, and the retail investors stay on board.
The retail investors create a floor for the price and their crowing about their winnings creates a bubble. The institutional investors then see what's happening and hop along for the ride.
The result is the two most overvalued companies in history (Tesla and SpaceX). In theory, the whole pile eventually comes crashing down, but just like any bubble the fund managers who buy are unlikely to be the fund managers left holding the bag.
But that is a weak causal story compared with the much more direct variables everyone is living through: housing costs, wage stagnation, student debt, childcare costs, healthcare costs, delayed household formation, and wealth being increasingly captured by the top of the economy
That analysis is utterly wrong. Far, far worse than the smartphone theory.
It is, in fact, the almost exact opposite of the truth. The truth is that wealth is what causes fertility decline. Wealth and female education, actually, which come hand-in-hand. This story is strikingly visible everywhere around the globe. As a population becomes wealthier and its women become better-educated, fertility falls. Without exception, and the effect is so powerful it overrides culture, religion, everything.
This is the primary driver in the US, too. In fact, wages have not stagnated, not when you look at the full picture including government transfers, and every generation is wealthier than the one before. Somewhat surprisingly, given the current housing price bubble, each generation even has higher home ownership rates than the previous generations at the same ages. Houses and apartments are also significantly bigger and more luxurious (which explains most of their higher prices, actually; do some comparisons on a per square foot basis over time, then adjust for the higher quality and greater amenities we have today).
But if you look at how Americans spend their money over the years, the biggest change you'll find is that we spend less on housing, food and clothing as a percentage of our income (in spite of bigger, nicer houses, far more restaurant and delivered food, and much larger wardrobes) and much more on entertainment -- and that in spite of the fact that entertainment has gotten dramatically cheaper.
And then, you lose your country....the culture is lost, what makes your country YOUR country....disappears.
The US solved this problem 150 years ago. First with the observation that immigrants acculturate. Second with the acceptance that elements of their culture are going to get melded in to form a new culture. Culture is never static, anyway, it always drifts and morphs. Immigration just changes it a bit faster. But it's good! This ongoing immigrant-driven culture change is what made the US a superpower. Embrace it.
However, immigration is only a stopgap solution to the problem of population decline, because fertility is declining everywhere on the globe, fast. The global fertility rate is basically at replacement now, but the decline is continuing, and accelerating. We'll drop below replacement as a species in just a few years. Even then population will keep growing for a while due to the "filling out" effect, but then it'll start dropping, fast. And it will quickly become top-heavy (more old than young).
Our economic system does not cope with population decline.
Probably not just our economic system, our civilization as a whole, though AI may change that. A highly technological civilization depends on having a large population because it depends on a vast amount of knowledge, which requires a tremendous amount of specialization. Some of this is the obvious sort, such as the scientists and engineers who are focused on increasingly-narrow areas of expertise, but a lot of it is not at all obvious, especially in industry, where everything we make requires a huge amount of knowledge that was learned by doing and isn't -- and maybe can't be -- taught anywhere but on the job.
To some extent we could probably manage with a smaller population if more of the population became highly educated (not necessarily in the academic sense, though we'd need that, too), but that transition wouldn't be easy, in part because there are lots of people who simply aren't interested in highly-technical work. We'd need a lot more of them to become willing to learn and do it anyway. Obviously the first step would be to bring the whole remaining population up to what the developed world considers a basic level of education -- that would enable us to tap new supplies of scientists, engineers and technicians. But the population reduction that seems likely to come means we'd need a lot more than that to be able to maintain our knowledge base and production diversity.
AI might change all this, of course. It could make it completely unnecessary for humans to participate in any of the above. But without something like that, it seems unlikely that our technological civilization could survive with less than a billion people or so, and technological progress would likely take a severe hit long before we hit that level of population reduction.
It's a crime in the US to "shout fire in a movie theater". Guess Americans live in Soviet times too.
NO. It is a crime to FALSELY shout fire in a theater. Huge difference.
Even that isn't true. The correct statement is that it's a crime to falsely shout "Fire!" in a crowded theater in circumstances where doing so is reasonably likely to incite a panic that would cause death or injury.
If you shouted "Fire!" in an American theater today, odds are that everyone would just ignore you, and a reasonable person would expect that to be the outcome. That's because fires in modern theaters are (1) quite rare, (2) much more likely to be announced by an alarm and (3) not difficult to escape safely due to the plethora of clearly-marked exit routes that are required by law. So everyone knows that in the event of a fire trampling people in an effort to escape is not necessary for self-preservation, that calmly walking to the nearest exit is better.
It's worth noting that the truth or falsity of the shout isn't really even much of a factor in the legal question. What matters is whether a reasonable person would expect that the shout would be likely to cause unnecessary death or injury as opposed to not shouting. The original Schenk v. United States case opinion in which Holmes used "Fire in a crowded theater" as an example, did argue that truth was an absolute shield, but later rulings, especially the 1969 case that overturned Schenk, removed that protection. If you shout intending to cause real physical harm and expecting that to be the result of your action it's a crime, even if what you shouted was true.
I'd have a lot more accidents if my reaction time was 0.5 secs.
The number of accidents you have depends far more on how you drive than how fast you react. If you and Claude are correct about AVs having slower reaction times that just highlights the importance of driving style over reaction time, because AVs have fewer and less severe accidents than human drivers.
The main thing you can do -- and AV systems do, generally -- is leave yourself more space and therefore time to react, which includes driving slower in areas where sudden incursions into the roadway are likely.
That said, I expect AVs to react faster as their systems improve. All of our AI is excessively compute-heavy right now, but we know that isn't essential because our brains do more with less in spite of the fact that our wetware runs on a significantly inferior substrate. As we learn how to build more efficient AI systems, that should reduce AV compute requirements and make it feasible for them to "think faster", and therefore react faster.
The one thing that is clear is that AVs will continue improving on their already better-than-human safety records.
I recall that this was discovered a long time ago when sales and marketing people realised that people would tend to turn right after they enter a store. I also seem to recall that this didn't hold true for left-handed people.
It would be interesting to see data from countries that are left-hand traffic. Streams of people in left-hand traffic countries tend to walk on the left side, and tend to move to the left if someone is walking towards them - which tends to be fun when walking about a right-hand traffic country! Though given these results were also tested in Japan, which is left-hand traffic, I'd expect there isn't a difference.
Would make sense, if you're right handed and turn to the right you're leading with your dominant eye (and hand) so better able to handle threats that emerge.
Similar if you choose to pass someone on the left, you present them with your stronger (right) side.
But this study supposedly found no correlation to handedness, so either something else is going on or maybe a sample size issue.
Sorry, let me clarify - the modern conceit of applying that term to the Arab (Jordanian, Syrian and Lebanese) and Egyptian people who left instead of becoming Israelis, was invented by Arafat (an Egyptian). The word Palestine was derived from Philistine, referring to the child-sacrificing enemies of the Jewish people who were famously defeated by David and ended up founding Carthage. Those people no longer exist.
It is thought that Rome renamed the region as an insult to the Jews who were driven from their land after they revolted in the first 70's. Arafat resurrected the term in order to unite the ethnically diverse "refugees" - which I put in quotes because they are not refugees, they chose to leave instead of becoming citizens because that's how much they hate the Jews.
Drop the bigotry. They left because their land was being given away by Western superpowers to a bunch of foreigners. They did exactly what any group of people would do. And they had a far better claim to that land than Jews who hadn't lived there for centuries, especially considering those Palestinians were actually descended from the ancient Israelites.
On the other hand the ethnic cleansing in the West Bank being carried out by Israeli Settlers, with the support of the Israeli government, is not typical or excusable.
Makes sense. The same standards apply to humans. If we were to tweet something completely made up, there is a chance of legal troubles. So should be the same for AI
Have you ever tweeted something completely made up? What happened? Or, if you haven't done it, what do you think would happen? Suppose, for example, that you tweeted out a claim that "Coca-Cola contains extract of ground-up baby brains". What do you think the legal consequences of that (horrendous!) claim would be?
There is an important legal distinction that this court chose to ignore, which is that you're only liable for incorrect information if it's reasonable to expect that people would believe that you are providing correct information. If you, bubblyceiling, tweet false information, you will not, in fact, be held liable for it, because courts would rightly reject the claim that readers had a reason to believe they should trust you.
Obviously, Google's statements are held to a higher standard that bubblyceiling's. But everyone understood that web search results weren't Google's statements. The question at hand then is whether people believe that Google's LLM's statements are, in fact, statements made by Google, the corporation.
No one could seriously believe that. This court was dead wrong.
Headline is that Solar produced more power in May than Coal in the U.S. Yet most of the comments here are about how evil Trump is and how he's destroying the environment or what not. Which is it? Is Solar increasing electrical production share under this administration, or not? Conflating whether or not people's political preferences align has nothing to do with the other.
I think you missed that most of those comments about Trump are gloating that he is demonstrably failing in his effort to destroy renewable power generation and favor fossil fuels -- and especially Beautiful Clean Coal (just like he's failing at approximately everything else, except this failure is good). Once you have that context, it makes a lot more sense.
The US could turn off all electricity and cars and use zero energy and Indonesia, India, and China would solely continue to destroy the environment at just about the same rate. Just to put things in perspective.
Well, China, for one, is building renewable energy generation far faster than we are. They're also building a lot of coal plants so it's going to take them some effort to push their emissions down to the global average per capita. However, note that we're far, far above the global average, and also well above China.
As for India and Indonesia, their emissions are already well below the global average, so they're not really the problem. Once we and China get down to their level, then we can all start pushing the average (and therefore total) down further. We need to cut our emissions by 85% to get to that point. Or keep them constant while importing about 1.7 billion people.
Within a computer, natural language is unnatural.
