"Launch cadence" has been a standard term in the industry for a long time. https://books.google.com/ngram...

Yes, Canadian farmers in this case. Other countries may be in similar conditions. It is hard to compete with a country that subsidizes so much industry, which includes farms.

Those tariffs and the supply and quota thing is in response to a large player who heavily subsidizes their farmers to protect our farmers, who are generally small family operations.
There's a few things about Canada that Trump hates, from Margaret (Pierre Trudeau's wife, Justin's mother) not fucking him way back to us not supporting having Trump towers. At that we got rid of his towers before he was President, we never liked Trump.

Well that resulted in a spiralling tariff war. Weirdly this time every country except two have folded with "yes, we will pay the tariffs". It's weird as someone like Trump just escalates their demands. Its depressing being alone against the idiot with power.

If any president, regardless of party, declares a national emergency and that the only way to stop it is to block imports of something that we critically must have and cannot produce in adequate quantities for our country to not fall apart horribly, we will question that president's judgment, intellect, and/or character, because at least one of the three almost certainly must be compromised in some way to reach such a flawed decision.

Look, it's really simple. The word emergency means that something A. is sudden and unexpected, and B. must be dealt with right now, or else the consequences will be immediate and severe. This is not an emergency.

The U.S. power grid requires large quantities this hardware continually, both for expansion and repairs, on an ongoing basis. Not having adequate quantities is a national emergency with immediate implications for the country. Not having the manufacturing infrastructure to build critical power infrastructure hardware in this country could *eventually* be a national emergency if, for example, China decides to declare war on the U.S., but right now, it is just a long-term *risk* of *future* national emergency, not an actual emergency, because doing nothing will not necessarily cause immediate harm to the country.

You can't solve a long-term and likely mostly hypothetical risk to the country by creating an *immediate* national emergency. You can't put the entire country in *more* short-term, medium-term, and potentially long-term peril in the hopes that it somehow causes people to reduce some future risk that might never even materialize.

Basically, this policy is the import/export equivalent of saying, "My neighbor might steal my car someday, so I'm going to push my car off a cliff so he can't take it away from me in the future. Sure, I'll lose my job because I can't get to work, but having a car that he can steal is an emergency." See how stupid that is?

Worse, if this were an actual emergency (e.g. if China invaded Taiwan and the U.S. decided to shut down all imports from China in retaliation), there would still be a bigger problem, which is that it will take years to build the infrastructure to make the hardware needed in the United States.

What this means is that unless the emergency is so severe that every party agrees to maintain the ban, the minute the next president takes office, the policy will immediately be overturned in the name of national security, and it will no longer be financially viable to keep building those plants, and they won't ever get finished. Moreover, in the absence of such an exigent threat, the companies know that this is the case, and will drag their heels as much as possible before they even start building anything, so at best, all he will get in the next two years are a bunch of promises that will be soon broken.

So a policy like this cannot realistically have a positive impact on anything. If he had done it in the first year of his first term in office, it might have produced some success, but at this point, it's almost a decade too late.

That said, if he doesn't stop those imports, then companies might build wind turbines off the coast that he'll have to look at, so I can see why banning those imports could be an emergency for the president, and thus a "national emergency". [rolls eyes]

Not running a malicious AI doesn't do anything to protect YOU.

You are at risk from OTHER people running a malicious AI.

This report came from a security company, not a LLM developer. And their conclusion is pretty much the opposite of what you just said: any model that downloads unverified content from the web and treats it as instructions should be viewed as fundamentally insecure. The problem is models that trust unverified content.

Every competent engineer knows you should keep commands and data separate from each other. Mixing them in a single stream is begging for problems.

But is that possible for an LLM? I suppose every input token could come with a binary flag to indicate whether it should be interpreted as a command or not. When training a model with reinforcement learning, only reward it for following commands that are marked as commands. Also include counterexamples that have commands in the data stream, and give a negative score if it follows them.

But I suspect there will always be loopholes. The commands are interpreted in the context of the data, and if you can manipulate the data, you can probably make it misinterpret the commands in the way you want. Given a sufficiently large model whose behavior is unspecified and far too complex to understand, there will always be ways to manipulate it into behaving maliciously. It's impossible to guarantee safety.

And of course, the lack of separation between commands and data is part of what makes the models useful. You ask it to do something, it doesn't know how, so it looks online for directions. If you only want it to follow instructions you specifically approve, it will be much more limited in what it can do.

The labour participation numbers have dropped from something like 95% to maybe a third. Some of that is good, children taken out of the labour force, old people getting to retire, women becoming stay at home wives without having to spend the day spinning yarn for a penny a day and such. On the other hand, the streets are full of homeless people, basements full of unemployed young people, all kinds of people on disability though they could often be working etc.
Even back in the time of the loom, it took 70 years, 3 generations for jobs to bounce back. There was saving things like massive emigration to the new world and free land to counter the massive unemployment.
Perhaps your right and about the year 2100 employment will bounce back, it is still a lot of unemployment until then.

Ooh. I've always wanted to be a Geocities warlock. Or is that Geocities Warlock(tm)?

But seriously, loving the irony of Elon's petulant rebranding coming back to bite him in the a**.

The fact that twitter.com still exists as a redirect actually strengthens and weakens their claim to abandonment. A mere URL with no content does not generally qualify as use in commerce. And the existence of that redirect without Twitter branding on the resulting site makes it clear that abandonment is willful, because it indicates that the old name is no longer intended to be used, and that they don't have any future product plans with that name (because otherwise they would have a landing page saying that the new Twitter was coming soon).

IMO, X clawing back the Twitter trademark should be rather difficult for them at this point, famous or not, barring concrete documentation created prior to the 3-year limit showing future plans to reintroduce the mark in the very near future.

In relative terms, they're marking it up more. But ask yourself which hurts people's pocketbooks more: the $110 that the cheap Android tablet maker takes as profit or the $400 that Apple takes? Consumers don't care about percents. They care about dollars. And Apple's profit margin on their tablets is almost 4x that of their cheapest competition.

Alternatively that's three spare tablets for when the first one breaks. As soon as you're more than 2x the price of the cheapest competition, unless users absolutely have to have the capabilities of the expensive one, a lot of stuff like build quality ceases to matter, etc.

It sure sounds like they're trying to downplay it: "an outlier scenario", "a rare and unexpected confluence of events". To make you think this was something unusual that's not likely to happen again. But what are those rare events? "the presence of impossible tasks". You think it's going to be rare for models to get impossible tasks? Expect it to happen constantly. "model persistence over long task horizons." That's the whole point of these models, that you can give them a task and they keep working on it for a long time! "messages to peer models." They did that spontaneously. Since they did it this time, it's likely to happen frequently.

This sounds more like a test of how it behaves in a very typical situation.

And of course they emphasize all they things they've already done to prevent it from happening again, but with little evidence that they'll be effective.

Well, good luck. Part of what made Canada Canada is American political refugees.

Netflix knows your IP address and therefore your country (ignoring VPN's etc) and only delivers content that is licensed to be shown in a country. Lots of stuff we don't get due to licensing. You see it a lot on YouTube too where you click a link, maybe on Slashdot, to a video and a black screen comes up with writing saying something like the content creator hasn't made it available in your country.
As for the Canadian content, its in the catalogue and maybe high lighted. You don't have to watch it but the choice is there.