FermiEstimate

joined 1 year ago
[–] FermiEstimate@lemmy.dbzer0.com 4 points 11 months ago (5 children)

Large, non-nuclear EMPs mostly use explosives. Covering a large battlefield means you're essentially bringing a massive, single-use explosive charge to the battlefield, staying uncomfortably close enough to benefit from it, and trying to set it off at exactly the right time, because they're not reloadable. And your enemy is probably thrilled you're doing this, because it saves them from hauling their own explosives there. (On that note, why are you sitting on this thing instead of dropping it on the enemy?)

This is in addition to whatever shielding you brought, which is likely bulky and conspicuous. And you're probably not doing combined arms, because shielding infantry and light vehicles from massive explosions is, it is fair to say, something of an unsolved problem.

But wait, you might be thinking. I know there are non-explosive ways to generate EMPs. Yes, there are, but you need a power source for those, and if you have a really good, portable one of those and a consistent supply of fuel to run it, you probably have better uses for it, like powering a modest laser. Oh, also, you're 100% sure your shielding works perfectly, right? You'll find out quick if you don't.

[–] FermiEstimate@lemmy.dbzer0.com 38 points 1 year ago (2 children)

Environmentalists are fond of saying that “There is no second Earth“. They are wrong! Here’s why: 

There is an entire second Earth right here on Earth.

Second Earth is a waterworld. It’s the vast Pacific Ocean that covers half the planet.

Well, he's a little fuzzy on the concepts of halves and wholes, but let's hear him out on colossal geoengineering projects.

Nobody:

Absolutely nobody:

The ghost of Sam Hughes: Okay but have you considered

[The AI]’s going to fall in love with you

Fortunately for everyone, they went out of business before a mandatory reporter had to make the weirdest call ever to CPS.

[–] FermiEstimate@lemmy.dbzer0.com 7 points 1 year ago (3 children)

Ah, right, I guess that's why other vending machines never caught on. Why spend $2 on a Snickers at work when a quick trip to the grocery store can get you candy for way less?

What you're overlooking this time is vending machines sell convenience, not just single-serving portions. The fact that very few customers really need ammo without leaving the store/mall is indeed why this is a questionable business model and not just a sketchy one.

I'm puzzled, though, by the belief that hunters are more likely to make overpriced, impulse purchases of ammo than mass shooters. I'm even less inclined to buy that than ammo from a vending machine.

[–] FermiEstimate@lemmy.dbzer0.com 15 points 1 year ago (5 children)

You're forgetting mass shooters, i.e., the people who don't care if they're identified or if they're getting a good price. Safe to say they're not worried about their credit rating if the plan is to take on a SWAT team in 20 minutes.

[–] FermiEstimate@lemmy.dbzer0.com 1 points 1 year ago (2 children)

Is it actually good against tanks now? I always liked it, but it still hurt to finally get a shot off just for the tank to shrug off the hit.

[–] FermiEstimate@lemmy.dbzer0.com 30 points 1 year ago (3 children)

American Rounds

What, was the Circus of Values brand too expensive to license?

Oh, hey, I've run into this in the wild--the Kalendar AI people keep ineptly trying to start a conversation to sell some kind of kiosk software by referencing factoids they scraped from our latest press release. They've clearly spent more effort on evading spam filters and rotating domains than they have on anything else, but they helpfully use "human" names ending in "Kai," so creating a wildcard filter wasn't too hard.

Credit where it's due: I'd never heard of Kalendar or the software company who hired them, but this experience has told me everything I need to know about both of them. If you don't sweat the details and rate sentiment change using absolute value, that's kind of impressive.

[–] FermiEstimate@lemmy.dbzer0.com 9 points 1 year ago (1 children)

Addressing the “in hell” response that made headlines at Sundance, Rohrer said the statement came after 85 back-and-forth exchanges in which Angel and the AI discussed long hours working in the “treatment center,” working with “mostly addicts.”

We know 85 is the upper bound, but I wonder what Rohrer would consider the minimum number of "exchanges" acceptable for telling someone their loved one is in hell? Like, is 20 in "Hey, not cool" territory, but it's all good once you get to 50? 40?

Rohrer says that when Angel asked if Cameroun was working or haunting the treatment center in heaven, the AI responded, “Nope, in hell.”

“They had already fully established that he wasn't in heaven,” Rohrer said.

Always a good sign when your best defense of the horrible thing your chatbot says is that it's in context.

I'm just going to pretend that's one of the researchers from Where Oaken Hearts Do Gather.

[–] FermiEstimate@lemmy.dbzer0.com 29 points 1 year ago (1 children)

I conclude that scheming is a disturbingly plausible outcome of using baseline machine learning methods to train goal-directed AIs sophisticated enough to scheme (my subjective probability on such an outcome, given these conditions, is ~25%).

Out: vibes and guesswork

In: "subjective probability"

view more: ‹ prev next ›