Although it’s been used for a fairly wide array of algorithms for decades. Everything from alpha-beta tree search to k-nearest-neighbors to decision forests to neural nets are considered AI.
Edit: The paper is called
Avoiding fusion plasma tearing instability with deep reinforcement learning
Reinforcement learning and deep neural nets are buzzwordy these days, but neural nets have been an AI thing for decades and decades.
Suppose one year you spend $60k, but only earned $50k. You lost $10k.
The next year, you spend $57k, and earned $53k. You lost $4k, and your losses narrowed by $6k.
Disney+ lost 1.3 million subscribers in the final quarter of 2023 amid a hefty price hike that went into effect last fall, but managed to narrow its streaming business’ losses by $300 million during the October-December period.
That doesn’t really sound like it backfired to me. They lost subscribers but made more money.
Oh, yeah.
If your point is that ICE car batteries have problems in the cold, so cold batteries is a problem for everyone and worse for ICE cars, that’s fair.
If your point is that ICE car batteries suck therefore EVs suck, that’s not really valid logic.
Lead-acid batteries aren’t lithium ion? And the car starter battery isn’t equivalent to that of an EV?
You might as well say that I have trouble starting my gas weed wacker, therefore cars are hard to start.
The main problem is just that getting a product from a one-off in a lab to a cost-competitive mass-market product is hard and can take a lot of time, to say the least.
For example, Don Sadoway initially published about a molten metal battery in 2009. He gave a Ted talk in 2012. They’ve run into assorted setbacks along the way and are apparently just starting to deploy the first commercial test systems this year.
It’s less that these breakthroughs are bullshit, and more that commercializing these things is hard. The articles about the breakthroughs are often bullshit, though, or at least way too rosy.
Keep in mind, though, AI progress is often more like punctuated equilibrium.
Each new approach gets you much further, and polishing each approach gets you slight improvements until the next approach comes along. Improvements to chatgpt might plateau until the next big breakthrough architecture. Or maybe not.
Levelized cost averages the fixed costs over the lifetime of the generation
They’re generally comparing utility scale installations, not home rooftop solar though.
The problem is when you mention Nazis to Jews, the first things they’ll think of are Zyklon B, Babi Yar, the piles of children’s shoes at Auschwitz, Mengele, that sort of thing.
And while what’s going on in Israel is terrible, it’s not Mengele terrible, Babi Yar terrible, or Treblinka terrible. So they write you off as just another antisemite, rather than listening to your point.
The problem with solar is that the sun doesn’t shine overnight. The good thing with that is that we use much less power overnight than we do during the day.
If you’re relying a lot on solar, you need to build a big-ass battery that you charge during the day and use at night.
Alternatively, you build a nuclear or gas plant sized to overnight usage and run them 24/7. Then, you build way smaller batteries to handle dispatchability and smoothing demand over the course of a day. Nuclear is good for baseline power, and doesn’t come with the environmental costs of a gas plant. It has a niche.
Yeah. Power plants are nowhere near 90% efficient.
It’s worth emphasizing, though, that they’re still way, way more efficient than car engines are.
Also, regenerative breaking saves a lot of energy. Basically, instead of using the motor to increase the cars speed, you use it as a generator to recharge the battery.