DennisP

32 ms·

721.
by DennisP 2y ago
As I said, the combined reaction generates 5% of its energy as neutron radiation. The reason it's only 5% is that D-He3 is a much more energetic reaction. The main issue with D-He3 is that it's more difficult.
722.
by DennisP 2y ago
Plenty of people do want longer range though. Also, trucks could use longer range, and aviation definitely needs longer-range batteries.
723.
by DennisP 2y ago
Standard procedure in physics is to round to the closest number with the same number of significant digits. In this case, rounding down to 600, so people have a sense of how approximate the number is.
724.
by DennisP 2y ago
Yep, for D-T 80% of the energy is in neutrons. Helion is using D-D/D-He3, and for that it's about 5% in neutrons, and most of the rest in fast-moving charged particles. So they have a simple way to extract electricity directly. Th
725.
by DennisP 2y ago
It's a pulsed reactor by design, it's not supposed to create a "fully stable" reaction. They're actually pretty much on schedule, once you account for the several years it took them to get the necessary funding.
726.
by DennisP 2y ago
The original fuel is deuterium, which is absurdly abundant on Earth. Pure deuterium fusion results in half helium-3, and half tritium which decays to helium-3 with a 12 year half-life. Then their main reaction is D-He3. The D-D reaction is
727.
by DennisP 2y ago
Elites are not the only ones who get cancer treatments. Since the diseases of aging are extremely expensive, it's even likely that national health insurance programs would pay for anti-aging treatments. Longer lifespans would also help
728.
by DennisP 2y ago
The desire for immortality goes back at least as far as the epic of Gilgamesh. Medieval alchemists tried to achieve it. In China, Daoists attempted to dramatically lengthen their lives by various esoteric means. Tibetan Buddhism also has pr
729.
by DennisP 2y ago
Maybe. But you could have said the same about the smallpox vaccine, but look how that turned out.
730.
by DennisP 2y ago
I guess the idea would be to send FTL probes into space in all directions, so they can see the light heading our way from novas and report back. However, accomplishing this seems considerably more difficult than finding an underwater buoy.
731.
by DennisP 2y ago
The top answer addresses the butterfly effect, by using the time and position of novas, which are unpredictable.
732.
by DennisP 2y ago
Ingredients include carrots, cabbage, kidney beans, potatoes, dairy, and poultry, all blended up and baked. Sounds reasonably nutritious.
733.
by DennisP 2y ago
It's our default because for most technology, we have high confidence of a high p(survival).
734.
by DennisP 2y ago
This article says that we can't trust the estimates of p(doom), therefore we should take no action. But it assumes that "no action" means spending many billions of dollars to develop advanced AI. But why is that our default?
735.
by DennisP 2y ago
If the hardware is improving exponentially, then increasing the block size on the same curve is also an exponential improvement. Layer 2 is helpful but does not entirely remove the need for space on chain. In LN for example, every channel o
736.
by DennisP 2y ago
Depends on the mask. N95 protects you pretty well.
737.
by DennisP 2y ago
Well you can use intermediate nodes, it doesn't have to be a direct payment channel between sender and receiver. But to make that work, you have to be online in real time to receive funds. It's not like on-chain transactions that
738.
by DennisP 2y ago
Do you understand that the block limit was set fourteen years ago, and that bandwidth and storage have increased exponentially since then?
739.
by DennisP 2y ago
We know microgravity is terrible for us but we don't have data on 1/3 gravity. We might be fine. Or maybe we'd be fine if we took this drug, did some weight training, and added twenty minutes of centrifuge every day. For the
740.
by DennisP 2y ago
There's also someone building a factory to make iron-air batteries for grid storage. They're way cheaper than lithium or even sodium, with one of the most common materials on the planet, but only about 50% efficient. They're
741.
by DennisP 2y ago
But with several math errors, which make him off by a decade. He multiplies by 1.59 annually and gets a 25X increase in total battery in seven years. That actually takes eight years, but whatever. Then he says that's eight doublings, b
742.
by DennisP 2y ago
Part of the problem was the lawsuit from CoinLab, which took years to resolve. There wasn't anything like that with FTX.
743.
by DennisP 2y ago
Don't FPGAs have some overhead? How much better could this be on a custom ASIC?
744.
by DennisP 2y ago
The study itself was discussed here five days ago: https://news.ycombinator.com/item?id=40667786 One point of discussion was making satellites out of other materials that won't deplete the ozone.
745.
by DennisP 2y ago
Version 11 had lots of hand-coded behavior. Version 12 is entirely a neural network, trained on human driving. Partly it comes from people running FSD and intervening sometimes, and partly it comes from just passive observation of a people
746.
by DennisP 2y ago
They do plan to run their own robotaxis. But there are several million Teslas on the road already. They're just leaving money on the table if they don't make them part of the network, and doing so means they have a chance to hit c
747.
by DennisP 2y ago
There's no bandwidth requirement, it runs locally on the car.
748.
by DennisP 2y ago
Learning that sort of context is what neural nets are good at. The whole point is that none of this is hardcoded; the AI just learns what humans do in similar situations. At the point they're not even hand-labeling anything. The only i
749.
by DennisP 2y ago
FSD already understands some human gestures, and it learns this stuff by seeing how human drivers do things. Understanding a headlight flash doesn't seem like a big ask. And recently, the biggest complain from FSD users was that it dro
750.
by DennisP 2y ago
It's not enough for robotaxis yet, and Tesla doesn't claim that it is. They just think they'll get there. What they do claim is that with human supervision, it lowers the accident rate to one per 5.5 million miles, which is a

More ›