SpaceX makes Mars spaceships and other space things, Tesla makes electric cars. Neuralink makes BMI (brain-machine interface).
SpaceX
This section is summarized from SpaceX’s Big Fucking Rocket – The Full Story, and How (and Why) SpaceX Will Colonize Mars.
The main story:
- Humans might go extinct. That's bad.
- To stay alive, humans should spread across the universe.
- The first step of space colonization would be to physically move some humans to another space rock like Mars, or the Moon, and build a living place there.
- To do that, humans need cheap big rockets to push things into space and get to the space rocks.
- Cheap big rockets are hard and expensive to make, so Elon Musk started a company to develop them.
- The company will develop rocket technology and get customers who would pay them money to send their stuff into space, and earn money to fund the Big Fucking Rocket.
- The BFR, and maybe better rockets, would send 1 million humans and a lot of other stuff to Mars, to build a self-sustaining colony there. Then the destiny of SpaceX will be done.
The BFR is made of two parts: a Big Fucking Spaceship on top of a Big Fucking Booster.
 |
| Note: the design has been updated in 2018, and would probably be updated yearly |
Fermi-style numbers:
- length: 120 m, about half in the booster and half in the spaceship.
- diameter: 10 m
- 1 raptor engine thrust: 2 MN
- total thrust of booster: 100 MN
- total mass: 4 kiloton
- maximal cargon payload: 100 ton
- maximal passenger payload: 100 humans
- cost per ton to Mars: $\$140k / ton$, according to slide 41.
How it gets to Mars and back:
Future timeline of BFR:
- suborbital test flights: 2019
- cargo to Mars: 2022 synodic period of Mars
- human to Mars: 2024 synodic period
BFR can also be used for rapid flight around the world. Imagine ICBM, but carrying humans, not nukes. It can go anywhere on earth within one hour.
Another project is Starlink: a constellation of satellites in LEO and VLEO that covers the entire earth with 4k-12k communication satellites. It gives cheap Internet (about $50 per month) over the entire world, with satellite receivers the size of "a pizza box". Details are here.
- bandwidth: 240 Tbps
- launch numbers: with BFR, 350 satellites per launch, 40 flights.
- cost: $10 billion
Timeline:
- 2018 to early 2019: Second set of test satellites put in orbit
- 2021 or 2022: First 800 satellites available to consumers.
- 2024 to 2029: Full system in place with 4,425 operational satellites
Hyperloop
Put small capsules in a long straight tube in near vacuum, and use electric induction on the bottom for propulsion. The capsules will float on a air cushion on the bottom. The tubes will be covered by solar panels for energy.
The main line would be between LA and SF. It would run a 80-minute loop of 35 min from LA to SF, 5 min at the SF station, 35 min back to LA, 5 min at the LA station.
Speed: about exactly speed of sound.
Cost of construction: $6 billion.
Cost of a one-way trip on the Hyperloop: $20.
Elon Musk et al published the design in a paper, Hyperloop Alpha, in 2013. There are now projects to build it.
Tesla
From the post How Tesla Will Change The World. The basic story:
- As mentioned before, humans need to turn away from fossil energy to renewable energy to avoid collapse of society.
- Cars are the biggest users of fossil energy.
- So humans should change to electric vehicles (EV).
- To make a good EV, they need to be cheap and fast.
- Build good batteries (gigafactory) and solar energy plants (solar city) and electric motors and recharge stations (superchargers).
- To get funds, start with building luxury cars, then use the money to build good cars, and finally use that money to build cheap cars.
- The fossil fuel industry and the fossil car industry are making fear, uncertainty and doubt of EV (and climate change), to save their profits, similar to how tobacco companies fought to make people not know that tobacco cause cancer.
This one is rather straightforward so I won't talk about the details, other than quote this picture which is supercool:
We clearly see that decreasing coal and petroleum consumption would do great to slash "rejected energy" (waste). Using EV and solar energy would accomplish that.
Also there's a joke somewhere that Tesla is realizing Edison's dream.
Neuralink, and OpenAI
Neuralink is Elon Musk's company for creating a Brain-Machine Interface (BMI). OpenAI is a non-profit company that does research on human-friendly AI and publishes openly. Elon Musk was a co-founder.
This post gave a thorough description of what the brain does, then sketched the current technologies of BMI, and gave some sketch of what BMI can be used for, and finally noted the motivation for doing this: to deal with AIG.
Basically, BMI is hard, very hard, and the whole post exhaustively details why it's hard.
Basically, BMI is hard, very hard, and the whole post exhaustively details why it's hard.
OpenAI is made to make sure enough people are working on AGI at more or less the same speed so that when it happens, it won't be a monopoly situation. This multipolar alternative, however, might not be better, see Nick Bostrom's Superintelligence (2014), chapter "Multipolar scenarios".
Neuralink is made as an alternative approach to AGI: instead of reconstructing AGI by engineering on silicon computers, just keep upgrading the human computers. This approach is considered better, because it allows the AGI to be an outgrowth from Human General Intelligence, so it's probably a more certain way to make a human-friendly AGI.
I don't have much to add, other than that compared to creating a silicon-based AGI, BMI might be too hard and would thus come a later than AGI, so that it probably doesn't matter, but still, a crude form of BMI should be possible before AGI comes.
Why is BMI hard? Hardware engineering. There will be enough computing power for AGI, and it'd be a matter of programming it, which can be done with typing with fingers. BMI requires some extremely miniature technologies that seem to be more technically challenging.
Philosophically, I'm not that concerned about winning the Darwin award for humans collectively. The obligate anaerobic organisms destroyed themselves when they terraformed the earth. Without their stupid collective suicide, humans would never have arisen.
Like these creatures back then, humans today are cosmically outstanding, but not necessary. Becoming a biological bootloader for another form of life, if it happens, will be just another chapter in the saga of the universe.
As such, the only motivation for me to develop a human-friendly future is trivially selfish: for a CelestAI that loves me as a true pony. All other philosophical arguments will be unconvincing for me.
Why is BMI hard? Hardware engineering. There will be enough computing power for AGI, and it'd be a matter of programming it, which can be done with typing with fingers. BMI requires some extremely miniature technologies that seem to be more technically challenging.
Philosophically, I'm not that concerned about winning the Darwin award for humans collectively. The obligate anaerobic organisms destroyed themselves when they terraformed the earth. Without their stupid collective suicide, humans would never have arisen.
Like these creatures back then, humans today are cosmically outstanding, but not necessary. Becoming a biological bootloader for another form of life, if it happens, will be just another chapter in the saga of the universe.
As such, the only motivation for me to develop a human-friendly future is trivially selfish: for a CelestAI that loves me as a true pony. All other philosophical arguments will be unconvincing for me.
No comments:
Post a Comment