Elon Musk talked about the capabilities of AI. The examples he cites were AI putting words in your mouth, AlphaGo beating the top 50 human players at the same time, dealing with radiation on trips to Mars and the Brain to computer Neuralink.
Elon says the degrees of freedom to which Artificial Intelligence is able to apply themselves is increasing by ten orders of magnitude a year. It is not clear to me exactly what Elon means with this statement of improvement. He talks about a double exponential improvement. Exponential increase in hardware capability and an exponential increase in software talent that is going into AI.
A lot of the hardware used for AI is often not optimized for AI. A GPU is order of magnitude better than CPU for AI. New neural net optimized chips are an order of magnitude better than GPUs. There are many neural net optimized chips coming out soon.
AI puts words in your mouth in a convincing way.
AlphaGo can beat the top 50 human Go players at the same time with zero chance of them winning. I could not find a video of AlphaGo beating the 50 top human players. There is some analysis of a recent AlphaGo vs AlphaGo game.
Elon talked about generative adversarial networks. Those systems competed to make a better and better video reproduction. One system would try first. The other system would then find flaws and fix them. The first system would then work on flaws and fix them. Eventually you could not find the difference between the original and competitively created copy.
What are Elon Concerns About Health Risks of Going to Mars
Radiation risks of going to Mars in the worst case are like smoking on the way there. He believes he can improve the radiation protection so that radiation will not be a show stopper.
Solar wind comes at you following along the magnetic field lines. This means you will need all around protection.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Ahh yeah, that makes sense that forging would create smaller crystals out of a big one. There is always some ongoing debate on blade forums about which method produces a better blade- stock reduction or forging to shape. I always thought forging to shape is better because it conforms the steel grain structure to the shape you made (I think). Care to weigh in on stock reduction versus forging? In a bend test (vice clamping tip of blade between wood blocks) my bet would be that the forged blade outbends the stock reduction blade everytime…
Steel…is really complicated stuff (to us materials science guys). however, it can be abstracted too as bulk materials of relatively straight-forward (and reproducible) characteristics.
Single-crystal steels don’t remain “single crystal” very long when forged into things like knife blades. Indeed, the strongest and best (for the application) steels are just the opposite: nearly amorphous, with the tiniest crystal domains all thoroughly ‘mixed up’ (distributed) through a compliant (ironically, soft!) matrix.
It is the combined matrix strategy that delivers durability, toughness, and yet doesn’t short change ‘sharpen-ability’ and the ability not just to make a keen knife edge, but to keep it.
For instance, using pretty ordinary high-carbon, high-vanadium, high-molybdenum stainless steel, one can make darn-tough knifes. You can hack wood, pry things, they don’t get damaged. They also can be sharp enough to shave with. And the crystals are sub-micron.
GoatGuy
Divorce, eschewing material things, new GF, less time “alone” to read.
I wish Elon Musk was the man I want him to be haha. But, I am still glad af he is who he is precisely. And by the man i want him to be, I mean a guy who responds to me on twitter like “cool idea about the trombone dude..” or something. Just anything really
Hi GoatGuy! Was wondering what you thought about how SX superalloy monolithic crystal steel would compare against other stainless blade steels being used for pocket knives? Do you think monolithic crystal would be better in terms of blade edge retention than normal steels? Don’t mean to blast you with random questions from the sidelines, but I was just reading about how SpaceX will use inconel sx300 and then 500 for their raptor engines and starship parts and it sounded like really cool stuff.
Mat… More infastucture is needed before sending people to mars. Strategically placed reflective material on phobos to create more light and heat on mars. Like the idea i would rather have two dim bulbs in a room rather one. Tons of water preferably collected from our moon incase of an emergency. Designs for habitats that can bury themselves. Or be buried in a small crater as the meteor has already done the digging. To protect against wind that can rob away heat. To protect againsts radiation and to help insulate against the cold. Solar powered magnet field cables could also be placed on both mars moons to help cut down on solar radiation. And more.
Well here is my thing on experts. Up until I Believe it was the middle of the last century that the top economist world wide always went with what was then the most well thought out way of planning economies.
They simply believed that people would make the most logical best decisions on things…..
Let that sink in.
Finally! Someone with brains on here.
The Musk Fluffer worship has gotten WAAY outta hand.
How come you stopped reading SciFi?
To die from cosmic rays ripping into the colonists of Musk Eisley?
Musk Fluffers get upset when you say perfectly logical statements like that, btw. Be careful.
PS: As per GoatGuy, I like your avatar too. Unfortunately, most Millennials will have no clue about it. Wait! That’s a BENEFIT! Sorry!
Question: Was he tripping on acid when he said all of this?
Hey, anyone is fully justified in asking that very question given how he’s done it before.
Oh come on. That is very disingenuous. He was being asked the questions.
Elon Musk is not an expert on AI so I put zero trust in what he says on the topic.
It would have to be fusion powered of course. Plenty of ice available, a nice dwarf planet would provide resources for both fuel and building material. Maybe too much – wouldn’t do for the kids to get ambitious and disrupt the pastoral ambiance.
He’s running a business, plus F9/FH is only partially reusable. He will never charge $100/lb unless some other provider is willing to charge $101/lb to LEO.
If the ship was 100% reusable then the only real cost after that would be the fuel. It’s not like you can buy a rocket from SpaceX and run your own “airline” like you can with wide body aircraft from Boeing/Airbus.
Cheap or expensive space travel is only a matter of choice.
Annoyingly, the Oört cloud is so darn far away that Sol for all intents is just a really bright star. To put it into quantitative perspective:
500 AU (the nearer Oört objects)
… (¹⁄₅₀₀)² Sol’s light
× 1,360 W (earth)
→ 0.0055 W/m²;
Which is pretty darn low. Moonlight (Luna) on Earth, at a full Moon phase, directly overhead (minimum atmospheric extinction), runs about 0.0017 W/m². About 3.3 times Luna’s wan light.
Clearly not something one is going to collect with mirrors, concentrate, and power PV cells with. Its teeny-tiny small.
All Human far-reaching “living in the interplanetary clouds” business is either going to happen closer — where Sol’s massive output becomes a meaningfully useful power source — or will be nuclear fission/fusion powered. No other way about it.
Not trying to be a buzzkill, but people just tend to forget how distance-squared really taxes energy coverage.
Just saying,
GoatGuy
My guess is that the number of AIs in space will greatly outnumber the number of humans out there – given a few decades anyway. Colonies on Mars (or O’Neil type colonies near asteroidal mass) may provide a good refuge (or reservation) though for “Human” types.
I wonder if Musk’s vision of enhancing humanity by interfacing with computers isn’t just another variation of “AI” with its own set of dangers. A rogue superhuman intellect may just be far more dangerous than any AI we construct.
As far as the long term future goes “Humanity” will probably diverge into a multiplicity of different species, some with various degrees of integration with cybernetic enhancements and a myriad of species with strictly biological genetic engineered differences. Possibly parasites in an AI dominated construct.
If I live to see it (who knows?) I might just settle into a nice “Amish Style” Plain Jane “Pure Human” genetic preservation reserve somewhere in the Oort Cloud. Nice quiet lifestyle.
One of these days I’m just gonna have to get around to writing my novel.
PS: love your avatar. Of all modern Sci Fi movies, Fifth Element holds some position in the top 5 or 6. Along with Galaxy Quest a true comedic masterpiece, the first Star Wars 1977, The Martian, 2009 Star Trek, Alien, …
Cheap space travel is definitely up in the air. Musk’s premise is “by making the booster reusable, and using cheap compact uncomplicated fuels, and with deploying 21st century materials and manufacturing, we can make a rocket that’ll take stuff — in any amount — to LEO for less than $100 a pound”.
Revolutionizing, to be sure, this claim.
The rockets launched so far have missed that by a fair fraction.
$90,000,000 to get FH 63,800 kg into LEO … $1,400 per kg → $640/lb.
$90,000,000 to get FH 16,800 kg to Mars… $5,500 per kg → $2,500/lb.
The BFR / SHS is supposed to get the cost per kilogram down substantially. But you know… until we see it, for real, paid in, at $250/kg. … I’m just not going to suck up the hype.
Just saying,
GoatGuy
I feel your pain.
Dialog on the net has the growing ominous feeling of talking with raving lunatics or solipsistic philosophers (depending on the sophistication/politeness of your peer on the dialog), where no one can even agree on the most basic stuff that makes blue, blue and the water wet.
So effective our bubbles of information have become, that we only see what we want to see.
The themes of Science Fiction have evolved over the decades. Until perhaps the last 5 years, … strike that … 10 years, I was an avid SciFi reader and first-edition collector.
Fascinating to read the “old stuff” to see what the Masters of SciFi thinking had to say about the future then (present now), and the future yet-to-be. Equally fascinating to keep up reading the evolving SciFi compendium of memes and themes. Groundbreaking somewhat dystopian futurists such as Gibson, or Baccigalupi. Stewart, Aldiss, Asimov, Bradbury, Bester, Simak, Herbert, Harrison, Dick, Le Guin, Niven, Pohl, Foster, Bujold, Robinson, Stephenson, Scalzi, Doctorow, Melville.
Just saying.
If there are (say) 100 main themes in Sci Fi, then between them, they’ve been written over and over in various incarnations 25,000 times.
GoatGuy
Hard AI and affordable Mars travels?
Both need hard evidence to be believed. AI doesn’t seem to be converging towards general intelligence, but towards highly specialized, excellent pattern recognition. Current cloud based approaches join several specialized tools to make a broader one, but that’s not making an artificial sentience, just better services.
Mars (or cheap space) travels are still up in the air, depending if SpaceX or other delivers them. But this one at least is somewhat converging to the goal with every technological milestone, and we can very clearly see if there is progress or not (are there rockets that can do the job? are they cheap(er)?).
So, I’d reign my techno-rapture utopia dreams until we have the goods, or a very believable trend taking us there. On this, cheap(er) space trips are in much better shape than general AI IMHO.
Perhaps. (I get the reference to Old Codgers in every film.)
Perhaps I’ve seen enough in the last 6 decades to know that our dependence-on-technology is generally OK-to-good, but it also has and continues to breed a kind of mental myopia amongst not just the technorati, but also the Common Lass and even more Common Bloke.
It may be quite dark-hearted, but watch Black Mirror, the series. Its on Netflix (or Amazon), I don’t remember which. Quite the eye-opener, to what the modern dystopian film wright thinks may bode for the future. 1984 all over again, but even more insidious.
Perhaps I am beginning to miss the day when I could hold a conversation — in public — with most anyone, and use my regular vocabulary to have a lively two-directional chat. Without the recipient looking like I was speaking Martian or Klingon. The era of Tweets on Twitter has lead to a generation of verbal Twits, I am afraid.
And sure: our vocabulary changes all the time.
And I’m old.
But when all ideas are “fact checked” against the endlessly banal google-search algorithm, to be debated as false without any basis-argument, just references … well, long is lost the art of rhetorical debate. Might as well be communicating with opinion less AI’s.
Just saying,
GoatGuy
You should have signed that … and get off my lawn!
Won’t be long till AI will reach mars. Oh wait it will travel WITH the humans on the spaceship/ rovers smh…
Go to Amazon and look up a game called Olympica: The U.N. Raid on Mars, 2106 A.D. It was a game produced in the 70s and featured the Martian colonists organizing into a hive entity through computer-human networks, the individual members of which were called “webbies.” It was a prescient concept except that maybe the U.N. troopers should have been the ones trying to forcibly integrate the Martian settlers into their own collective.
Wow… such dependence.
The year The Twin Towers fell, Google was a search tool to find websites that might be of interest. YouTube hadn’t been invented. Google Maps hadn’t been invented. There was no über, Lyft, all that. People took taxis to get around, if they didn’t drive. I actually had a real, paper map (pack) in each of our family’s cars. They worked. One either learned the skills of reading maps, or relied on good old “its gotta be here somewhere” driving.
Anyone who drove knew how long — more or less — it was going to take to get between A and B. Was it a Big Game night at the Coliseum? (freeway congestion). Was it a big event in San Francisco? Blue Angels? The Garlic Festival? Was CalTrans fûqueing up the roadways again, to “repair them?”
Anyway.
Now we’re complaining that the Google “AI” isn’t getting trip-times accurate.
Woo hoo.
Cry me a river.
GoatGuy
Maybe that’s the answer.
To why we need to go to Mars.
Mars: AI-free.
See how attractive that is? The one place in the Solar System that is free of AIs, of any type or kind. The last refuge of Humans, should their AI’s go nuts, and take over the World.
Just saying,
GoatGuy
Google maps also makes wrong predictions for me but only when there is much traffic. I always take longer and some alternative routes even take double the amount of time (30 minutes instead of 15).
Hi, google maps again missed my arrival time by 2 times. Where is the AI if you cannot accurately predict 15 km ride?