Elon Musk’s problems are bigger and more important than yours. While most of us are consumed with our daily activities, Musk has been anointed by a Higher Power to save us all from ourselves.
He’s here to make sure that we eliminate traffic accidents, that traffic is a thing of the past, that we solve autism (his words, not mine), that we connect the human brain to machines , that we fill the night sky with satellites so that everyone can access the Internet and colonize Mars.
He’s not sure exactly how we’re going to accomplish all of these things, but he has more than enough money to turn every great idea he’s ever had into an industry that works.
Who cares if Tesla is 10, 20 or 100 years away from solving the driverless car problem? Financial experts are almost unanimous that $ TSLA is doing well with its current rise.
People who have the opportunity to invest in Neuralink will earn money until ELON will maintain the current trend going.
Never mind that the remote technology he claims will someday use the common BCI his company makes today and turn it into a magical telepathy machine in 2021.
The reality is, artificial intelligence can’t do them. things Musk needs to keep Tesla and Neuralink on their promises.
Here’s why:
- AI has a serious “mapping” problem that Tesla, Neuralink, Google, Amazon, Facebook, Microsoft, OpenAI, DeepMind, and the rest of the players in the field currently have no idea how to fix.
2. Elon’s money is useless here.
AI’s “mapping” problem
When we talk about a mapping problem, we are not referring to Google Maps. We’re referring to the idea that maps themselves cannot be made possible by a one-to-one representation of a particular area.
Every “card” automatically suffers from an enormous loss of data. In the “real” territory you can count every blade of grass, every pebble and every mud puddle. You only see a small representation of the distance on a map, maps are useful for getting directions but when trying to count the number of trees on your property or to determine exactly how many wolverines are in a nearby bush hide, they are pretty useless.
When we teach a deep learning system to “understand” something, we have to feed it with data. And when it comes to massively complex tasks, such as driving a car or interpreting brain waves, it is simply impossible to have all of the data. We just draw a tiny approximation of the problem and hope that we can scale the algorithms to the task.
This is the biggest problem with AI. This is why Tesla can use the Dojo to train its algorithms in millions, billions or trillions of iterations and give its vehicles a more driving experience than any human being who has ever existed together, it still makes inexplicable mistakes.
We can all point to the statistics and shout: “Autopilot is safer than unaugmented human driving!!” Similar to Musk, but the fact remains that people without an autopilot are much safer than Tesla’s full autonomous driving skills without a human.
Building the safest, fastest, and most efficient production car ever is an incredible feat that Musk and Tesla should be commended for, but that doesn’t mean the company will even remotely solve driverless cars or any of the artificial intelligence problems that plague the entire industry.
No amount of money will go to human-scale brute force algorithms, and Elon Musk may be the only artificial intelligence “expert” who still believes that deep learning-based computer vision is key to self-driving vehicles.
And the exact same problem applies to Neuralink, but on a much larger scale.
Experts believe that there are more than 100 billion neurons in the human brain. Despite what Elon Musk may have tweeted recently, we don’t even have a basic map of these neurons.
In fact, neuroscientists are still questioning the idea of regionalized brain activity. Recent studies show that different neurons light up in changing patterns even when the brain accesses the same memories or thoughts more than once. It happens when a person thinks of ice cream, the next time they think of ice cream, the old card could be completely useless.
We don’t know how to map the brain, which means we have no way of creating a dataset to teach the AI how to interpret it.
So how is AI trained to model brain activity? You fake it. You teach a monkey to press a button to summon food and then you teach it how to use a brain-computer interface to press the button, as Fetz did in 1969.
Then you teach an AI to interpret all of the monkey’s brain activity in such a way that it can tell whether the monkey tried to push the button or not.
Note that the AI is not interpreting what the monkey is trying to do, it is only interpreting whether or not to press the button.
So you need one button for everything. You need enough test subjects using BCI to generate enough generalized brainwave data to train the AI to perform any desired functions.
The equivalent of this would be if Spotify had to build robots and train them to play the actual instruments that are used to create every song on the platform.
Every time you wanted to hear Michael Jackson’s “Beat It” you had to send a training request to the robots. They picked up the instruments and began making absolutely random noises during thousands or millions of workouts. Hours until they “freaked out” something like “Beat It”.
When the AI changed its version of the song, its human developers gave their feedback to indicate whether they were getting closer to the original melody or further.
By now, a half-talented human musician could play the entire composition of almost any Michael Jackson song after just a few hours of listening.
Elon’s money is no good here
Robots don’t care how rich you are. In fact, Artificial Intelligence doesn’t care about anything because it’s just a set of algorithms that are combined with data to produce personalized results.
People assume that Tesla and Neuralink will solve the AI problem because they have essentially unlimited support.
But as Ian Goodfellow at Apple, Yann LeCun at Facebook and Jeff Dean at Google can tell you: If you could solve autonomous cars, the human brain or AGI with money, it would already be solved.
Musk may be the richest man in the world, but even his fortune doesn’t dwarf the aggregate value of the biggest tech companies.
And what the general public doesn’t seem to understand is this: Facebook, Google and Tesla and all the other AI companies are working on the same AI problems.
When DeepMind was founded, the purpose was not to win chess or games of Go. Its purpose was to create an AGI. The same is the case with GPT3 and almost all other multimodal artificial intelligence systems currently under development.
When Ian Goodfellow reinvigorated deep learning with his vision for neural networks in 2014, he and others working on similar challenges sparked a fire in the tech world.
Since then, we’ve seen billions of dollars worth of neural networks evolving, pushing the boundaries of computers and hardware, and even with all of that, we could still be decades or more away from self-driving cars or algorithms that can interpret human neuranal Activity.
Money can’t buy a technological breakthrough (it doesn’t hurt, of course, but scientific miracles take more than money) And unfortunately, Tesla and Neuralink, many of the world’s brightest and most talented artificial intelligence researchers, know that meeting Musk’s demands may be huge promises a futile effort.
Perhaps this is why Musk has expanded his recruiting efforts beyond seasoned AI researchers and is now trying to attract whatever computer talent he can find.
The good news is, no serious review can cool the mood of tireless Musk fans. Whether or not he can deliver the products does not affect the amount of worship he receives.
And that’s just as likely to change as Tesla’s ability to build an autonomous car or Neuralink’s ability to interpret neural activity in the human brain.
Source link
