We are at the end of an era, in a much grander scale than you’re probably imagining.
As I often do, I dove deep into a Wait But Why article. They usually take me an hour or so, but this one–about what Elon Musk is building with Neuralink, and more importantly, why–took me three hours. Everything Tim Urban writes is insightful, but this one changed the way I see the future. It hit me so hard that I decided to distill the post’s 38,000 words into about 3,800–still quite a read–but well worth it. I still recommend the full piece with funny illustrations only Tim can offer, but if you’re pressed for time, I hope you find this summary eye-opening.
We’ll start with the basics, with the foundational elements of this piece touching on the way human knowledge grows, the way we communicate, and the tools we use for doing so. Next, we’ll cover how progress happens, how Elon Musk builds companies, and examine where those things merge today around brain technology.
That’s where the fun begins. We’ll get into both fun and scary applications of it that are likely to be ubiquitous within the next few decades, and uncover a potential threat that we probably aren’t taking seriously enough. And finally, we’ll conclude with how Neuralink, and in fact all of Elon’s (pre-Twitter) companies, are really about helping maximize humanity’s prospects for a bright future.
Let’s dig into it.
1A) The way knowledge grows
- Thought experiment #1: Humans today are not smarter than humans of 50,000 years ago. If you took a baby from 50,000 years ago and raised them today, they’d be fine. Likewise, if you put a baby human born today in a group of chimps and ask them to raise him, that human as an adult will know how to run around the forest, climb trees, find food, and masturbate. That’s who each of us actually is. (In case you’ve ever wondered by you’re so unimpressed by humans and yet so blown away by humanity’s achievements, remember this).
- Humanity, on the other hand, is super intelligent. The invention of language allowed each human to brain dump what they learned onto a pile, which got bigger and bigger. The Human Colossus (Tim’s term for collective humanity) has been inventing things ever since, optimizing over time, driven by the desire to create value. Humans today aren’t smarter; they are just starting from a place of the most advanced knowledge of the generation before them. This is an increasingly steep escalation; our collective knowledge is currently moving at an unprecedented pace, which is why we live in an unprecedented, completely anomalous time in history.
1B) The way we communicate
- Humans started developing language, which is technology, about 100,000 years ago, with fully developed language systems around 50,000 years ago. Consider this a huge “aha” moment that enabled knowledge growth.
- If you take a caveman from 50,000 years ago (or even 200 years ago), they would be utterly shocked by how we live: airplanes, submarines, skyscrapers, the Internet, iPhones. These would all be incomprehensible; our world has become magical.
- But how we communicate, through language, has largely stayed the same. That caveman would wonder why we haven’t come up with a super cool, incomprehensible way to communicate in the same way we have developed amazing technology for how we live, work, and travel.
- Before language, there was no good way to get an idea from your brain into mine. Language transformed vocal cords and ears into communication devices, with air being the first communication medium, and we have progressively added layers.
- We can group this into clear eras of communication:
- Isolated brains → language → writing → telephones → Internet → mobile/ wearable devices → ?
- Isolated brains → a bunch of communication hacks using different kinds of middlemen → ?
- Era of isolated brains → era of indirect brain communication → era of ?
- The era of indirect brain communication lasted 100,000 years, and we are in the very last moments of that era. We are now living on the dividing line between Era 2 and Era 3, and the “?” will be digitally communicating brains.
1C) Our communication tools
- Tim goes into a lot of detail on how our brains are constructed, far beyond my ability to comprehend. But the key takeaways are:
- We don’t have great knowledge of the brain; if full knowledge of the brain is a mile, we’ve made it about 3 inches so far. (But here’s a cool short little video on how our brains work).
- Our brains have many complex parts. For the sake of understanding this article, we’ll group these into two: a primary limbic system that controls our basic animal survival functions (breathe, eat, shit), and a secondary advanced cortex that can process more complex thoughts (ability to reason).
- Our cortex is a thin, outer layer that would be about the size of an unfolded dinner napkin if spread out.
- (Not critical for understanding, but cool to know: our brains are wrinkly in order to provide more surface area for brain stuff (neurons). We have limited real estate in a head that has the size constraint of being born through a vagina; the wrinkles are nature’s way of accommodating that).
- Technology already exists to attach onto the cortex as a kind of tertiary layer (google Cochlear implant if you like). These brain-machine interfaces, or BMIs, are somewhat primitive today due to a variety of hurdles (it’s not easy to get inside the brain, for one). But more clever solutions are in the works, including thin layers of silk, mesh or ink that are lined with electrodes and fit our cortex like saran wrap.
- In the communication example from (1B), your phone is as much “you” as your vocal cords, eyes, or ears. Each is simply a tool to move thoughts from brain to brain. Does it make a difference if the tool is held in your hand, your throat, or your eye sockets?
- Because of this, we are already cyborgs in a sense. What makes a cyborg a cyborg is their capabilities, not from which side of the skin (or the skull) those capabilities are granted.
- Our external devices are essentially an extension of our brains; a tertiary layer that currently lives in our hand, pocket, or on our desk. This can eventually live in our brains, as BMI technology improves; In Era 3, we won’t need devices anymore.
- Our communication speed is limited when human elements are involved. We can only communicate as fast as we type, talk, move our thumbs, etc. This is why computer-to-computer communication is much faster in comparison. Era 3 will remove this limitation, making peer-to-peer communication as fast as a computer (your brain → my brain without any intermediaries).
All good so far? Good. Now we will start stacking concepts onto these foundations to build a deeper understanding.
2A) How progress happens
- Repeat from (1A): the Human Colossus is always driven to create value. Where opportunities to profit exist, innovation will happen. Just like teaching a dog to sit involves correlating sitting on command with a reward, if you can create rewards in whatever context, humans will chase after them.
- Typically, this involves a few innovative pioneers succeeding (Nikola Tesla, Henry Ford, the Wright Brothers, Alan Turing), then everyone else jumps on board to make progress go faster.
- Humans first created machines to take over physical tasks, essentially outsourcing the work of our hands and legs. The computer represented our first attempt to outsource the work of our brains. The first computers were memory upgrades, remembering machines, which we upgraded into processing machines, which we are now working feverishly to evolve into thinking machines.
- And so, a new colossus has been born: the Computer Colossus, a worldwide web that allows isolated computers, or “machine brains,” to communicate with each other.
- If there’s one thing we should learn from the past, it’s that there will always be ubiquitous technology of the future that’s inconceivable to people of the past (think about the iPhone in your pocket as compared to the far less powerful Pilot ACE computer Alan Turing built in the 1950s that filled up a room).
2B) How Elon Musk builds companies
(Important note: this article was written about Elon’s companies in 2017–so let’s temporarily forget the ongoing Twitter fiasco and his fraught 2022 and focus on his virtues as a visionary. The key takeaways still apply).
- Elon built Tesla and SpaceX for a greater purpose than most entrepreneurial ventures; he built those companies to spur humanity into action toward a better future by creating the incentive referenced in (2A), in the form of a ripe new industry exploding in both supply and demand.
- His model looks like this (credit: waitbutwhy.com):
- Let’s make a funny analogy: Elon saw the Human Colossus dog continually peeing on the floor in the form of adding deeply-buried carbon into the carbon cycle (overdependency on fossil fuels).
- Rather than pleading with the dog to stop peeing on the floor (which a lot of people waste their breath doing), or threaten the dog (which governments try to do), Elon created an electric car so rad that everybody wanted one. Since 2008, when the first Tesla came out, the number of major car companies with an electric vehicle went from zero to pretty much all of them. Reduced dependency on fossil fuels → better future.
- With SpaceX, Elon isn’t trying to convince humanity that it’s wise to build a civilization on Mars as life insurance for the species. Instead, he is creating an affordable cargo & human transport route there, creating enough reward potential to spur the human colossus into action (think real-estate development, tourism, etc.).
- Tesla Model (credit: waitbutwhy.com):
- SpaceX model (credit: waitbutwhy.com):
- He’s applying this same model to Neuralink, which we will get to a bit later.
2C) Where we are today and (immediate) tomorrow
- Brain Machine Interfaces are already able to help with many disabilities; efforts so far have mostly focused on restoring lost functionality. This is in large part due to where demand is, and that human experimentation is only possible for people trying to alleviate an impairment.
- In Alzheimer’s patients, for example, memories themselves are often not lost, just the bridge to those memories. BMIs can help restore this bridge or serve as a new one.
- A limiting factor is bandwidth; as this improves perfect vision or hearing (among many other abilities) will be restorable.
- Other first applications will likely be to help with mood disorders. Whereas pharmaceuticals enter the brain and spread out randomly, neural interfaces will be able to isolate and stimulate just one area, can be tuned in real-time, and can carry information out about what’s happening inside.
- This era is starting with disabilities, but eventually BMIs will begin to emerge which people without disabilities want. The early adopters will likely be rich, but this was true of cell phones too. In 1983, a two-pound cell phone cost $9,000 in today’s dollars, and now well over half of living humans have a much better one.
Still with me? Now let’s get to the mind-bending, sometimes fun, sometimes terrifying stuff. In the very near future, we will essentially be able to use our thoughts as a remote control (a company called BrainGate is already enabling paralyzed patients to control a cursor, or even play a video game, with their thoughts).
3A) Here are some super fun applications that are pretty hard to fathom today.
- A deeper understanding of each other: Emotional communication is lossy… we lose a lot in translation. Ten people may describe feeling “sad,” despite experiencing ten different things. Brain-to-brain, feelings can be uniquely understood.
- Sensory communication #1: An NBA player will be able to invite you to be able to see and hear through his eyes and ears while on the court. A surgeon will be able to control a machine scalpel with her motor cortex instead of holding one in her hand; she could receive sensory input from that scalpel, making it feel like an 11th finger and giving her finer control over incisions. Or a rookie surgeon can have a mentor watch her work through her eyes and think instructions or advice to her. If something goes wrong, they’ll be able to “take the wheel” and connect their motor cortex to her outputs and take control of her hands. The porn industry will explode. The NFL may be able to phase out brain injuries by keeping biological bodies on the sidelines controlling the motor cortex and experiencing the eyes and ears of an artificial body that is either uniform or identical to that physical player. And all of this is without the need for screens; your visual cortex is the screen.
- Sensory communication #2: think about why we binge on sugary candy. We have a battle raging in our heads between our cortex and our limbic system; both are trying to serve us well, but our limbic system is stuck in a tribal era 50,000 years ago. It makes you binge candy because it thinks anything that sweet and chewy must be rich in calories and that we might not find food again for 4 days, so it’s smart to load up now. A good brain interface will be able to decouple our sensory desire of eating from the survival purpose (nutrition), enabling us to experience eating chocolate cake while putting something healthy into our bodies.
- Pain removal: pain is simply the body’s way of telling us we’d better check something out. Via BMIs, we will be able to choose to get that information in less unpleasant formats.
- Supercharged intelligence: and, of course, the idea of googling something will be so seamless that it won’t even feel like it’s not within our own natural capability.
3B) New technology always comes with real dangers and harms a lot of people.
- Hundreds of thousands of people have died from plane crashes, and millions more from cars, but if we had the chance to go back to horseback and steamship, would we? Despite the to-be-expected drawbacks, advances in technology almost always turn out to be a net positive.
- Trolls who’ve had a field day since the Internet will have an even greater ability to disrupt by literally getting inside our thoughts.
- Computers have bugs, and they crash, and you can CTRL-ALT-DEL to restart it, or just buy a new one. You can’t get a new head.
- Computers can be hacked, and this is absolutely terrifying. It’s one thing to consider a hacker stealing information from your brain, but BMIs can also put information in. A clever hacker may be able to change your thoughts, your vote, your identity, or make you do some other crazy thing you don’t want to do, without you being aware.
- There will be terrible catastrophes, but again, to perhaps lesser degrees of fright, this has been true of every technological advancement. A much larger group of good guys will fight to keep bad actors at bay, a giant “brain security” industry will emerge, and if given the option, people in this future era will likely not for a minute consider coming back to today.
3C) But, we are vastly underestimating the threat the Human Colossus is creating toward itself with Artificial Intelligence.
- Intelligence has given humans godlike powers over all other creatures on Earth–which has not been a good outcome for those creatures. We’ve built major industries processing and selling any creature’s body parts that are possible value creators, and we sometimes kill them for sport. Being selfish is what species do, but the great danger to other species comes not from our selfishness, but from the tremendous power we have over them. This power comes from our intelligence advantage.
- And now, we are teaching the Computer Colossus to think, intentionally creating something that will have (already has) far more intelligence than we do. We’ve never done this before, making us amateurs at it. There is a significant risk that either a small group of bad people are able to monopolize AI power, or AI goes rogue. Scary stuff.
- The Human Colossus is not going to quit making AI smarter; there is too much value to be created. Warnings that we shouldn’t build superintelligent AI won’t be heeded, just like we said earlier that you’re wasting breath telling the dog to stop peeing on the floor.
- Remember from (2A) how a few innovators typically jumpstart the Human Colossus? Well, that’s one thing if the innovation is an automobile. But today we are talking about brain control, a literal magic wand that could give whoever owned it unchecked power over everyone, including the power to prevent anyone else from making a magic wand.
- That first innovator could be someone good. It could just as likely be someone bad. Leaving that to chance is not a risk in anyone’s best interest.
To wrap this up, let’s put it all into perspective. We are currently on the edge of more history timeline boundaries than just the transition into digitally communicating brains outlined in (1B). After 1,000 centuries of human life, this century will be the one where we become multi-planetary, and the one where we manage to wrestle genetic code from the forces of evolution so it can reprogram itself. People living today may actually witness the moment biotechnology frees the human lifespan from the will of nature (our physical body decomposition). Quite simply, we’ve built an entirely new level of power–powerful enough to overthrow Earth’s 3.8 billion-year history.
This puts us on a superhighway towards unimaginable change whether we like it or not, and generations alive today, whether they realize it or not, are the first in history to face real existential risk. But these generations could also be the first with an actually realistic hope of a genuinely utopian future. We need to quickly figure out what the “unimaginable” is, and that’s where Elon Musk comes back into play. Which of these outcomes we’re headed for is not at the whim of chance, but at the whim of the Human Colossus, and depends on what we do over the next 150 years, or maybe just the next 25 years.
4A) Elon is all about trying to prepare us for that “unimaginable,” regardless of what it turns out to be.
- Since we can’t stop people from trying to make the magic wand, Elon, Sam Altman and others have built OpenAI as an open-source research lab. The idea is that, while a legitimate concern that bad guys can also benefit from collective knowledge shared by the lab, it will ultimately boost the efforts of millions of other people trying to create magic wands, thus making it less likely that any one inventor can achieve it long before others do.
- This increases the likelihood that whenever someone does finally create one (which Tim affectionately refers to as a “Wizard Hat”), there are thousands of others near completion as well, decentralizing the power and reducing the risk of an ill-minded, all-powerful sorcerer. And keep in mind, these would be different wands, with different capabilities, made by different people, for different reasons, making it more likely that the world’s aggregate mass of magic wands would truly represent the needs of the masses.
4B) This “needs of the masses” sounds a lot like democracy, by definition.
- Abraham Lincoln referred to democracy as “government of the people, by the people, for the people.” It is imperfect because “the people” are imperfect; it can be unpleasant because “the people” are often unpleasant, but decentralized power is in everyone’s best interests.
- OpenAI is about addressing the “by the people” and “for the people” parts of Lincoln’s equation.
- This leaves the “of the people” to cover, which we’ll do in the last section on Neuralink.
4C) Neuralink is, like Tesla and SpaceX, ultimately about increasing our chances for a better future.
- The Human Colossus is creating superintelligent AI for the same reason it created cars, factory machines, and computers–to serve as an extension of itself to which it can outsource work.
- Computers that can think will be our greatest invention yet, enabling us to outsource even important, high-impact work. But unlike those other machines, which are programmed with narrow intelligence, we are essentially building a ridiculously smart thing, far smarter than us.
- As history reminds us, when there’s something on the planet way smarter than everyone else, it can be a bad thing for everyone else. If AI becomes the new thing on the planet that’s way smarter than everyone else, and it turns out not to clearly belong to us, it means that it’s its own thing. Which drops us into the “everyone else” bucket. The possibility of AI becoming uncontrollable is likely a far greater threat than even bad people getting control of AI. It may not happen, but it could.
- We are novices at this; nobody has ever built superintelligent AI before. Since we don’t fully understand what kinds of AI will be around when we reach the age of superintelligent AI, we need to be prepared for any scenario.
- Neuralink is working on the idea that the best way for us to be prepared to fight AI in the event of a bad scenario is for all humans to be able to become AI. It is about having human-AI integration for the precautionary protection of our species.
- If everyone is Superman, it means that one rogue or bad-acting Superman can only do so much damage. If we are successful, the near-future world will have millions of people controlling little pieces of the world’s aggregate AI power. We will be much less vulnerable with the ability to think with AI, defend ourselves with AI, and fundamentally understand AI because of our own integration with it.
- Communication bandwidth is the key factor in determining our level of integration with AI, and our pace of progress in this direction matters a lot. Elon believes it will be catastrophic to develop superintelligence too far before we are able to do a merged brain-computer interface.
I hope that gave you something to wrap your head around. I’ll let Tim bring it home:
The reality is that we’re whizzing down a very intense road to a very intense place, and no one knows what it’ll be like when we get there. A lot of people find it scary to think about, but I think it’s exciting. Because of when we happened to be born, instead of just living in a normal world like normal people, we’re living inside of a thriller movie. Some people take this information and decide to be like Elon, doing whatever they can to help the movie have a happy ending—and thank god they do. Because I’d rather just be a gawking member of the audience, watching the movie from the edge of my seat and rooting for the good guys.