We don’t Need Super Intelligence for AI to be Super

In The Impossibility of Intelligence Explosion, François Chollet writes:

“Arguably, the usefulness of software has been improving at a measurably linear pace, while we have invested exponential efforts into producing it. The number of software developers has been booming exponentially for decades, and the number of transistors on which we are running our software has been exploding as well, following Moore’s law. Yet, our computers are only incrementally more useful to us than they were in 2012, or 2002, or 1992…In this case, you may ask, isn’t civilization itself the runaway self-improving brain? Is our civilizational intelligence exploding? No. Crucially, the civilization-level intelligence-improving loop has only resulted in measurably linear progress in our problem-solving abilities over time.””

Using technological advancement as a corollary for “civilizational intelligence”, if the number of transistors is “exploding”, this means our intelligence is exploding as well, because it is increasingly difficult to maintain Moore’s law, yet we have. The amount of progress since 1750, is equivalent to the amount of progress from 12,000BC to 1750. Saying “our computers are only incrementally more useful” reminds me of Peter Thiel in his famous “What Happened to the Future” manifesto coining “We wanted flying cars, instead we got 140 characters” while also, might I add, touting Facebook, a company he is an investor in, as an example of true technological progress. Using the smartphone as an example, the advances of these devices are far more than simply “incrementally useful”, even since 2012.

“A smart human raised in the jungle is but a hairless ape. Similarly, an AI with a superhuman brain, dropped into a human body in our modern world, would likely not develop greater capabilities than a smart contemporary human. If it could, then exceptionally high-IQ humans would already be displaying proportionally exceptional levels of personal attainment; they would achieve exceptional levels of control over their environment, and solve major outstanding problems— which they don’t in practice.””

Yes, they don’t in practice because many with a high IQ, among other inhibitors, have difficulty with social interaction, something computers don’t need to worry about. Even if we can only replicate the intelligence of an average brain, this would be a huge breakthrough. Just look at all those things on your TODO list. What if you had 0 distractions, were always motivated, never had to sleep, eat etc. The amount you could get done would be amazing. And then what if we duplicated that average brain as much as we wanted, wouldn’t that be super?

Big Tech is Eating the World

In 2011 Marc Andreessen declared “Software is Eating the World”. It is now evident that the software doing the eating is owned by a smaller and smaller set of companies. These Big Tech companies (Alphabet, Apple, Amazon, Facebook, and Microsoft) are backed by serious infrastructure, profits, data, custom silicone, ecosystems, and are increasingly vertical. You argue Big Tech is just like any other powerful sector with a lot of big players. My point is that this is different because Big Tech isn’t limited by vertical, it is limited only by the problems software can solve, which is an ever increasing amount of problems. The Google reorg to Alphabet is interesting because it is an admission of Big Tech’s future, which is to do everything. Software is so flexible and so powerful, that it is giving the large companies who specialize in it an unprecedented amount of wealth and power. We call them “tech companies”, soon that will be like saying “color tv”, because all companies will be tech companies.

Think about it, a large tech company has the power to make a large portion of the human population believe the world is at nuclear war. Facebook and Google could simply promote fake news stories. Apple could spoof text messages to their devices and ensure they can only access websites promoting their fake news stories. With how much internet traffic run through Amazon and Microsoft, they could easily spread news of a nuclear war that doesn’t exist. Am I afraid this is going to happen? No. Am I worried this is even possible? Yes. What’s scarier though is the fact that two tech companies, Facebook and Google, own algorithms that shift the thinking of billions of people. I don’t care who owns the keys to that kind of power, that kind of power should not exist.

This sounds crazy, but just 10 years ago you would have thought it crazy for the company behind the iPod to be building a car, an online book retailer to earn an Oscar, a search engine to be solving aging, and your your mom to be on Facebook. It used to be every tech company could grab their own piece of the pie, now it’s about who is going to own the pie shop. Ok Google take search, Apple take laptops people actually like using, Microsoft own the enterprise. Now Big Tech is competing fiercely with each other on multiple fronts, squashing or purchasing smaller players, and making component makers like Qualcomm, Intel, and Nvidia obsolete. But now, these Big Tech companies are breaking out of their bubble. As I alluded to above, there is no such thing as a “tech problem”, there are only problems. Big Tech has chosen to eat media as an appetizer, transportation looks to be next, and I predict banking will come after that.

Big Tech has three possible outcomes. The first, is they get broken up with antitrust legislation. The second, a paradigm shift e.g. blockchain, that levels the playing field. The third, Big Tech keeps getting bigger and bigger until Big Tech is only one company, which we worship as our God and would never dare wish ill on it.

Clark Goes There

You can’t make this up, a Marxist who attempted to join the United States Marine Corps. Clark is a friend from high school and one of the most interesting people I know. He has thoughts on everything. Sit down, strap in, and get blasted by the Clark experience as he talks movies and tv, the Marines, Marxism, politics, and religion. Like I said, he goes there.

31:30 Southern California, not a lot of hicks or rednecks to burn Qurans with so you gotta hide yourself. But you can be proud of how Republican you are amidst all the Democrats. Because you see you’re special. You’re Republican. You’re important. Everybody else, they’re just stupid. Those academics, those poets, virtually every person, every artist, everything that doesn’t have to do with the CMAs, they’re all stupid.

34:00 Fantasy was part of my religion. I remember going to church down at St. Francis and Jesus Christ just dying on that cross, that’s not a hero I wanna read about. I wanna read about the dark elf Drizzt Do’Urden, fighting vampires and demons and all sorts of stuff.

59:30 The last Marxist that went through the Marine Corps was Lee Harvey Oswald. So I definitely tried to keep that silent.

Snap Inc. and the Brave New World of Internet Centralization

Snap Inc., one of the highest valued tech IPOs of the last decade, is built on cloud infrastructure from Amazon and Google. It has committed $2 billion to Google Cloud Services and $1 billion to Amazon Web Services over the next five years. This is a tech company that has outsourced a large portion of their tech and it makes me fear for the future of the internet. I don’t blame Snap at all, Google Cloud and AWS are excellent products that has allowed Snap to reach millions of users without significant up front costs associated with building out an entire backend infrastructure form scratch. It is also very possible that like Dropbox, Snapchat will wean themselves off an external infrastructure and take it all in house. But I suspect that over the course of the next decade it will become increasingly difficult from both a technological and cost perspective for companies to build their own infrastructure that can compete with Google Cloud, AWS, and Microsoft Azure. On top of this, solutions like Google’s Firebase, which I expect to become even more popular and even more advanced, ensure the switching costs are even greater because these types of solutions supply more than just computing primitives like storage or computation. These types of solutions force your software to comply with a very specific API, making switching infrastructure even more difficult. Does this mean we will be left with three companies that control the vast majority of internet traffic? I think yes. Though there are other companies like Cloudflare that are increasingly expanding their infrastructure, economies of scale and the momentum behind Google, Amazon, and Microsoft make me think this game is already over. The only way unseat these behemoths is to change the game. Maybe it’s blockchain, maybe it’s something else. One thing is for sure, it’s a brave new world.

Trust to Trustless

Humanity has undergone 3 major paradigm shifts in transacting. Bartering was the first paradigm shift, allowing humans to trade one good for a different good. The second made trade even easier, allowing all goods to be assigned a value using a standard measurement, i.e. money. The third shift allowed trade to occur without the two parties ever having to meet, by trusting a middleman. This allowed for cross continental trade and allows us today to order millions of products on Amazon. With the arrival of blockchain technologies like Bitcoin and Ethereum, we are currently undergoing the 4th major paradigm shift. Where Amazon and Visa are single points of failure, blockchain offers what I call a level 1 trustless system. There is no single point of failure that we must trust, the nodes of the network ensure consensus, without even having to trust each other. Level 1 trustless systems can still fail. Though difficult, it is possible for malicious actors to compromise the network. There are also humans in control of these projects that act as another point of weakness. The next paradigm shift will be to a level 2 trustless system, which realistically cannot be compromised by a malicious actor. A level 2 trustless system does still rely on humans to steer the direction of the project. A level 3 trustless system is what I consider to be fully trustless. That is, malicious actors cannot compromise the network and there are no humans in control of the direction of the project.

First they Came for the Backend Developers

Parse showed us what the brave new world of serverless development looked like. Before that, using tools like AWS allowed developers to build without managing their own servers. And before that, Joe in dev ops was on call to restart the company server if something went wrong. Parse offered an SDK that offered push notifications, authentication, and storage. Facebook bought them and shut them down. Google bought Firebase and picked up the torch. Albeit some pain points at the beginning, they have kept iterating. This week they announced Cloud Firestore, addressing a pain point around Firebase’s limited Realtime Database. It is easy to see how this can be the future. It makes development much easier, handling scaling and offering several turnkey solutions to common problems. Realm offers a similar product which doesn’t quite have all the features Firebase does. Even with cloud computing solution like AWS, backend developers are needed to get everything wired up. With solutions like Firebase, backend developers are not needed. Frontend developers and native developers, it’s only a matter of time before Firebase solves that problem too. By 2023, any web app or native app that only needs to parse simple data, without doing anything technically special, will be able to be built by someone with limited technical expertise.

Google, All in or Fall in

Benedict Evans, writing in his newsletter:

“Google paid $1.1bn for part of (struggling) HTC’s smartphone team (~2,000 people) plus some non-exclusive IP rights. It remains very unclear how far Google wants to go into making its own phones. Android has lost the greater part of the high-end to Apple, and that seems unlikely to change.”

What do you mean it is “unclear how far Google wants to go”? They just spent a billion dollars. Google poached one of the top Apple chip architects in June. The Pixel phones are priced identically to the iPhone. Not only is Google serious about making their own phones, it is only a matter of time before Google starts building their own chips for those phones. iPhones are great because the software and hardware are built together. There is no equivalent of this on Android. It’s why Apple’s chips blow the competition out of the water. Especially as phones start doing more specialized tasks like AR/VR, machine learning, etc. Chips need to be purpose built to optimize for these use cases. Software is eating the world and now hardware is eating software. As for Benedict’s point that Apple’s dominance in the high end “seems unlikely to change”, I agree. The battle may already be over and Apple very well could own the high end phone market until the next paradigm shift in computing. But on the eve of Google’s largest hardware event ever, let’s see what they got.