The course of lectures "Startup". Peter Thiel. Stanford 2012. Lesson 19

Original author: Blake Masters
  • Transfer
  • Tutorial

In the spring of 2012, Peter Thiel ( Peter Thiel ), one of the founders of PayPal and the first investor FaceBook, held a course at Stanford - "Startup". Before starting, Thiel said: “If I do my job correctly, this will be the last subject that you will have to study.”

One of the students of the lecture recorded and posted the transcript . In this habratopika, mg1 translates the nineteenth lesson, astropilot editor .

Lesson 1: Challenging the Future
Lesson 2: Again, Like in 1999?
Session 3: Value Systems
Session 4: The Last Step Advantage
Session 5: The Mafia Mechanics
Session 6: The Law of Thiel
Session 7: Follow the Money
Lesson 8: Presenting an Idea (Pitch)
Lesson 9: Everything Is Ready, But Will They Come?
Session 10: After Web 2.0
Session 11: Secrets
Session 12: War and Peace
Session 13: You Are Not a Lottery Ticket
Session 14: Ecology as a Worldview
Session 15: Back to the Future
Session 16: Understanding Yourself
Session 17: Deep Thoughts
Session 18: Founder - Victim or God.
Occupation 19: Stagnation or Singularity?

During the classes, three guests joined the conversation:
  1. Sonia Arrison , Technical Analyst, author of 100 Plus: How the Coming Age of Longevity will Change Everything , and co-founder of Singularity University
  2. Michael Vassar , futurologist, former president of the Singularity Institute for the study of Artificial Intelligence (SIAI)
  3. Dr. Aubrey di Gray , expert in gerontology and director of science at the SENS Foundation .

Lesson 19 - Stagnation or Singularity?


I. Prospects


Peter Thiel : Let's start by saying that each of you will outline your vision of what technological changes await us in the next 30 or 40 years.

Michael Vassar : It’s much easier to talk about how the world will be in 30 years than what it will be in 40. A period of thirty years seems visible. Today, we have gone from how to calculate one or two genes to calculate the complete human genome, and it costs thousands of dollars. Paul Allen is currently conducting a $ 500 million experiment that seems to be going well. The path taken by technology is both breathtaking and scary.

Imagine that in 30 years we will have computing power that exceeds modern millions of times and whose algorithms are a hundred times more efficient. At that moment, we will come closer to the fact that we can simulate the brain and all that. And after that, everything will be wrapped.

But such progress over the next 30 years is by no means something that can be taken for granted. Overcoming bottlenecks - for example, energy restrictions - will be difficult. If we do this, we will be at the very end of the journey. But I suppose that this path will not be so smooth.

Aubrey di Gray : We have a clear idea of ​​what technologies can be developed, but much more vague ideas about the timing of their development.

Perhaps only 25 years will pass, and we will be able to achieve a second cosmic velocity.

But this assumption has two warnings: firstly, it is a matter of substantial resources that are required for development, and secondly, even in this case, the odds are 50 to 50; those. we have a roughly 50 percent chance of achieving the goal. But there is an approximately 10 percent chance of not achieving the goal in the next 100 years or so.

In a way, all this is unimportant. Fuzzy deadlines should not affect prioritization. We must do what we do, no matter what.

If you look at such approaches to artificial intelligence, you will understand that in order to succeed, you need to have both a great understanding of how the world works and much greater computing power.

They are valuable even with a 10 percent chance of success over the next 30 years.

We must be sympathetic to the fact that we will have to consider very difficult approaches. Planning for technology is not easy. In fact, this is a process of pushing the boundaries of the unknown and planning manipulations of nature, begun with an incomplete knowledge of nature at the starting point.

Achieving a full-blown load - and its timing - is such a speculative topic that it probably makes no sense to talk about it as something really likely. But our priorities should remain the same: to develop revolutionary technologies in biotechnology, computing, hardware, etc.

Sonya Arrison : Most of the time I do biotechnology, so I will talk about this segment in the first place. Clearly, biology is rapidly turning into an engineering problem.

I became interested in biotechnology a few years ago when my IT friends started buying books on biology. They believed that the next big thing in programming would almost certainly be biology, not computers. Now this point of view has become mainstream. Bill Gates once said the same thing, along with the rest. The best engineers go to biotechnology. After 30 or 40 years, the approach to biology as an engineering discipline can radically transform the world. There is a feeling that genomics is developing faster than Moore’s law says.

Prices are falling. The calculation of the first human genome cost about three billion dollars. Now it can be done for about $ 1,000.

In the field of genome compilation, work has been done that allows us to study all kinds of genomes that organisms only have, and this opens up a bunch of possibilities. Today, the main claims boil down to the fact that, despite the fact that the first human genome was obtained in 2000, 12 years later, not much has been done towards new treatments or treatment methods based on this technology.

The position of such critics is rather weak, because they overlook an important thing: for most of this 12-year period, calculating the genome was so expensive that only very few scientists could use the genome in their work. Of course, now that prices have fallen seriously, this barrier is also falling.

Something will definitely happen - because people continue to work on radically new things. Gene therapy as a treatment method looks promising.

Perhaps we can develop new types of fuel. Kickstarter has a project for combining firefly and oak genes. The result should be trees that will glow.

This is more than just a cool idea - perhaps you are using these glowing trees to illuminate the streets instead of streetlights. It's great. And there is so much more that we cannot even imagine now. Much can and will happen at the intersection of biology and engineering.

Of the things not related to biotechnology, moving learning online seems to be able to radically change the scope of education. Things like the Stanford Artificial Intelligence classes, Udacity , Khan Academy - we don’t know exactly how it will all end, but it can be argued that there are many things that are expected on this front.

Peter Thiel : Let's get the cultural component involved: why do most people think you are crazy?

Michael Vassar : Having an opinion about the future - it seems strange anyway.

Only a very small minority is trying to imagine the future, even the near future. Perhaps because thinking about the future is somewhat uncomfortable and difficult. People prefer to work with models in which there is only one variable, and everything else remains in the same state. Of course, we know that this is nonsense, because the world is structured differently. But this serves to simplify. Reasoning in such a simplified manner, we can focus on one thing and work together. Spraying onto 100 unknowns would in some sense deprive us of this dynamic.

But thinking about the future is very important, and this is what can isolate you.
Removing from people means there are fewer people you can talk to. There are fewer common related meanings; people no longer understand where you got it from.

But this does not mean that people believe in something different from what we believe. Usually not. Usually you do not face the fact that someone is sharply against your dissimilarity. Perhaps those on the idea of ​​global warming or the apocalypse are really firmly convinced that other views on the future are unacceptable. But most do not think too much about all this. What is perceived as insanity is not the content of the views, but rather the fact of putting faith in the first place.

Aubrey di Gray : I somewhat disagree. People are somehow inclined to have some idea of ​​the future. This is usually an expectation of relative stagnation.

People tend to think not only that most things will not change, but that even that which will change will not change too quickly. People who criticize my point of view in biotechnology and aging, for example, do not recognize bad logical steps and do not grasp the essential points. On the contrary, they prefer not to believe what I am saying, because this contradicts their tendency to stagnate. They walk away quite convinced that progress in anti-aging technology and life expectancy will never accelerate. And it is amazing.

I try and contrast this, paying attention to the fact that if you were to ask someone in 1900 what time it would take in 1950 to cross the Atlantic, they would predict based on the speed trajectories of ocean liners. They would not have been able to anticipate the appearance of the aircraft. That is, they would be mistaken in calculations by orders of magnitude. Of course, everyone knows how many technological changes have occurred in recent centuries and decades. Everyone knows what the internet has done in recent years. But there is a great reluctance to use any of this as a precedent for what might or is likely to happen in the future.

This can be approached from the point of view of desirability. Fear of the unknown is a very deep-seated emotion. When people come up with radically new opinions, they tend to think that everything will go the wrong way. It is very difficult for people to consider the reasonable possibility of implementing such scenarios, so they exaggerate the risks. More rational discussion approaches disappear.

Sonya Arrison : I note that no one considers me crazy.

Peter Thiel : You disguise well ...

Sonia Arrison: Well, it’s hard to call me “crazy,” since I am focused on reality-based technologies. I write about the cultivation of artificial tissues, regenerative medicine, the disclosure of the secrets of nature, for example. It all exists now, and continues to evolve, and, I believe, is really changing the world. There are three reasons why people have problems with this.

First, they do not understand this. Secondly, they do not believe this. Thirdly, they are afraid of it.

Think for a second about lantern trees. Some people are excited only by the idea itself. This is very different from what it is now. Some react automatically, without really thinking: “Do not interfere in nature! Do not play with God! ”Such a reaction can be understood, but it blocks the way to progress. This is not a good reaction. Often it is not productive.

Peter Thiel : Then the best approach is to ignore these people?

Sonya Arrison : Than to ignore, it is better to train them. It is important to convey things in an accessible language. Technology that people don’t understand is often more like magic. And the magic is scary.

But if you clarify - “this is responsible for this” - you can “sell” this idea to them. This is just a matter of explaining the benefits and value. “It will displace dirty fossil fuels,” for example, could be one of the convincing lines of argument in favor of firefly-wood streetlights.

Peter Thiel : There is a very interesting case that we are likely to see unprecedented or accelerating progress in the coming decades.

So why not just sit down, grab some popcorn and enjoy the show?

There is another cut of this question: in R. Kurzweil’s book “ The Singularity is Near ”, progress follows an exponential growth curve. This is the law of nature. In a way, a singularity happens regardless of what individual people are doing for its appearance right now.

The assumption was that there will always be enough people who are trying to do something new, so personally you do not need to do anything, and you can just wait for something to happen on its own. Is there anything wrong with this argument?

Aubrey di Gray : Yes. The role is played not only by the fact that technology is evolving.

When they develop, it is no less important. Take the fight against aging, for example. Approximately 150,000 people die every day. About 100,000 of these deaths come from old age. (Probably about 90% of deaths in Western countries are due to old age). That is, every day that you do not delay will save 100,000 lives. From this point of view, it does not matter how inevitable the singularity is.

The fact that it is inevitable is a little comfort for people who are dying or are losing their loved ones now. We want to defeat aging with the help of medicine as soon as possible, for the simple reason that the sooner we achieve this, the more suffering we will ease the suffering.

Michael Vassar: I completely agree. It is important to work towards the realization of benefits and avoidance of harm. Inevitability can be double-edged: sometimes you want it to come (if it entails good consequences), and sometimes you don’t want things to happen. By focusing on inevitability, many other important things can be lost. If death is inevitable or seems to be such, then we will proceed from the fact that we all will die sooner or later, but still there is some chance of survival, and we must fight for it. In addition, popcorn is harmful. Although, I think, Aubrey could figure out how to make it not so harmful ...

Sonia Arrison: Focusing only on inevitability is dangerous because it allows people to be complacent about the bad systems that are already in use ... People can ignore the many perverse incentives that often interfere or demotivize scientists working on radical technologies. Too few people think about how the FDA (US Food and Drug Administration (US Food and Drug Administration), the US drug market regulator, without whose consent drugs cannot be released to the market - approx. .rev.) can block very important developments. If all this happens in any case, then it makes much less sense to reform what we have now, that is, we can better realize our goals. But of course,

Peter Thiel : Well, so who do you think will do this? Who will forge our technological future?

Michael Vassar : You. (laughs)

Peter Thiel : (pause) Michael ... it is assumed that you will motivate people in this audience ...

Michael Vassar : But I am speaking seriously. There are few such people. You, Elon, Sean ...

Sonya Arrison : In my opinion, innovations come from two sources: from top to bottom and from bottom to top. There is a huge DIY community in biology. These lovers work in laboratories that they set up in their kitchens and basements. At the other end of the spectrum is DARPA(Defense Advanced Research Projects Agency (Agency for advanced defense research projects; an agency that seeks and finances promising technology projects - approx. Transl.) Which spends a ton of money trying to bring out new organisms.

Scientists from different countries communicate with each other, jointly working on complex projects in the field of biology. And this interconnectedness plays a very important role. Together these interactions will bring the necessary changes.

Aubrey di Gray : I do not agree. My answer is Oprah Winfrey.

Yes, there are few people like Peter. There are very few such diviner people who can really make important changes at an early stage in the formation of projects. But there are still quite a few people with this way of thinking, like Peter, who do not. This does not mean that these people do not understand the problems or the importance of technology. They understand all this very well. But they are held back by public opinion. They may not be able to clearly explain this to themselves ... But they feel inwardly the emotional blockade that other people raise around them. A good financial situation does not mean that you are not afraid when people laugh at you. Many potential soothsayers refrain from acting only because they do not resist the pressure of society ... That is why those people who form the majority opinion are so critical. Maybe,

Overpowering public resistance and influencing discourse, these people can encourage anyone to build technology. If we can change public opinion, large benefactors can launch this mechanism.

Michael Vassar : I do not think that progress will be initiated from above or from below, well, the truth. Individual patrons who focus on specific things, such as Paul Allen, of course, do a good job. But they do not really accelerate the future; they accelerate a separate stream to a greater degree in the hope that this will accelerate the coming of the future.

The feeling is that these people do not compare watches with each other. Historically, the top-down approach has not worked. And the bottom-up approach, too. Changes are taking place somewhere in the middle - a Quaker community type, the Royal Society and the founding fathers (referring to a group of politicians standing at the origins of the US base, http://en.wikipedia.org/wiki/Founding_Fathers_of_the_United_States - ca. Perevi.). These effective groups totaled several tens or several hundred members. Almost never there are geniuses working alone.

And these are almost never defense departments or large institutions. You need addiction and trust. These features cannot be in one person or in millions.

Peter Thiel: There are three points of view on who determines the future: the combination of “top down” and “bottom up”, the creators of public opinion and community. Let's work with Michael's idea of ​​a community. Imagine that this is just a small group of people working in the technological field that promotes their ideas.

Aubrey Di Gray: I think the community argument is correct. Michael is right that one person will not change anything. Here a lot depends on the infrastructure. In biology, work costs a fair amount of money. Developing algorithms can also be quite expensive. People need to embed themselves in a cash flow network, no matter if it comes to financing from entrepreneurs, philanthropists, or the public. But the really radical technologies that were discussed in this lesson are so early that philanthropic help will probably play a key role for some time to come. This can change quickly if these technologies advance dramatically and more people see commercial survivability in them. When public opinion changes, people who want to be elected will finance the things that the public wants,

Sonya Arrison : In a sense, the desire to see only one source of progress is wrong. Progress can appear, and usually happens, from many areas. Things are interconnected. Ideas grow out of each other, and often ideas that seemed unsuitable could later work.

Question from the audience : We know that there has been progress in the past. But quite rarely, this progress looked like what people assumed at the beginning. So how do you know that your statements about how progress will be made in the future are correct? What do you think about the conventional wisdom that “most discussions about the future are either fantasies or crap”?

Michael Vassar: People are used to predicting the future in a rather definite way. Imagine that you are looking for oil. This implies a rather specific forecast: in such a place there is so much oil, which will end in so many years. The bulk of people stopped doing this. Recent science fiction is about to do a bit more than science fiction of the past.

It was usually difficult to predict the distant future. It may be relatively easy to predict what will happen by the end of the 2020s, based on previous experience. But it’s unusually hard to make any statements about the year 2040.

Before the era of cinema and the media, people knew how to predict the future much better. They used logic and trend analysis, not what looks cool on the big screen. Current forecasts of the future are more focused on looking likely than on making informed and accurate forecasts.

Look at things like Neil Stevenson’s Avalanche has some good abstraction, some satire here and there. There are many details that probably will not be so in the late 2020s. But we can perceive them as justified as the Kurzwellian descriptions of the possible future of technology.

Sonya Arrison: This question ultimately says “Okay, a lot of people in the past made a mistake about the future, so why should we even talk about this now?” This is nonsense. Yes, people will be wrong. But we are not talking about guesses from the “finger to heaven” series. We talk about what we have now and build on this. This is not science fiction. Gene technology and therapy already exist.

Yes, we can create viable code, as Craig Venter demonstrated . The question is how long it will take and how fast we can move forward. These are questions that are difficult to answer. But this does not mean that you do not need to think about it. We must think about it. The fact that people have different points of view does not invalidate the project.

Question from the audience: Will the future be a scientific or engineering problem?

Aubrey di Gray : If this is the question, then we are exactly in the middle. In medicine and computing, for example, we are seeing a shift from research and science-based approaches to an engineering approach.

Michael Vassar : Science is more important than engineering.

But it’s easier to talk about the latter. That is, someone should use engineering to remove 99.9% of people who have no idea what is going on. But then this one should immerse himself in science along with the remaining 0.1%. This is where success comes from.

Sonya Arrison: There is still the problem of aggregating knowledge. It is difficult or impossible for one person to know everything. And it turns out that people do not know what others are doing, so sometimes they work on the same or something redundant. With computers, knowledge can be better organized, whether it relates to science or engineering.

Question from the audience : In the field of iron, Moore's law seems to remain relevant. But in the software field, the development and collaboration process seems to be improving only linearly. Is this the problem of finding acceleration tools, or are we dealing with some kind of hidden limit in this area?

Michael Vassar: A linear increase in opportunities will help you overcome major barriers. There is a circle of feedback. Linear growth may be sufficient to control the process, speed it up and get positive feedback to see the changes that will lead to exponential growth. And then you go back to linear growth. This is probably true for all psychology and artificial intelligence (which, in essence, is psychological engineering).

Peter Thiel : We know that in practice, time planning is very important.
Therefore, at the very time when we do not know exactly when the revolutionary technologies of the future will be embodied, time planning plays a very large role. This is all science fiction that just looks like the truth, and maybe it makes no sense to work on it now. It will be like the Chinese who tried to launch a rocket in the 11th century. No one worked and could not work on supersonic flights in the Middle Ages.

Aubrey Di Gray: I do not think that the aspect of time is so critical. On the way to the final goal there should always be intermediate points. In the 11th century, the goal could be to fly to the moon. But technology then only allowed, say, one foot off the ground. That is, at that time you could get a degree if you developed a system that would allow you to get 10 feet off the ground.

That is, the question is which paths will lead to the ultimate goal and which will not. We must recognize good paths and give them a green light. But without a long-term goal, you will not be able to organize a competitive path and you will never achieve anything.

Peter Thiel: That is, perhaps, the goal of 20 years ahead with many intermediate goals is a good approach. The problem is that the more milestones, the more abstract the question of attainability of the goal becomes.

Aubrey di Gray : You are forced to watch how it comes and avoid bad turns. And there are still humanistic reasons to look at things more broadly.

We must remember that 100,000 lives are saved every day because the solution to the aging problem comes sooner. In this world, 20 years is much better than 21.

Sonia Arrison: People are usually scared away by a goal that seems too difficult or impossible. We cannot see in every tireless seer. That is, demonstrating the chances of success is a key thing. We can grow blood vessels, tracheas, and bladders in laboratories. So we can get to the hearts. Demonstration of intermediate steps is the key to success, because without them fewer people will be carried away by the prospects of growing new hearts.

Michael Vassar : Apollo Projectwas a gigantic project 10 years long, incorporating many technologies. That was over 40 years ago. In this case, we probably can no longer even fly to the moon. The creation of the US Constitution was an incredible success. The founding fathers knew how to do this. They wrote based on a specific socio-economic and technological context. They did not intend to write a comprehensive guide for the whole world at all times. And yet, when we seize power in the Arab countries today, we simply repeat our constitution. We have no idea how to do what the founding fathers did 200 years ago. We have lost the ability to make a system so elaborated in terms of cultural nuances. Applied history is greatly underestimated.

Question from the audience: No trend exists without any restrictions. Where is the asymptote of the future? At what point do we reach the borders of the physical world? How long does exponential growth last, and at what point does it end?

Michael Vassar : It's hard to say where it ends. It’s possible that it’s not yet - much more needs to be done. If something happens x times in a row, and there are no other variables, the only way to think about the chance that this will happen again is to evaluate using the formula (x + 1) / (x + 2). This is a very crude trick, but it can be quite useful.

Aubrey Di Gray: Kurzweil assumes you get S-curves. But these curves can be replaced by new S-curves, and each time the paradigm will be improved. Glue all these curves together, and you get an S-shaped megakriva. Obviously, within physical laws, this is the limit to which you can advance. But we have not yet reached these problems.

Sonya Arrison : In a way, things slow things down. But it normal. Necessity breeds all inventions. There will be other things that you can do. There will always be a new exponential curve.

Audience Question : We at the Stanford Transhumanist AssociationWe are interested in an open dialogue about the consequences of technological changes, and we conduct a lot of research on how basic emotions, such as fear or empathy, influence a person’s assessment of technology. What, in your opinion, are the most effective ways to interest people with the ideas of transhumanism?

Sonya Arrison : Sometimes you can just appeal to humanity. Some aspects of transhumanism, if well understood, could alleviate a lot of suffering. Some questions very well fall into this category. Therefore, if you designate them correctly, the conclusion will be obvious. No one wants to increase suffering.

Other things do not fit well on this plane. There are things that only look revolutionary new - we may think that they are wonderful, but for others it’s not. The emotional argument in such things is that people must be free to be individuals. But here the moment of fear is very important. Many people fear freedom. That is, the problem lies deeper.

Michael Vassar : You can appeal to the human sense of wonder. If you have ever talked to those who suffer from Alzheimer's disease or some kind of mental disorder, you might have the feeling that they are missing something. Also, like all of us. The gap between them and us is negligible. We may be missing a ton of things. Shouldn't we try and fix it in order to lose less?

II. Final Thoughts (from Peter Thiel)


This course was mainly devoted to the transition from zero to one. We talked a lot about how to create new technologies, and how radically improved technologies can bring singularity closer. But we can use far more tools to move from zero to one. Every new thing in the world contains something important in terms of singularity. Whether you are creating a new company or making a key decision in your life, this is a singularity in miniature. In fact, the life of every person is a singularity.

The obvious question is what to do with your singularity.

The obvious answer, unfortunately, is to go off the beaten track. You are constantly encouraged to play it safe and be ordinary. The future we talked about is only probabilities and statistics. You are only part of the statistics.

But the obvious answer is wrong. It’s like selling yourself cheaply. Statistical processes, the law of large numbers, and globalization are timeless things stemming from probability theory and possibly random.
But, like technology, your life is a story made up of events that happen only once.

By their nature, singular events are difficult to study and generalize. But the big secret is that there are many undisclosed little secrets. There are still a lot of white spots on the map of human knowledge. You can research them. So go and fill these white spots. Literally every moment is an opportunity to go to these new places and explore them.

In all likelihood, there is no special time, which is certainly right in order to start your company or your life. But some points seem more favorable than others. Now is just such a moment. If you do not take responsibility and do not open towards the future, if you do not take responsibility for your life - there is a feeling that no one will do it.

So go in search of your borders and explore them. Decide to do something important and different from everything. Do not limit yourself to the terms “luck”, “impossible” or “useless”. Use your power to build your own life, go and do new things.

Note :
I ask for translation errors and spelling in PM. Translated mg1, astropilot editor , all thanks to them.

Also popular now: