We should ask ourselves whether the right time to enter a tech industry is early on, as conventional wisdom suggests. The best time to enter may be much later than that. It can’t be too late, since you still need room to do something. But you want to enter the field when you can make the last great development, after which the drawbridge goes up and you have permanent capture. You want to pick the right time, go long on tech, succeed, and then short tech.

Microsoft is probably the last operating system company. It was also an early one, but there’s a sense in which it will be the last as well. Google, the narrative goes, is the last search engine company; it wrought a quantum improvement in search with its shift to an algorithmic approach, and that can’t be much improved on. What about bioinformatics? A lot seems to be happening there. But whether it’s too early to jump in is hard to know. The field seems very promising. But it’s difficult to get a sense of where it will likely be in 15 or 20 years. Since the goal is to build companies that will still be around in 2020, you want to avoid a field where things are moving too quickly. You want to avoid being an innovative but non-profitable disk drive company from the ‘80s.

Some markets are like the automotive market. Should you start a new lithium battery company? Probably not. The time for that may have passed. Innovation may be too slow. The technology may be too set by now.

But sometimes seemingly terminal markets aren’t. Look at aerospace. SpaceX thinks it can cut space launch costs by 70-90%. That would be incredibly valuable. If nothing has happened in an industry for a long time, and you come along and dramatically improve something important, chances are that no one else will come and do that again, to you. 

Artificial Intelligence is probably an underrated field. People are burned out on it, largely because it has been overrated and overstated for many decades. Few people think AI is or will soon be real at this point. But progress is increasingly relentless. AI performance in chess is increasing. Computers will probably beat humans in Go in 4 or 5 years. AI is probably a good place to look on the tech frontier. The challenge is that no one knows how far it will go.

Mobile Internet deserves some mention. The question is whether there’s a gold rush in mobile. An important subquestion is whether, given a gold rush, you’d rather be a gold digger or the guy selling shovels to gold diggers. But Google and Apple are selling the shovels. And there may not be that much gold left to find. The worry is that the market is just too big. Too many companies are competing. As discussed above, there are various rhetorical tricks one can use to whittle down the market size and make any given company seem way more unique. Maybe you can create a mobile company that owns a valuable niche. Maybe you can find some gold. But that’s worth being skeptical about.

源于：https://blakemasters.com/post/21169325300/peter-thiels-cs183-startup-class-4-notes-essay