I think that it is nearly impossible to figure out the exact changes that new technologies will create in the world. That's problematic because investing in startups is largely about figuring out whether or not a team with a new technology - or application of existing technology - will create a change large enough to support a big new company.
Not only is it hard to figure out what changes will happen, it's hard to figure out what changes will happen over given periods of time and if those changes are good or bad. In fact, it's probably safe to assume that any given technology will create badness over certain periods of time even if the long term impact is highly positive. Computers are a good example of this. During WWII, IBM punch card machines (not quite computers, but getting there) were used by the Nazis to organize aspects of the Holocaust. At the same time, Alan Turing was building the foundation of modern computing to crack Enigma. Internal combustion engines gave us mobility and trade on a previously unimagined scale, but also led to anthropogenic climate change.
Since there's no way to know every change that a piece of technology will produce, I think there are two ways of evaluating what to build and invest in. I think of these as Utopia Bets and Apocalypse Bets. The most extreme example of this dynamic comes from AI. In one framing, AI creates a world in which all of our hard problems are solved and humanity devotes itself to exploration, art, and generally being good. In the other extreme, the AI wipes us out.
I've found that many people like to talk about their Apocalypse Bets. There is something emotionally satisfying about being cynical and painting a dark version of the future. I think one of the reasons people do this is because they don't want to look stupid and they want to hedge against bad outcomes. If you predict something will go well, and it doesn't, not only does a bad thing happen, but, worse, you are humiliated for being wrong! If you predict something will go badly, and it does, at least you were right. If it goes well, your life is better, and everyone forgets about you being wrong.
Certainly, Apocalypse Bets are popular in the public imagination and press these days, as you can see by looking at books and movies where technology gets away from man and destroys us. Turning on the news or reading the paper, you can find any number of talking heads discussing why our love of technology gave us unstoppable pollution, weapons we can't control, and epidemics ready to wipe out half of humanity.
Add into this that whenever the market starts going down or volatility picks up, talk about how much worse everything is going to get becomes more common because people are scared. There are investors that have made a lot of money off betting that things will get much worse, mostly in the form of shorting the market in one way or the other. No matter what the talking points are, though, I've never actually met someone who makes venture investments in Apocalypse Bets, regardless of how bearish they are about everything else.
I think this is because capitalism is a bet on the future. Investing over the long run has a positive expected return because markets believe, overall, that the economy will grow. If you didn't believe this, you'd never invest. If that's true, then I think venture investing is an exercise in optimism.
The whole process of betting on future return is magnified by startups. In startups, the founders and the investors are betting that a small number of people can change the way the world works, and make a lot of money in the process. It seems that most of the changes that founders are trying to create improve the world. I can't think of ever hearing a pitch where the founders argued that the world getting worse would be a net positive for their business. This might be an argument made by arms manufacturers, but I don't know as I've never been pitched by one.
Startups, then, are making Utopia Bets. It is rare to find a founder who argues that their company will, by itself, bring about Utopia, but they mostly believe that the aggregate force of technological change is pushing humanity to a much better place. This makes sense, seeing as how spending your life working towards making the world a worse place would be depressing for anyone that isn't a supervillain.
The best investors I know look for Utopia Bets as the direct rationale for investing in a given company. Investors rarely pick a specific solutions which they think will bring about a better future and then finding a company to do that thing. Instead, they'll often start with a question similar to: “In the future, cities will have 40mm people. What needs to be built to make those cities function well?” This leads to branches of sub-questions and hypotheses. The investor would probably need to think about transportation, communications, logistics, etc. Each of those areas gives rise to a potential set of companies, towards which those investors will be receptive.
This thought process can be useful to investors in two ways. On one side, having a distinct view of how the future will be better gives investors a way to publicly talk about what interests them. If they say intelligent things, founders who are thinking about the same ideas will reach out to talk about those problems. Some of those founders will be really good, and may end up building companies which the investors can put money into. There are also many people who have thought deeply about these problems, but didn't think anyone would fund the crazy ideas they had to bring about a better world. When those people come out of the woodwork and build companies, not only would the investor likely the first opportunity to invest, but, even if the investor passed, there's a net good of new interesting ideas being tried in the real world.
The other good thing that happens through this thought process is trickier. Thinking deeply about certain problems can be a really helpful filter when deciding to invest in ideas, but it can also be misleading. On the one hand, thinking deeply about a given set of ideas can help differentiate what is good and what is bad. Conversely, it can also create significant bias towards funding ideas that seem like a perfect Utopia Bet, without considering the founders. From this perspective, any bet is great if it stands even a tiny chance of making the world better.
That usually leads to funding companies that only seem good. It's much better to consider whether or not the founders are actually good and likely to build the company that will help make the Utopia Bet come true. You can get a read on this by working through how deeply those founders have thought through the idea as it relates to the way the world is going to change. In fact, the best Utopia Bets are about the founders and their views on how the world will change, rather than the a priori assumptions of the investor. These a priori assumptions can actually end up creating throwing false negatives, because the best companies often exploit something in the market that outside experts have dismissed as non-viable.
This fits well with the idea that the best founders should know far more about what they're making, and be far more passionate about it, and be thinking far more originally about it, than anyone else. These founders should continually upend how the investor thinks and the investor should be learning more from the founder than vice versa. If an investor actually knows more about the idea, and is so passionate about it, that investor should build the company!
Founders who find investors whose view of the future echo - but doesn't mirror - their own end up in a relationship that is far more collaborative than those who take money from investors motivated only by returns. When looking for investors, it's important for founders to understand what sort of Utopia Bets individual investors want to make, because it will help frame the conversation and create a real dialogue in which both sides learn. That's a far more effective route to raising money than one sided pitching.
This isn't to say that the relationships formed through this process will always be smooth. In fact, when two people with strongly held visions of the future get together, every difference in that vision can lead to conflict. However if both sides are truly pulling for making the future better, they should be able to find a way forward together to build incredible things.
 John Paulson (https://en.wikipedia.org/wiki/John_Paulson) almost $5B off of one set of linked bets against the rising housing market. George Soros “broke the pound” in 1992: http://www.investopedia.com/ask/answers/08/george-soros-bank-of-england.asp. That felt like an apocalyptic bet, but it also may have helped UK break a recession. He also then bet on the Pound and made more money! (thanks for the info, Elad!) I see these bets as different than shorting a company because of their basis on big macro trends.
 This isn't true of short term trades where investors are often betting against temporary pricing imperfections. That reflects a different kind of optimism - confidence that you are smarter than markets.
 Though I've met one or two. They're either crazy, working on AI, or both.
 Bill Gates talks about this quite a bit, and links it back to his and Paul Allen's original vision for computers: https://www.gatesnotes.com/2015-annual-letter?page=1&lang=en.
 We've actually noticed both of these trends in response to our Requests for Startups https://www.ycombinator.com/rfs/.
 False negatives are one of the scariest mistakes that investors can make because of the way returns in VC are dominated by the outliers. See: http://paulgraham.com/swan.html
Thanks to Andy Weissman and Elad Gil for helping me think about this.