A wild idea is one that many people think is obviously quite unlikely. I'd say at least a third of these wild ideas are likely true. (Parenthetic links are to more of my writings.)
The startling prediction of the "many worlds" interpretation of quantum mechanics is that when systems like your mind interact with small quantum systems, every possible quantum outcome actually happens in a different "world." Quantum mechanics is our most basic theory of physics, and surveys of prominent physicists reportedly find majorities favoring this interpretation. (More here.)
In the 1970s, the RAND Health Insurance Experiment randomly assigned 5000 adults to free or full-price health care over 3-5 years. Free care folks got more eyeglasses and teeth filled, and spent ~30% more, but were otherwise no healthier. This result is consistent with typical time-series and cross-sectional analyses. I'm willing to extrapolate from this 30% change to a 50% cut. (More here.)
We learn more about brains and making smart computers, but we seem to have run out of major architectural innovations -- better ones won't make a huge difference. The big stumbling block seems to be how much "common sense" a system knows, like that things tend to fall down when you bump them. One group has been writing these down for fifteen years with moderate success; a century more effort may be plenty.
Your mind is a pattern of activity in your brain. The ability to induce that pattern is encoded primarily in your neurons -- in which neurons are of which type, and which neurons are connected together. Freezing a brain today in liquid nitrogen destroys many things, but seems to preserve this type/connection info. By 2100 we should be able to scan this info from a frozen brain. If we scan your brain and then build and run a computer simulation of it, someone who remembers being you would wake up and feel alive. (More here.)
Simulated brains are potentially immortal, just as all computer data is. And the ability to cheaply simulate brains will revolutionize labor economics; wages should fall to near the cost of making brain simulators. The population of such "uploads" should expand very rapidly, allowing huge increases in both economic growth rates and inequality. (More here and here.)
Some uploads could have robot bodies, while others could live in simulated computer worlds. Our descendants may place some of them in historical simulations, with simulated people who do not realize that they are simulated. How sure can I be now that I do not live in a future historical simulation? The more such future simulations there will be of this era, the higher a chance I must assign to this possibility. (More here.)
Our best long-term time series of world product is well fit by a sequence of three exponentials, each of which grew over a hundred times faster than its predecessor. If growth modes are similar in terms of how much the economy grows when they dominate, we will be overdue for a new mode by a century from now. So if a faster mode is coming, it should have come by then. (More here.)
This claim is on this list not because of the usual concerns like nuclear war, ecological collapse, or bounded resources, but because so many wild claims seem plausible if we long continue our economic, technical, or spatial/material growth. Two percent growth two thousand times is 1017, for example. So either this claim is likely, or several others on this list are. Alas, the most likely no-grow scenario seems to be obliteration.
If our growth does not stop, it must continue. And it cannot continue this long without enabling and encouraging massive space colonization. Spatial/material growth requires it, technical growth enables it, and economic growth induces technical growth. (It won't really start for many decades, though, until space costs come way down. More here.)
Astronomy suggests that the chance of intelligent creatures arising in any one largish region a billion years ago are similar to the chances today. So if many such creatures arose then, and if each one had a non-trivial chance to start a space colonization wave within a few million years, then most everything should be colonized by now. Yet things do not look colonized around here. (More here.)
If colonization waves are far enough apart, competition among colonists selects for the fastest possible colonization wave speed. When such colonists stop somewhere, they grow as fast as they can and then leave in a big hurry, taking everything of portable value with them. (More here.)
Limits on freedom of contract are today said to fix problems of monopoly, externalities, asymmetric information, and local irrationality. But if limits are good fixes, then people should voluntarily choose such limits in early contracting over contract law itself. If competing private laws offered law and enforcement packages, contracting early with each other on inter-law disputes, we could each better tune our laws to our individual circumstances and preferences. (More here.)
Under futarchy we would vote on values, but bet on beliefs. Legislators would define and monitor an after-the-fact measurement of national welfare. The rule then is: if market speculators estimate a policy would increase national welfare, that policy becomes law. Because speculative markets aggregate info well, choices would be based on our best available info about policy consequences. (More here.)
On matters of fact or morality, honest rational truth-seekers cannot agree to disagree. Even if highly computationally constrained, they should not be able to anticipate the direction of others' opinions relative to their own. Yet virtually no pair of humans is like this. Thus virtually no humans are truth-seekers, and since most humans think they are truth-seekers, they are self-deceived. (More here.)