composing quantitative predictions generating intelligent understanding composing precise contingencies aggregating critical futures composing precise contingencies mapping the future predicting probable contingencies mapping intelligent forecasts mapping quantitative insights modeling accurate estimations composing quantitative understanding modeling precise forecasts calculating critical futures computing contingent estimations

Question

Metaculus Help: Spread the word

If you like Metaculus, tell your friends! Share this question via Facebook, Twitter, or Reddit.

If and when the first whole human brain is successfully emulated, how much will 1 hour of subjective run time cost?

Whole brain emulation (WBE) is the possible future one-to-one modelling of the human brain. The basic idea is to take a particular brain, scan its structure in detail, and construct a software model of it that is sufficiently faithful to the original that, when run on appropriate hardware, it will behave in essentially the same way as the original brain. [1]

If whole brain emulation is possible, then one factor that might influence how it develops after being invented is how expensive it is per emulation. If it is cheap at the moment of discovery, there may be a rapid proliferation of ems quickly replacing human economic activity; if it is very expensive, then growth in the number of ems may be initially much slower.

Sandberg (2014) calls this the "overshoot" scenario, and proposes that the cost per em will be low if the last constraint to be solved is something other than hardware (such as neuroscience knowledge or scanning technology), since whole brain emulation will be delayed to a time when hardware is cheaper.

How much computing power is required to run a whole brain emulation is not known (although rough estimates have been made by Sandberg (2014), and by fellow Metaculites). What is known is that the required power will almost certainly depend on the necessary level of resolution at which the copy of the brain needs to be modelled in order to emulate the human brain at a high level of fidelity.

If and when the first whole human brain is successfully emulated, how much will the necessary computational power to replicate human mental capability for an hour of subjective run-time cost in 2019 USD?


Methodology

In line with the methodology used in AI Impacts (2015), hardware costs will be amortised over a period of three years. That is, for a given computer we consider the fixed costs of purchasing and the cost of operating it over a year. Operating costs may include maintenance, rent of related facilities (power generation/distribution, cooling systems, etc.), utilities costs (i.e. electricity) and staffing. Then, hardware costs per hour is given by:

This figure is then adjusted to 2019 prices using a commonly used producer's price index of the country in which the supercomputer was purchased.


Resolution

If a whole human brain is successfully emulated, as per the criteria in When will the first whole human brain be successfully emulated?, the question will 365 days after the first successful emulation as the lowest hardware cost-per-hour (in 2019 prices) that is enabled by a supercomputer that is shown to run an emulation.

This cost-per-hour is not necessarily that of the the first system that runs an emulation. Instead it will be the lowest cost of the system that runs an emulation within 365 days of when the first emulation is run successfully.

By "cost-per-hour" we mean the cost to run an emulation for an hour of "subjective time", to adjust for the possibly compressed or expanded simulation time. That is, if the emulation processes inputs times as quickly (or slowly) as a typical human, cost per an elapsed real hour (i.e. the cost per wall-clock hour) will be divided by to convert this into costs for an hour of "subjective time".

{{qctrl.predictionString()}}

Metaculus help: Predicting

Predictions are the heart of Metaculus. Predicting is how you contribute to the wisdom of the crowd, and how you earn points and build up your personal Metaculus track record.

The basics of predicting are very simple: move the slider to best match the likelihood of the outcome, and click predict. You can predict as often as you want, and you're encouraged to change your mind when new information becomes available. With tachyons you'll even be able to go back in time and backdate your prediction to maximize your points.

The displayed score is split into current points and total points. Current points show how much your prediction is worth now, whereas total points show the combined worth of all of your predictions over the lifetime of the question. The scoring details are available on the FAQ.

Note: this question resolved before its original close time. All of your predictions came after the resolution, so you did not gain (or lose) any points for it.

Note: this question resolved before its original close time. You earned points up until the question resolution, but not afterwards.

This question is not yet open for predictions.

Thanks for predicting!

Your prediction has been recorded anonymously.

Want to track your predictions, earn points, and hone your forecasting skills? Create an account today!

Track your predictions
Continue exploring the site

Community Stats

Metaculus help: Community Stats

Use the community stats to get a better sense of the community consensus (or lack thereof) for this question. Sometimes people have wildly different ideas about the likely outcomes, and sometimes people are in close agreement. There are even times when the community seems very certain of uncertainty, like when everyone agrees that event is only 50% likely to happen.

When you make a prediction, check the community stats to see where you land. If your prediction is an outlier, might there be something you're overlooking that others have seen? Or do you have special insight that others are lacking? Either way, it might be a good idea to join the discussion in the comments.