Backcasting is a prediction methodology that attempts to predict the technology development path based on the expected final outcome. This methodology assumes that there are some predetermined attractors (caused by laws of nature of human needs) that the technology will gradually approach. This assumption is called technological determinism.
To predict the future using backcasting you need to take a final "extreme" result that you expect to be true (either because it's physically possible, such as a nanoassembler, or because humans are bound to want it, such as perfect virtual sex).
For example, you state that we will acquire "the ability to simulate all neurons in a human brain in a computer". Sometimes you may be able to put a timeframe estimate on this prediction as well (e.g. "the technology for brain simulation will arrive in 25 years"), because enabling technologies are likely to be developed by then (sufficient computing power, accurate brain scanning technologies).
The question you ask next is "What approximations of that technology would arrive in the interim?". Then you work backwards, trying to "retroanalyze" this final result.
Backcasting can be useful to think about discontinuities brought by the arrival of disrupting technologies. For examples, the nanoassembler will probably to be a disrupting technology. Its effects on the society will likely be determined by the situation right before its design is finally perfected. It is hard to predict that situation starting from today, but backcasting may help imagine what nanotechnologies will then be available and what effect will they have.
Similar ideas exist under different names:
- Backward chaining (Eric Drexler)
- Horizon mission methodology (John Anderson)
- Retrosynthetic analysis (Elias J. Corey)
- Shortest path and other search algorithms in computer science
- “Meet in the middle” attacks in cryptography