For decades, many scientists thought the best way to create an AGI was to make a computer simulation of a human brain. Unfortunately, the brain’s workings have proven less tractable than they hoped, and our understanding of how it produces thought remains basic. The rapid advances in LLMs show that the first AGI could have an alien architecture that only shares passing similarities to the human brain. If AGIs take over the world and can do whatever they want, they might never try creating accurate simulations of human brains since the resources could be better spent developing newer kinds of AGIs that were smarter than we are and possessed whatever comparative advantages of cognition we have. The human brain computer simulation could become a “stranded technology.”
Paper books might never become obsolete. They’re cheap ways to store information, don’t need electricity to work, and can stay intact for hundreds of years if not abused and kept in the right conditions (e.g. – a dark room at a constant 68-72° F, with 40-50% relative humidity). As the books finally start wearing out, you can cheaply make new copies of them without losing any information and then reset the clock for several hundred more years.
On the other hand, let me say that I think that some of the advantages of paper books will wane once books with paper-like “e-ink” pages are introduced. Everyone agrees that paper-thin display screens are coming, so when they do, why not use them to make books? The front and back covers and spine would contain a processor and battery, and between them would be several hundred pages of paper-like displays. You could download a library of thousands of different book titles, and, at the tap of a button, have your one e-book reconfigure all of its pages to be whatever book you wanted to read at that moment. I predict this device will be available for less than $1,000 during the 2030s, and should get very cheap in the 2040s. I’m unsure if it would make paper books obsolete, but it would sway a lot of people who still cling to books because they like the physical feel of them and don’t like getting eye strain from glowing screens.
Paper books might truly become obsolete if intelligent machines replace humans. They’ll have instant mental access to nearly all information through the internet, and will have individual, internal memory banks with the equivalent knowledge of several million books. (All of this stuff could be made EMP proof as well, so paper books lose that advantage as a storage medium.) For advanced beings like that, it would be a waste of time to even grab a book lying on the table, open it, and flip through the pages to do speed reading because the same book could be downloaded and read in digital format even faster.
That said, I don’t think non-digital media will become obsolete. At the very least, microfilm offers important advantages as a data storage medium that AGIs will appreciate.
A major roadblock to building lifelike androids is their skin and hair. Even if an android’s frame perfectly mimics a human skeleton and its “flesh” perfectly mimics human muscles and fat, the hair and skin will gradually stop looking human due to wear and tear. The lack of healing mechanisms and hair follicles will mean that every cut, puncture, or burn, or every hair strand that is pulled out or cut short, will become permanent. Over time, this damage will get so noticeable that an android won’t be able to pass for human.
The first solution to this problem will be making the androids’ skin out of removable components. For example, the scalp would be its own skin component that would border the face component, and there would be a seam along the hairline. Each limb would have its own skin component, and the torso would have at least one skin component. If one area of an android’s body accumulated enough surface damage, it would have that skin component replaced. The problem with this is that the presence of seams would prevent the androids from looking fully human.
The second solution will take much longer to arrive and will be for androids to have self-healing abilities that repair damage at the microscopic level. This might involve having internal self-repair nano- or micromachines, or having those repair bots in a separate container that and android would open and use to repair damage as needed.
In many sci-fi depictions of androids, they have colored, glowing eyes. I doubt the real ones will look like that (except customized machines), but at least early on, androids might have appearances that deliberately set them apart from humans for the sake of our comfort and security.
Once we have androids capable of the full range of human facial motion, using machine learning to train an AI to move the face realistically will be easy thanks to the millions of hours of footage human faces moving realistically to compare it’s work with. If that isn’t enough we can get volunteers to wear sensors to track muscle movements in real human faces.
A future plane built with advanced materials might have “feathers” and the ability to flex its wings in the same ways birds can (the F-8 Crusader had wings that could “pivot” seven degrees, helping it to take off and land, and the F-14 could also independently retract or extent its wings to maneuver) I don’t think most planes will ever have these features, but a minority of them designed for specific roles might.
A biological human “caste” system like that imagined in Brave New World and other works of science fiction will not happen. This is because, by the time we have the genetic and medical technology to reliably and cheaply instill specific traits into offspring, intelligent machines will be better than even the most highly engineered humans at those tasks. For example, what’s the use of caste of humans designed to be manual laborers (they’re probably stronger and more endurant than average) in a future where robot workers are everywhere?
The closest we could get to castes would be creating highly engineered organic life forms for specific tasks. For instance, by virtue of its organic construction, brains made of nerve tissue could have comparative advantages in certain types of data processing over silicon chips, so there could be “server farms” of highly engineered brains floating in tanks of liquid just to perform that mental work. Likewise, intelligent machines might find it useful to keep Earth and other planets inhabited by general-purpose, generally-intelligent organic, servant life forms that could restore civilization if a calamity disabled all technological life. However, any such life forms would be so different from each other that they wouldn’t be the same species, and hence could not be thought of as belonging to “castes.”
People worry about intelligent machines exterminating the human race, but they could also neutralize us as a threat by altering our minds, or at least by altering the minds of the most troublesome people. There have been countless instances of people’s minds being permanently because of brain damage (Phineas Gage is the best example) or brain surgery (lobotomies are the best examples), and even noninvasive procedures like transcranial magnetic stimulation can alter how people think. With more research and better technology, the future powers that be could find the medical means to affordably and permanently control how humans think and act, removing the need to kill to retain control.