4
$\begingroup$

Hooking up a computer to the human brain or somehow connecting the brain to a computer is nothing new in the realm of sci fi literature. But say you want to merge the human brain and a computer together (think basic chip implant) and you do it successfully.

Research

In computer science there exists a division of problems (P-NP) where computers can come solve a problem using an algorithm.

  • Some problems can be solved in a reasonable amount of time with a deterministic approach (P) in polynomial time.
  • Others can't be solved within a reasonable amount of time (NP-Hard) but can be verified in polynomial time if given information about the answer or information surrounding the answer (obviously oversimplifying a lot to a very great deal but this is the general concept for framing).

I am ignoring the differences between NP-Hard and NP-Complete for question simplicity, but if you know the differences feel free to use that to influence your answer.

Examples

The human brain is more than capable of predicting possible future outcomes based on incoming information, and short circuiting decisions based on patterns and behaviors.
Food for Thought

  • When working together, what types of problems does the human brain help computers solve better?
  • Essentially, could the brain help or enhance a computers range of solvable problems, especially for real time/dynamic problems (think military/racing/surgery/ language processing etc). If so, how would it aid a computer?

In this case I'm more concerned about the conceptual mechanisms than the actual how.

  • Would it help in solving short circuiting based on past behavior/memories, race conditions, deadlocks, thread pool starvation, access/index out of bounds, preventing crashes or segfaults, program recovery from bad exits or inputs, prediction and confidence (confidence in a decision/the human version of the Viterbi algorithm) etc.
  • Say someone is driving down a road and they see pebbles rolling down a hill in a very familiar way. Maybe the last time they fell like that, it was the warning sign for a landslide that was about to occur later. The driver would take that past memory and immediately use that to calculate possible outcomes while he is driving. The risks of a landslide, where he is, how likely it is to occur in his path etc. While a computer could also calculate this probability using a combination of algorithms and then do analysis to calculate the confidence of the outcome it created; a computer would need more than just sheer image processing.
  • Someone weighing whether or not they should continue a conversation to change someone's mind when their target is saying questionable things but hasn't gone fully off the deep end yet. Sure, one could use natural language processing to break down the sentence, but that brings in a whole host of issues surrounding cultural usage of words and phrases versus their textbook meaning.
  • Interpreting sentences differently from how humans would because certain words can have multiple cultural contextual meanings or even grammatical meanings (e.g. A computer can hear "She saw the ant" and interpret it as "She saw the ant" or "She saw the aunt". Casual swearing between friends is another example.) But a person listening to the conversation would be able to immediately pick up on the context (I am aware that humans are capable of misunderstanding each other, but I'm choosing to ignore that subset of problems for now since such a situation can't even be verified for correctness by a human let alone an NLP classifier)

The Question to be answered by you

How would the human brain (mechanically) aid a computer in problems previously outside of its solvable range in real time/dynamic problems, given a hybrid between human and machine was possible? Think mechanisms here, not scenarios.

Improve this question
$\endgroup$
13
  • $\begingroup$ As written this seems more like an open ended discussion prompt intended to promote brainstorming and idea generation rather than a specific ask about some issue you're having with your world. Questions with many valid answers like this one are not suitable for this site. $\endgroup$
    – sphennings
    Commented Jul 21, 2022 at 11:47
  • $\begingroup$ Agree with @sphennings OP set the stage fairly well but the actual question past the post title is unclear. "What jobs would a hybrid human brain / computer machine be suited for?" Is my interpretation of the question. However that answer would also involve brain storming. IMHO this is a perfect example of a sad failing of legitimate and interesting question where SE WB fails its purpose. Though this is what SE is. So until a solution is found little can be done, except limbo under the rules bar to reach your goal. $\endgroup$
    – Gillgamesh
    Commented Jul 21, 2022 at 12:54
  • 1
    $\begingroup$ I would suggest modifying either the question in the body of your premise, or in your subject title to better reflect what you're trying to ask. There seems to be some confusion on that topic. $\endgroup$
    – shootbuildthink
    Commented Jul 21, 2022 at 15:06
  • 1
    $\begingroup$ I hope that my edit matched the spirit of your question. I spent a good amount of time looking over it to find your core question. $\endgroup$
    – shootbuildthink
    Commented Jul 21, 2022 at 20:28
  • 1
    $\begingroup$ @shootbuildthink It very much did, thanks. The question at the end is far clearer now compared to before the edit, the mechanisms is really what I wanted to home in on but my floundering in writing got in the way. $\endgroup$
    – FIRES_ICE
    Commented Jul 21, 2022 at 20:32

2 Answers 2

Reset to default
4
$\begingroup$

Human for the plan. Machine for the execution of the plan.

It is not so different from how we have always used machines. I want to go to the lake and see what the ducks are up to. I sit behind the wheel of my car and with my guidance the car gets me there fast and dry. Maybe if I have a fancy car I can just tell it where I want to go. The car is not interested in ducks; that was my plan.

The human can perceive the need for action based on the human's long and short term goals. The machine can implement the action much faster than the human body can.

Improve this answer
$\endgroup$
1
$\begingroup$

Speaking generally, nearly any job done by the current manifestation of how AI might operate would likely benefit from input from a human intelligence. Though areas were ambiguous outcomes from from non-uniform inputs would be where a human type intelligence would contribute the most in such an arrangement. The best example of that would be IMHO would be Law. Not only would the administration on the judicial side benefits from the application of human morality and its intricate complexities. But weighing the sheer volume of written law, in any case would be daunting but applying it with the spirit as written has been a struggle since there has been written law.

Improve this answer
$\endgroup$

Not the answer you're looking for? Browse other questions tagged .