It would seem as more time passes the more we are in control of the restrictions that have been placed on us by our ancestors. Natural Selection did a pretty good job of giving us a foundation of a large brain relative to our body mass, giving us a leg up on cognitive functions. Yet the more we progress we also find ourselves stuck with problems that are intrinsically human and as such will probably not be solved technologically.
The Turing test is a simple example of one such problem. In essence the test seeks to develop a computer program “intelligent” enough to fool a human observer into thinking it’s human. At its heart this is a human problem, communication with another being at a meaningful level. If you’ve ever tried to talk with a chat bot as if it were human you’ll notice some characteristics after a short period of time:
- Lack of memory: Most chat bots won’t remember the conversation you had with them past the last sentence. This can lead to some silly conversational loops since they’ll ask you to clarify points they appeared to understand not two sentences earlier.
- Can’t be taught: A lot of chat bots are based on Natural Language Processing where they’re using algorithms based on grammatical constructs and the sentence structure. Whilst you can use this to form some sort of idea about what the other party is saying you can’t teach this dog any tricks once you’ve programmed it. I had a lovely conversation with the “winner” of last years Turing test about regolith. After teaching it what it was I then asked it to describe it back to me. Needless to say it couldn’t, and asked me what I was talking about.
- Spout nonsensical gibberish: If you hit one of the mid points in some of the NLP algorithms you’ll notice that most chat bots come back with a question which takes part of your sentence and regurgitates it back at you. If you said something like “Well I do enjoy opera whilst knee deep in peanut butter” they will come back with something like “What makes it that you enjoy opera whilst knee deep in peanut butter?”. Forgive the awkward grammar and you might be able to pass that off as normal conversation, but no one talks like that normally.
No program has yet completed this test successfully, although there are many good candidates.
The old saying “when all you have is a hammer all your problems look like nails” is something that applies here. Using technology to solve our day to day problems and better our lives is the hammer, and our problems arising from the human condition can all look like nails. Whilst I believe that no problem is above being solved, given enough time and resources, there are some things about life that are just so intrinsically human that technology will struggle to overcome them. Sometimes it is easier to adapt ourselves to overcome such problems; akin to a kind of artificial evolution.
So, what’s the point of trying then? Well I can put it down to two reasons. The first being our insatiable desire to overcome any obstacle that is placed in front of us. Routinely in our past we humans have been faced with problems that appeared monumental. Just on 100 years ago international travel took many months across the sea and was frought with danger. Now you can walk up to almost any airport and choose a country and be there in less than a day.
Secondly, on a time scale that we experience evolution makes negligible changes to us. We are at the stage where we have evolved to a point that if we want to go any further mother nature’s course will take thousands of generations for us to get there. If the human species is to survive and thrive in this barren universe we have to learn to master the world we exist in and then continue the process throughout the solar system and beyond. This is the only way to ensure that our race can survive through catastrophic events such as the loss of the entire earth.
Overall, we seek to overcome our shortcomings due to our innate desire to thrive.
