So when I talk to programmers or some university students who are about to get their degree. one of the first questions that I ran into was is it future-proof. or will AI (artificial intelligence) just come in and take over right as I’m about to graduate from university and get my first job.
So by the principle that if I had this concern then there are probably others that are having it too. I decided to make this article in order to kind of share my views on the topic as a software developer. Obviously I did think that it made sense to keep going.
I think people here AI and think that everything is going to get automated that all jobs will be replaced by robots or machines and that the world would go into chaos. Phone companies like Apple, Samsung, and Google are constantly posting their newest achievements in AI. Google shows off its Google assistant calling up restaurants and making pretty complex dinner reservations.
Despite encountering difficulties like bad reception and language barriers. now this is super impressive and it is really a true feat of Technology but is it really AI. when determining what AI actually is something that often gets brought up is the Turing tests that was conceptualized by our Alan Turing.
you can read about Alan Turning here
Alan Turing is what I would call the first computer scientist of our time and he’s also credited with having solved the German Enigma machine during World War Two. There’s a great movie about all of this that I recommend watching if you haven’t that’s called the imitation game. Anyway, the Turing test is essentially what Google does here if you put a person and a machine behind a curtain and you start talking to them your machine has passed the Turing test when you’re not able to figure out when you’re talking to the machine versus the human. So technically I would say that Google passes this test in this showcase.
However, I believe the Turing test is far more complex than this for a machine to truly pass the Turing test. You would have to spend a really long time talking to it while switching subjects and trains of thought and most importantly you also have to be aware that you’re actually trying to figure out whether you’re talking a machine or a person.
Which if we look at Google’s display here they’re calling restaurants where people are picking up answering that is not at all expecting to be talking to a machine. So they’re not trying to figure this out and so, therefore, these people are the easiest to fool and the reason for bringing this up is just to explain that the AI that we see in our phones that make it seem as if the singularity is only a few years away, is not actually as advanced as it may seem.
The second part of this is that the Turing test may not be the test that we should be too concerned with I think that. The Turing test is more of an indication that we’re getting closer to general AI than an actual test that proves that we’ve reached general AI.
This is where I think a lot of the general public get confused it seems as if we’re getting closer and closer to a general AI that could take over all of our jobs but in reality right now we’re only perfecting the skill of Predictions based on lots and lots of previous data. The recent Google’s assistant is able to reply correctly is that it has the ability to remember millions of conversations both in the written word and in the spoken word and by being able to remember all that it can make a calculation a fairly simple calculation I would assume which is this. Based on what’s been previously said in this conversation what other conversations have I heard that match this one the best and then what was the next sentence in those conversations that has the highest match rating to this particular question.
So it doesn’t necessarily understand why it’s replying yeah let’s do five o’clock for four people but based on the millions of previous conversations that it’s heard it’s able to make a guess that the sounds that produce this sentence is the most likely appropriate response here.
You can imagine this as if you go to a local park and the first time you go you see a dog the second time you go you also see a dog and now you’re about to go for the third time which animal are you more likely to see between a cat and a dog.
you’re more likely to see a dog. So, if you had to guess you should guessed dog and that is essentially how Google’s assistant makes conversation but instead of only having gone to that Park two times Google has been to that Park a billion times at every millisecond of the day and seeing every animal that has ever visited that Park since it first started going. So Google can make the guess of dog or cat with vastly more information than you could.
Now I am adding several layers of abstraction to this in order to make it sound extremely simple but it is a very complex thing and I don’t want to take anything away from the engineers that actually develop this. but it’s not complex and the way that most people would think it’s a complex thing to take all of that data and sift through it in a short amount of time it’s a complex thing to create the calculations that actually determine whether a sentence is more appropriate than another. it’s a complex thing to develop all these algorithms that actually do all of this.
But the base principle behind it is still simply making predictions. This order system may very well be able to replace jobs like customer support. which to a large extent has been going on for years already pretty much anywhere you call today there’s a voice directing you to press one for English or to press three to get to the Department of Defense for enquiries about the jail tons of crack in the area 51 database and Google’s assistant is essentially just a better version of this.
Again the jobs that will be replaced are the more boring repetitive jobs and that will leave room for new more creative jobs.
Automation has been happening as a dawn of time man when electricity came along that put a lot of people out of work. lot of people carrying buckets of water , lighting lamps and all those kinds of things
So the question is not is automation going to eliminate jobs. there is no finite number of jobs we’re not like sitting around dividing up the same jobs that were around since the Stone Age.
So obviously new jobs are being created and they’re usually better jobs more creative jobs.
So the question is how quickly is this transition going to happen and what kinds of jobs will that be eliminated what kinds of jobs will be created. it’s impossible looking forward to predict what kinds of jobs will be created.
Programming in that sense is contrary to what many people think and it’s actually very creative building an app involves so much creativity that I don’t know that it can ever be automated even with general AI and if we get to that point then I don’t think you’ll be needing to write code anyway because that would mean that the AI can do everything for us. meaning that either it takes over and we all become slaves or some government controls it and we all become slaves or it does everything for us so that we are free to do whatever we feel like. Which can still be writing code and building apps but it may as well be mountain climbing or free diving.
So writing code is extremely complex and very difficult to automate because it requires so much decision-making but I do think that parts of the programming process can be automated. like for instance when you want to create a button there are already extensions available for this where they provide you the boilerplate code for that and that’s kind of how I think that AI will work for us in the future.
That these sort of systems will get even more advanced which will really only be good for us because they’ll mean that we’ll have to spend less time doing the boring repetitive tasks and can spend more time on the more difficult and usually more interesting tasks.
So if you’re considering becoming a programmer and you’re kind of worried that AI will take over your job before you’re done learning then don’t worry about that. because I think it’s very far off and it’s probably even questionable whether we can reach that point.
So I definitely think that you should try it