Why wouldn’t you love this debate?
Elon Musk vs. Mark Zuckerberg.
The debate of AI.
Who is going to win this debate? And I don’t mean who is the person that is going to correctly predict the future of AI, but rather who is going to win this debate for the sake of this debate. Well, I think you’d have to go with Elon. For a guy who market to the world about immigration to Mars but in substance building a very down-to-earth business of satellite launching (as NASA de-monopolises the business).
We all know computers have been progressing at a crazy pace and we have all witnessed what it has accomplished in the last 20 years. And we have no reasons to believe that it can’t do even more spectacular things going forward.
But there are a few questions I have with AI.
What is intelligence?
To talk about artificial intelligence, one should start with intelligence. What is intelligence? Computers are basically very dumb machines (yeah, they only know 0s and 1s) but they can do very dumb things (instructions) at an incredible speed. What’s more is that they have very good memory. With these two advantages, they are great at automation. Speedy (repetitive) automation. And if the algorithm (a smart collection of dumb instructions) is great, that speedy automation is that much more impressive.
Yet, it’s just speedy automation.
For example, there is no significant difference in recognising a barcode and a physical grocery product, from an intelligence standpoint. The only difference is that barcode is simple and structured so the design of that algorithm is relatively easy. But ‘recognising’ a grocery isn’t all that impossible, as long as computers can continue to develop unbelievably quick and cheap processing power and storage, some kind of ‘structure’ will emerge. At the end of the day, they are only different from a technical standpoint (no disrespect, you still need great engineers to make this a reality) but essentially the same from an intelligence standpoint.
So what’s intelligence?
I think the biggest benchmark is the processing of language. No, not the chatbot that people are talking about these days – all of them are still just speedy automation to a pool of complexly associated responses – but the real-time translation of language (including reading, listening and speaking).
And as we know today, Google Translate still suck really really bad (for those of us who are bi-lingual). The intelligence of translation is still really really low. Sure, voice-to-text (converting human form into machine form) is improving, but text-to-voice (converting machine form into human form) still sound very robotic. In another words, it’s easier for us to explain to computers rather than computers explaining to us. So who is more intelligent?
It’s ironic that even the most advanced robots in any sci-fi movie has a robotic sound (think terminator).
But let’s take a moment to imagine. Imagine one day we have real-time language processing. We won’t have language barrier. We won’t have illiterates (everything can be ‘read’ to them in real-time). And we won’t have inequality of access to information (yes, most good Internet materials are in English). We are talking about a truly liberated world.
A real democracy.
So only when I see real improvements in language processing would I believe in this AI hype.
The next big thing
If we look back at history, we have never been blindsided by huge technological change. Big companies do, but as a human race, we don’t. Most technological changes impact our human race much like a huge change program at large multi-national companies. Sure, technology is there but it takes a lot of time and ‘change’ management’ to make things gradually happen to us.
You may say, how about the Internet?
Yes, we are all addicted to our mobile phones. But think about what we do on our phones. We listen, we read, we have fun and we communicate. From a human development standpoint, these technological advancements are best ‘sustainability innovation’ according to Clayton. Instead of paper, we stare at screens, instead of mails, we have emails.
So what has really changed at the macro level for us human beings?
This brings us back to the question of intelligence.
Have we become more intelligent? Or may be all these technologies have made us less intelligent? As computers will always going to be controlled by a minority of human race, maybe this is how computers are going to take over from us (the majority) – a minority dominating the majority?
So what has changed?