When people contest what is going to happen with technology, do they really know what they are talking about?
This is something that is a double-edged sword. Of course, those who are making predictions all think they know what is going on. However, as this video shows, someone was apt to be wrong.
Give this one a view since it explains what we are dealing with.
https://inleo.io/threads/view/taskmaster4450le/re-leothreads-2mhmxvyvj
Poor Bryant found himself on the wrong side of history. It seems that, even back then, the media was pushing an agenda. Gumbel was convinced computer sales peaked. When he was told that wasn't the case, his retort is we are looking at different data.
He also mentioned "it is not rising at nearly the rate it was going". Of course, Gumbel turned out to be wrong.
Gumbel can be forgiven since he is a media personality. The other one, who supposedly was an expert, was off by a few orders of magnitude.
The home computer akin to the home video camera. It didn't quite turn out that way.
It is easy to see this in hindsight. The problem is Dertouzos explained how the value would be driven. It wasn't as if he made a baseless prediction. He described what would be required.
Technology Changing The World
Either one understands technological progress or he or she has no idea what is taking place.
We are immersed in technology curves that are going to have a profound impact upon what we are doing over the next decade or two. Those who look at things as they are now and believe this is going to be reflective on the world ahead of them is akin to the analogy in the video.
A lack of understanding of the convergence that takes place is one of the biggest flaws people can make.
Whether it is Hollywood, automotive, real estate, finance, or jobs, the trend are becoming clear. Most contest it until it becomes abundantly clear.
It is like discussing the rise price of film for camera when the switch to digital is underway. Who cares about film prices in a digital world?
AI is going to disrupt the world. To think otherwise is going to get completely run over by it. By the end of the decade, if AGI is achieved, most knowledge work will be done by computers. The best analogy I heard is we are dealing with a 100 story building and, at the moment, we made it to the second floor.
In other words, we are in the truly early stages.
Some technologies take a long period of time to emerge. This is especially true for anything to do with atoms. Bits move a lot quicker. That is why digital platforms can far outpace this industries mired in the physical realm. It is also why companies, especially in manufacturing are moving towards the use if digital twins.
Moore's Law On Steroids
The reason why we can make bold predictions about AI, as an example, is due to the basics of what is taking place. As the headline about shows, this is accelerating at an incredible pace. Of course, we can see this is not a small sample size, covering the last 14 years, since 2010.
This puts Moore's Law to shame.
For those who don't know, that was an observation that said compute basically doubled every 2 years. That basically held for the last 60 years. It is also why we saw so many gains in computing over the last 40 years.
We are dealing with something that is doubling every 6 months. Granted we are dealing with total compute which is then rolled into the size of the models but it does show the impact.
Here is how the result look on a chart:
Source
Notice the bend in the tend line around 2010. What happens since then? We saw massive acceleration. Ironic that one of the lowest things on the list is robotics, something that is about to be supercharged.
This is why we will see a huge difference a decade from NOW. When it comes to chatbots, probably the form of generative AI that most are familiar with, we are around 20 months since the release of ChatGPT 3.0. We are not even two years into this and look at the progress already made.
Go back to the video.
The assertion was that personal computers would amount to nothing. Consider what is a computer these days. If we go with the simple view that anything with a processor can qualify as that, we are talking about a miss of tens of billions of devices.
How far off will people be who state that AI is nothing more than hype? What about those who believe it is being pumped and doesn't have much utility?
The reality is that AI is another form of computing. People want to correlate it to the Internet was the connecting of computers. AI is actually an advanced for of computing. Like the early programs, there is still a great deal of room for advancement.
That is what gets overlooked.
And just think, if the long term trend holds true, in two years, the amount of compute dedicated to training these models will be 16x where they are now.