Have you ever wondered what generative AI really is? In other words, what is the product?
We know it can generate text answers. There are image generators which output pictures in just a few moments. The early stages of video creation is also upon us.
What does all this mean?
To me, gaining a grasp on what we are dealing with is imperative. There is a lot of noise yet, when we get to the basics, things become clear.
For example, what is the Internet?
If you ask 100 people, you will get a lot of different answers. Because of this, we have a lot of confustion.
The Internet is really just a series of nodes, connected together, that send and receive data. That is it.
When we view the Internet in such a way, the mystery disappear. It is nothing more than a connection of computing devices.
That is pretty simple. From here, we can build whatever mental models necessary to opreate within this reelm.
Image generated by Grok
When we think about Grok or ChatGPT, what are we dealing with? If we try to break it down to the basics, where do we land?
Getting the proper framework is what allows us to develop other ideas about where things are heading. The concepts derived from our basic understanding become clearer.
For example, if we utilize that basic premise of what the Internet is, cloud computing makes sense. We can see how the idea of servers providing processing and storage can be accessed by smaller devices is feasible. It is no longer a mystery.
With this in mind, we go back to what is generative AI?
In the example mentioned above, there is one constant. No matter what the media of output, we are still dealing with the same premise.
Generative AI is nothing more than units of cognition. This is what is created. Cognitiva outputs, similar to what people do, is the result. This is true whether we are dealing with video, text, or the driving of an automobile. Cognitive units are created via an artificial means.
This is the baseline from which I always start. It is also what allows us to look into the future with some certainty.
What happens if you start a brainstorming team with 5 people? You will likely get some ideas about whatever the topic being discussed. next, double the size to 10 people. What happens at this point? It is likely that the number of ideas flowing increaes.
Hence, we get more cognitive output with more people.
Using the same scenario, what if the extra 5 people we added were graduate students from Harveard who majored in the topic at hand. At this point, we not only increase the number of ideas, it is probable the quality of them also moves expands.
Thus, we get better cognative abilities the more knowledge that is added to the equation.
Can you see where this is going?
With generative AI, we see the ability to produce a larger quantity of output. Due to the incredible processing speed, along with the hundreds of millions of users, output is increasing significantly.
The key is what happens as the congative ability increases. With each new interation, the models improve significantly. This is due to the addition of more data, better algorithms and more compute.
Instead of dealing with an elementary school student, we move to high school, college, and eventually post-grad work.
Accroding to AI metrics, this is the path that we are following. Depending upon the model, and the target area, we have some operating akin to a grad student. Of course, this is early 2025.
What happens by the end of the year when we have some models that are two updates further than now?
To me, that is when everything changes. The entire economy is going to be upended as all knowledge work is targeted. Workers in many of the fields will find they are left behind since the AI will simply eclipse this over a short period of time.
Do not size up what AI can do today. Instead, look at where it is going. That is the key to grapsing some of its capabilities.
