HOW MUCH BETTER DO YOU WANT TO GET?
Is the fear of AI an issue of us not wanting to change, or is it an issue of poorly designed interfaces that leave us frustrated? But if you want to change, how much better can we get? We are taking a look at the financial industry, but the thinking can be applied to all industries.
If AI is becoming a competitor, then what makes it so special? We better get to understand it well, as most of us do have to pay the rent. We are putting the AIs in the following categories, based on the approach to data processing:
- Basic statistics, which nobody would have called ‘AI’ 10 years back. And this is the vast majority of offerings out there. It is all about decreasing the margin of error, but the machine is doing pretty much the same things that humans have been done for the last decades. Just faster and with fewer errors.
- Programs able to “fine tune” the data in order to make it “structured”, so that it can be used in various models for processing. There is an increased amount of offerings in this category, but still very few. These machines are able to adapt “on the fly” to changes in variables and environments, as well as dealing with imperfect data.
- Programs able to deal with unstructured data. These are very few, and here is where the current cutting edge research is happening.
- Brain like AI. This doesn’t exist yet. We know of dreams in this field and expectations, but nothing concrete that comes even close to reality. Many see this becoming a reality around 2040 or 2050.
Based on the above, there are very high chances the main AI competitor in the financial markets will be standard statistical models that can execute things automatically, like valuation screens, performance analysis, portfolio allocations, and so on. There are only two variables here that the “AI” is focusing on: maximizing speed and decreasing the margin of error. So how can you compete against such an AI? Well, you are not going to get faster than the machines. But what if you can use the machines to get faster and better?
And this is what we are trying to focus on. Since the human brain remains the best “processing machine” out there, why not focusing on increasing its productivity exponentially, instead of replacing it? The brain is already capable of processing huge amounts of data.
Here is the core of our focus: what if we had the tools to feed the brain with the right data and information, how much better could we get? I am talking about the categories 2 and 3 of AI above, and especially 3. Most of the brain activities, thinking or not, are used on very standard processes, because we actually need to find the right data in order to make important decisions. The so called “right” data is never structured, hence the usual “AI” (category 1) cannot process it, because it does not understand it.
How long was your research for the last great investment you made? Chances are that more than 70% of the “thinking activity” is used on the “discovery process”, another 20%+ on sorting the quality data, and only very little on actually deciding.
Now, what if you had the tools to maximize your decision process, to help with the discovery and sorting process? That is a 10x improvement overnight, but as your capability to decide gets faster, productivity can jump a lot more. Better still, you could be years ahead of the current “AI” applications, as well as competitors, because you can deal in unstructured data. So how much better do you want to get?
We are hoping to move into public trials (we are behind schedule) in October, and if anything goes well, will be updating you on the progress of such tools later in the year.