ANDREW FOSTER;332977 wrote:I asked a similar question rather more bluntly here...
https://moneyforums.city...p---prove-me-wrong.aspx
I'm still failing to see an AI 'killer app', while I am seeing a lot of routine programming being masqueraded as AI along with some for real brain deads like the 'AI Toothbrush' and 'AI washing machine'
"Open the washing machine door HAL"....
So I'm intrigued to know as well.
There is no doubt that LLM's appear to be clever, but that isn't the same as 'AI'. LLM's didn't 'learn' to give the responses they do, they were coded with a set of rules and given a vast ammount of data.
The picture/video generation is also very impressive. The poetry...not so much....
I was told by a know it all school child that
Just like mankind has evolved via information, knowledge, failure and rules; AI too will will evolve over time for behind it is all those same principles. The biggest and most effective change is likely to be in the area of automation and robotics.
When that phase is here people will need to make a decision, does mankind choose ultimate capitalism over an element of socialism or redefine mankinds role in society.
Think self driving or robots driving cars, at the moment around globe many are earning a living by driving people around from A-B. Now suppose you have some software or robot doing the same job what do all those drivers do (they cannot all be coders and software engineers - robots and cars can be built by automation).
On the other end, think of car owners, many currently spend a large sum on one, park on their driveway, use it to travel to and from work. Majority of the time that huge outlay is parked up somewhere (up to 8 hours a day for those who simply use it to drive to and from work - depreciating at a fast pace). Now imagine investing in a car and getting it to drop you to work and then the car also earns until you are ready to be picked up again from either a highly specialised job or a job for which robotics and automation has yet to teach itself due to the amounts of data available.
So yes AI is nothing but vasts amounts of data which is being processed with pre-coded rules and instructions. In much the same way that we human use data and turn it into information and knowledge to help navigate our surroundings. The more data we are able to process and the more relevant the question we pose/input/frame the better the response and better we are able to navigate at speed. Think of a black and white striped horse in far away lands - which today a Zebra.
From which my take away was that
- Man focuses on the question - "Why" (taught to me as the key to problem solving and real knowledge)
- AI focuses on the question - "if" followed by a an element of "mathematical probability" for extrapolation
- From which it gathers more data which only changes the "mathematical probability" element but not the core "if" starting point.
So the AI is ultimately as good as the questions it poses and the data it has access to - probably why Deepseek know nothing about Tiannamen Square.
Now which companies in the world has vast amounts of data which they can leverage and use for the current phase of this AI journey.