Intelligence vs Energy. Not Chicken vs Egg. The AI Native intelligent systems.
Since this is the season of abstractions, I have locked my marketing, sales and even investing tools away. Ofcourse I am loving it also as this is my natural inkling.
Now when I was building upon the argument of delivering the AI the world deserves yesterday, I wondered what had impeded it for so long. That led me to some more fundamental questions. And the q of "native intelligence". The kind that the human mind possesses. Or the DNA of any sentient system. Actually, I reached a question as to whether energy is a more fundamental quality in existence or is it intelligence as postulated in a thought, imagination, inspiration or even a basic calculation. Now don't you remember the burst of energy you felt when you were gonna start you company. Or start that project. The listening of a particular argument got you out of your lethargy and you became hyper energized. Don't you think the intelligence preceded the energy. I mean you were eating the same food earlier. Just inspiration or even everyday duality-ridden illusory happiness has the power to give you more energy. So you should get my question now.
The next step is to see how systems represent energy and intelligence. In the scheme of a power grid let us say. The water or wind moves the turbine and the coil in the generator moves fast to generate the multitude of electrons which then are volted up and down through the transformers to deliver the AC at your doorstep which you then convert and run your appliances. The native intelligence of a coil to throw out the trillions and gazillion of electrons was afforded by nature itself let us say. So that moved the needle. BuT. But there has to be an intelligence in the appliance now to recover that stream of electrons you so affectionately call electricity to make it function the way it ought to. The fan to rotate, the AC to cool, the bulb to emit photons and the coffeemakers to make you relax. So energy by itself cannot do much. There is a component of intelligence which makes the world go round. And that intelligence for me is the basis of creation. Our mind has it, out cells have it. The tree has it. And the recommendation system in you favorite social app has it now.
But the self driving car doesn't have it. Yet. It's tougher for AI systems to have it when there is a combination interplay of more devices like a car and the data in the case of driverless cars which sits outside and needs to be captured and collected separately. Or the much acclaimed Boston housing dataset I am hearing about. Or it's more advanced & alternative Ames Housing dataset. Then that data needs to be augmented or labeled. Then the product manager needs to understand how to integrate that data pipeline and make sense of it as an input. It is easier in the case of an app which just works with the ether like a streaming service. I mean the code has to deal with only the human mind as a target and which is in itself an intelligent system. So your AI will just replicate and talk to that intelligence. And the coder's life is so much easier. In contrast to a guy writing code in a silo for AI for a driverless car. So different AI systems shall have different requirements and would need to first define a native intelligence to fast facilitate the experimentation. You can call it customization like in an ERP. Like my friend said was necessary in case of a monorepo based code system which can't just write the code for a local maxima and not geared towards the company goal. The native intelligence has to encompass the vision of the founder and the management into the code repo so that the bad code can be flushed out from the system. They should continuously change the weights and biases to iterate the runs and get the visualizations and reports contunously in CICd fashion. And keep referring to that intelligence of the founder continuously like a CICD system. Only the code or model/dataset combination that has the intelligence vector pointing in the right direction with the company's vision automatically selects or is validated. Once these experiments can be done fast enough, the results would come. Just code or data is just unchanelled energy. What we need is a native intelligence on top of it which moves the set faster. And that is facilitated by newron.
What separates humans from apes is that intelligence. Apes and monekeys have raw unchanelled energy and hence they destruct. More so collectively. And the intelligent man harnesses the same energy. The most well lubricated isolated human intelligence brings about the harnessing of that energy towards greatness. Take for the sake of argument th case of all great scientists. Tesla, Newton, bohr and many more intelligence specimens harnessed that energy to greatness by channeling and invoking the intelligence component from the universe so easily afforded to man. And that is why Heidegger warned us against the idle chatter which is AntiDasein. For the DaSein understands itself. An efficient system shall capture the intelligent inputs quickly and at every step and make outcomes possible. An ML flow highway will do that for participants.
Point is, AI shall be delivered as the world deserves it when energy is augmented by intelligence. Bad code and data combination is just unchanelled energy and it needs the component of intelligence of that native system of participants delivered fast enough to channel it into an AI. Like native apps performing better than web apps. Like apps in iPhone.And that intelligence is harnessed/invoked by continuously asking the right questions till the answer reveals itself as captured beautifully by the founder at neuron in his rumination in another piece you can check. To change the weights and biases as a change in the questions. To keep iterating faster with intelligence based inputs from the top management as advised by my other friend. And the neuron highway facilitates that process at a fast clip. That is how we get you to build a native intelligence framework for your system. Then the answers reveal themselves. I think the argument is complete. To deliver the AI the world deserves, we need to give it the native intelligence in a code system it needs.