A Pythonic Evolution
​
Tomorrow's world will be powered by Python.
​
Why ? Because Python enables the fastest extraction of data to create meaningful value drivers. Extract fast, test fast, model fast, ffail fast, refine fast. Repeat.
​
Today's LLM's (Large Language Models) can hoover vast amounts of unstructured data, combined with essentially unlimited computational power.
​
Just like today's raw mineral extractors - oil, gas, coal - tomorrow's winning firms will be those who extract, mine, model and refine data in the fastest way possible. In the same way that natural organisms extract data from their environments in order to survive, grow and flourish - data becomes essential to the growth of tomorrow's organisms. This is not restricted to technology organisations either - all companies big and small need data to survive.
​
Python offers that.
Artificial Intelligence, Machine Learning, Deep Learning (and ofcourse Generative AI) rely on data to power their predictive and generative capabilities. Today's data leaders - Microsoft, Google, Amazon - are those who extract the most amount of data possible to power their models tomorrow. ​They also employ the best and brightest to get better data extraction than yesterday.
​
As data moves faster, so will the underlying algorithms powering neural networks and LLM's. The faster the data extraction, the more modelling potential will be possible - hence the predictive capabilities become even stronger.
​
Python's lightning quick language and libraries allow the extraction process to be performed efficiently and within a proprietary environment. The only defensible position will be to gather, clean, label, categorize and train models the most efficiently - no matter where the data comes from.
​
In the absence of relying on this gathering, the exploration of proprietary methods of gathering our own data will become the mission.
The future of data is exploratory. Trained models that can extrapolate in a variety of directions from information that is entirely fluid in its input. While there are stable recurrent aspects data, that state is in constant flux. We are quite a distance from an organisation that can relentlessly clean data, label data and train models no matter from where the data comes from - an organism that may be more based on physics than maths.
Not at speed - yet.
In the era of data abundance, the people who can build models, refine models and execute on them fastest are the ones who are going to win. They are the chaos agents in the ecosystem.
​
Propelled by advanced neural networks, anything digital - finance, entertainment, medicine, sport - is about to become hyper-personalised - down to the most intricate DNA of each and every personal transaction. That which can be measured can be mined and can therefore be predicted.
​
Exponentially learning with each neural iteration, models will augment human experiences to become infinitely more powerful than pure human judgement.
​
Welcome to Fintech.ai. Welcome to Pythonic Thinking driving data and value.
​
The companies that can move the fastest and eat as much data as they possibly can get their hands on are going to be the winners in the ecosystems.
​
​
Peter Toumbourou
Charleston Advisory Group
2023
​
​
​
​
​
​
​