Insights from “The Next Decade in Data, Intelligence, and Technology” a Keynote by Rafal Lukawiecki at ESPC 2023

In the ever-evolving landscape of data, intelligence, and technology, staying ahead of the curve is not just a choice, rather a necessity. Every year since 2014 I make a late fall trek to Europe to attend ESPC, the premier European focused Microsoft cloud technology conference, at least when they are able to host them. Each year, they invite a keynote speaker that looks outside of the direct sphere of the Microsoft cloud, past keynotes by Paula Januszkiewicz being some of my favorite.

This year, I had the privilege of attending an insightful keynote , “The Next Decade in Data, Intelligence, and Technology” by the thought-provoking Rafal Lukawiecki. His outstanding delivery, a combination of whit, deep tech, humor, and expertise, delved into the transformative power of artificial intelligence, the impact on our jobs, and the crucial role of data in shaping the future.

AI, in particular the latest round of generative AI, has peeked my interest over the last year, like many of you. I consider myself to have a general understanding of where we are with AI at the moment. Mr. Lukawiecki opened my eyes to so much more.

The Transformation of Jobs

The advent of AI will inevitably affect all our jobs. However, the key lies in preparation – those who are ready to embrace the change will witness positive transformations in their roles. Those who don’t are likely to fall behind, or will be a part of workforce that is replaced. As technology evolves, it becomes imperative for professionals to upskill and adapt to the changing landscape.

The Nature of Artificial Intelligence

The question of whether AI is truly intelligent sparked an intriguing discussion. Lukawiecki highlighted that generative AI operates on probabilities and does not possess true cognitive thinking. This insight challenges our perceptions of intelligence and prompts us to explore the boundaries of machine learning.

What is intelligence? If AI has two primary methods for “generating” something new, that being probability or logic, do humans not also use probability or logic to “generate” an answer? As Lukawiecki demonstrated, do we not use probability to determine if we should take an umbrella when we go out today? Do we not use logic when working on a problem? When would or could AI actually be, what we would agree, be intelligent or conscious?

Developer Productivity and Copilot

For developers, the integration of tools like Copilot is becoming increasingly essential. Lukawiecki warned that those who neglect to leverage such tools may find themselves lagging in productivity. For example, Github Copilot, an AI-powered code completion tool, is changing the game for developers, enhancing efficiency and reducing manual coding efforts.

Copilot can help developers very quickly generate boilerplate code we use throughout most of our code. Truly, how much of your code is actually interesting? Validating input, communicating with API’s, normalizing data, this is all mundane. Copilot can write really good code for much of what we do.

Therefore, thanks Copilot for handling all of the easy stuff, the code I used to spend 60-80% of my day coding and debugging. Now I can spend 100% of my day on all of the big problems! No time for a mental break☹

Unintended consequences? Possibly.

Embracing Challenges

Lukawiecki’s keynote encouraged us to “enjoy the ride” while acknowledging that challenges lie ahead. As advancements in data, intelligence, and technology accelerate, professionals are urged to brace themselves for a period of increased complexity and difficulty.

Paraphrasing the often used expression, “May we live in interesting times”, indeed.

Taking the Helm

In the rapidly evolving landscape, Lukawiecki emphasized that individuals are the captains of their own destinies. Being proactive and taking charge of one’s career trajectory will be crucial in navigating the uncertainties that lie ahead. This next year appears to be the “Year of Copilot” at Microsoft. I for one am a huge fan of the naming, Microsoft got this one right. Copilot is not the pilot, it is not the captain. That is us! These tools are here to help us, we are still in charge.

Logic vs. Probability

The keynote shed light on the fundamental difference between logic and probability and its relationship to paths to implementing AI. Logic driven AI requires constraints. Probability based AI, the kind used by GPT, OpenAI, and most other forms of AI we are hearing about, relies on data. Lukawiecki highlighted that future programming will involve creating guardrails that balance these two elements to achieve optimal results.

Probabilistic AI is able to provide a response to most any input because the response is based on probability of that response being good or not. Any probabilistic response to an input is by its nature inexplicable. No one can ever say why a given response was given when using probability. With the data set size we have today, GPT won’t generate the same response twice, for anything but the most simple inputs.

When generating a response to a casual input at ChatGPT, such a case is ok. What do we do though when an AI driven car using probability injures someone? No one can say exactly why the car did what it did, so who is responsible?

The thing about probability based AI is that it is rather inexpensive. Getting all of the data together can be challenging, though once you have the data, generating responses to input is not difficult.

Logic driven AI is something different all together. Logic requires rules. Creating a ruleset for a self driving car for every possible situation is essentially impossible. There are an infinite number of possibilities. If an AI model is based only on rules, then that model may only provide a response if there is a rule to dictate the output. That doesn’t scale well. As Lukawiecki said, current approaches to computer hardware isn’t up to the task for fully logic driven AI.

The Black Box Dilemma

Addressing the ethical concerns surrounding AI, in particular generative or probabilistic driven responses, Lukawiecki touched upon the black box dilemma. Probabilistic responses generated by AI may be inexplicable, raising questions about transparency and accountability. Implementing guardrails becomes essential to ensure responsible and ethical AI development.

This aspect of the current state of AI was one of my primary takeaways. Most of the AI we are hearing about right now is based on probability. In this form of AI, a model is trained based on lots of data, then this model can be used to provide a possible response based on a given input. The response provided is based on a guess, a tokenized response where each token was added into the response based on the probability of that token being the best token at that moment.

Quantum Computing in the Future

Looking ahead, Lukawiecki envisioned a future where quantum computing becomes indispensable. Current processing capabilities may not be sufficient for the complex requirements of AI and machine learning. Quantum computing, combining logic and probability, holds the potential to revolutionize the field.

I see quantum computing akin to fusion based power or the grand unified theory. I really want to see quantum computing work, based on the little I know, it sure seems possible. Yet we keep waiting, the same as we keep waiting for other really cool technology.

Trusting Human Common Sense

Despite the rapid advancements in AI, Lukawiecki urged caution and stressed the importance of not solely relying on machine learning. Trusting in human common sense remains a critical factor in decision-making and problem-solving.

Do not prompt ChatGPT to write an email to your boss based on some prompt, then copy, paste, and send. After you paste, read the email, edit it, use common sense, your own voice, and your own expertise. Then send!


As we navigate the next decade in data, intelligence, and technology, Rafal Lukawiecki’s keynote served me as a compass, insights that can help guide us through the challenges and opportunities that lie ahead. Embracing change, staying vigilant, and actively participating in the evolution of technology, in particular AI, will be the key to success in this dynamic landscape. Let’s embark on this journey with enthusiasm, armed with the knowledge that we are the captains (not the copilots) of our own destinies.

Speak Your Mind