top of page
  • Writer's pictureSaurav Bakshi

Journey to the AI Frontier: The Trio of Skills I'm Cultivating Next - Introduction

As an avid learner and a person with insatiable curiousity about technological advancements, I was completely blown away by the fascinating promises and hype created by ChatGPT. Now you are thinking - no way, another post about ChatGPT! Just the keyword alone produces 764,000,000 search results as of today on Google. Please bear with me and stay with me as I try to unfold the details through the eyes of a Technology Architect.

I guarantee a focus on genuine technological implementations that stand to gain from ChatGPT, sans any exaggerations or unnecessary hype.

To set the scene, let me share a short anecdote about my propensity for excitement when it comes to breakthrough tech.

My current enthusiasm for technologies like ChatGPT and Dall-e (incidentally, that image of the dog was crafted using Dall-e) still doesn't quite match the fervor I felt back in the summer of 2016. Note: by "summer," I refer to the Australian December, for readers from the northern hemisphere. During this time, I stumbled upon an exclusive training program by Udacity on Self Driving Cars Engineering, conducted in collaboration with Google. This coveted course was open to a mere 200 applicants globally.

Fast forward to now, and there aren't such stringent entry barriers. A solid grasp of numbers and python programming skills are the primary prerequisites. Despite my then-limited grasp of machine learning and terminologies such as convolutional neural networks, gradient descent, and adam optimization, I couldn't resist applying. I banked on what I believed were my exceptional capabilities. Two weeks later, while wrapping up at work, an email from Udacity dropped into my inbox, announcing my acceptance into the program. The elation I felt was astronomical, quite literally akin to what Neil Armstrong must've felt setting foot on the moon. With my head in the clouds, I set off for home, eager to share the news with my wife. But in my heightened state of joy, I mistakenly boarded a train on a different line. It was only at Newport station that I realized I was heading to Williamstown instead of my actual destination. My only option was to journey to Williamstown and then backtrack. The humor of this situation peaked later in the week when a friend informed my wife that her husband, a station cop, spotted me at Williamstown. It's a tale we often revisit for a hearty laugh.

Just like any tech aficionado, our emotions traverse the entire spectrum — from ecstasy to sorrow — in our pursuit of the latest in technology. There's a particular gleam in our eyes when discussing groundbreaking innovations.

Back to my ordeal above - I failed and dropped out. However the failure taught me a lot. A pivotal realization was my lack of proper planning and self-assessment. Did I truly understand my strengths and how they would aid my success? Or recognize the weaknesses that might hinder it? Was I adequately equipped resource-wise? I recall my computer's GPU being incompatible for model training, and upgrading wasn't an option due to financial constraints. In hindsight, I see I acted impulsively, much like a child drawn to a shiny new toy.

For tech consultants like myself, emerging technologies are akin to novel tools designed to address business challenges.

Just as businesses often employ speculative, explorative, or evaluative lenses when eyeing new tech trends, we, as technology consultants, must adopt a discerning approach to new waves of innovation. It's not about our ability (or lack thereof) to grasp new concepts, nor the effort it requires. In fact, my personal challenge is that I try to learn almost too much to remain at the forefront. Sometimes that becomes counterproductive but learning, to me, always means fresh perspectives, challenges, and problem-solving strategies.

Now, turning our attention back to Generative AI. Its meteoric rise is undeniable. Our days are awash with fresh insights, academic papers, expert viewpoints, tools, software, and courses centered on this tech.

John Maeda’s course on LinkedIn Learning titled "Introducing Semantic Kernel – Building AI-Based Apps" offers an apt analogy, likening AI advancements to ketchup being coaxed out of a bottle. It starts slow but ends in a sudden, substantial splash. This mirrors the progression of language-based AI. For context, innovations like Data Augmentation for NLP, the transformative power of transformers from Vaswani et. al.’s paper – “Attention Is All You Need,” and Reinforcement Learning with Human Feedback (RHLF) are all relatively nascent, all emerging within the last decade. From my vantage point, having delved into Machine Learning and Deep Learning, the pace of this technological evolution is unprecedented.

The rapid developments in the sphere of GPT/Generative AI are daunting for learners. Everyone seems to have unique learning advice or boasts their accomplishments on platforms like LinkedIn — something I, too, am fond of to maintain my online presence. The current discourse is saturated with Generative AI.

Guided by Benjamin Franklin's wisdom — "By failing to plan, you are preparing to fail" — I aim for a methodical approach:

· Setting clear timeframes and scopes.

· Striving for impactful outcomes.

· Ensuring the journey remains enjoyable.

My posts will chronicle this journey, a countermeasure to the overwhelming influx of information.

Every exploration begins with envisioned results. And while influenced by available resources and time constraints, the essence remains consistent.

Countless management consulting frameworks exist to aid problem-solving. A prominent example is the Seven Step Problem-solving process by Charles Conn and Robert McLean in their book “Bulletproof Problem Solving.”

But this piece isn't about problem-solving. It's about a technologist's quest to discern the merits (or lack thereof) of Generative AI as a panacea for technological challenges. What current solutions can it provide businesses, and which businesses stand to gain the most?

It's an expansive topic.

Given my tendency to delve into details, I'll channel my focus more sharply to avoid being swamped by the information surge. My ensuing post will define the core problem: Why, as a technologist, should I immerse myself in Generative AI, and how can it enhance my work?

Thank you for bearing with this extended post, and I genuinely hope it offers insights to fellow mid-career technologists and even those just starting their journey.



bottom of page