Tech Thoughts: Learning to learn about technology, focus on repeating concepts
A short summary of my recent mental focus, and how to build systems for keeping up with new technologies more effectively
“I have a dream for the Web [in which computers] become capable of analyzing all the data on the Web – the content, links, and transactions between people and computers. A ‘Semantic Web’, which makes this possible, has yet to emerge, but when it does, the day-to-day mechanisms of trade, bureaucracy and our daily lives will be handled by machines talking to machines. The ‘intelligent agents’ people have touted for ages will finally materialize”
Sir Tim Berners-Lee, 1999
Tech Thoughts: How to Identify What You Need to Learn
Trying to keep up with technology often feels like drinking form a firehose - there is just no way you can consume it all.
But it brings me back to my high school physics teacher, who always made us start at the end of any chapter. That way, we would work our way backwards to identify the critical concepts to learn in order to solve the really complex stuff
And I think the same goes for keeping up with technology. Learning everything from “first principle” can often take forever, but identifying the key recurring concepts can take you a very long way
E.g. Keeping up with all new Generative AI solutions out there is an impossible task (at least for a mortal human), but understanding the foundations of how an LLM really works from a technical perspective makes you infinitely better positioned to assess new investments than if you try to understand the whole market by doing company-by-company comparisons
To do this, I’m developing a list habits I try to be pretty disciplined about. Here are two of them:
Writing down every new technical term, but focus on the terms that repeat
I try to keep tab on all the technical terms that come up in the blogs, newsletters, pitch decks, investor presentations, founder calls, and twitter posts I come across
The bar for this list is very low: “Could I explain this term to a CEO in 30 seconds?”
If the answer is “no”, or “maybe”, it goes on the list
Once a terms ends up on the list multiple times - I research it
While share research questions very openly
I was close to starting my PhD in Medical Robotics after my undergrad, but ended up on a different path.
Thus the chance to pursue an academic interest far down the rabbit hole probably slipped
Instead, I try to find as many PhDs as I can to answer my longlist of questions
And every time I research something, I try to force a list of follow-up questions.
Then I try to share the list as widely as I can
And so far, I have found that people love to share what they know - especially if they are the top experts :)
What I’m reading
Technology
How Transformers Work - The Neural Network used by Open AI and DeepMind
The Eroding Technical Moat of AI and the Power of Open Source
Intel Announces Aurora genAI, Generative AI Model With 1 Trillion Parameters
Fine-tuning LLMs Made Easy with LoRA and Generative AI-Stable Diffusion LoRA
ColossalChat: An Open-Source Solution for Cloning ChatGPT With a Complete RLHF Pipeline
Transformers, Explained: Understand the Model Behind GPT-3, BERT, and T5
Software 2.0 - Andrej Karpathy
Startups / Venture
About Author: After training as a robotics engineer, I spent nearly a decade as a consultant working on strategy development for global corporations and startups
Now I spend my time strategizing on the Future of Tech, building systems to faster identify and understand new technologies, and thinking about how new technologies will impact sectors and business models across industries
You can also find me and more of my thinking on: LinkedIn, Twitter, Website
Thanks for reading! Subscribe to stay in touch


