Accessing the Universe: The Revolution of Knowledge in the Era of GPT-4
A PhD equivalent of research in a weekend, the new norm
Greetings, fellow knowledge seekers! Just yesterday, as I was delivering a talk to a bunch of spirited grade school teachers, I found myself reminiscing on the evolution of our search for wisdom. Believe me, it's a trip down memory lane that will leave you nostalgic for the smell of old books and dusty libraries, yet excited for the future that's unfurling right in front of us.
If you, like me, have ever found yourself knee-deep in the chaotic beauty of an old school library, then you'll remember what a mission it was to write a report. Let me jog your memory: you'd boot up the local computer system, scour the database for books that might be vaguely related to your topic, and then embark on a scavenger hunt through the towering stacks of literature. When you finally managed to assemble a promising pile, you'd haul it home and spend the best part of a week diligently pouring over your collection to finally write your report. It's hard work that built character, right?
But then, as the technological tides shifted, more and more content started migrating to this thing called the internet. Oh, how we embraced this new frontier in college! No more lugging books home, no more late fines; now, we had an abundance of resources right there on our campus computers. But let's not sugarcoat it – the search results were as clunky as a rusty old typewriter, and we were just scratching the surface. In this vast sea of information, too many valuable pieces of knowledge slipped through the cracks due to our lack of sophisticated search capabilities. Take, for example, the brilliant deep-learning research by Benjamin Graham, a name unknown to most in the field, but I'd wager his work could stand toe-to-toe with the likes of Hinton and Lecun.
Even closer to home, consider your graduate thesis – a labor of love, an intellectual triumph. How many eyes have seen it, truly appreciated it? I remember my own was quite niche: “METHODS FOR STREAMLINING EXPENSIVE FUNCTION CALLS". I stand by my belief that it holds some truly insightful nuggets of knowledge, yet how many people will ever benefit from it? How many people will be unable to find it even if it would be helpful for their research?
Times are changing ever more quickly now, with the advent of large language models (LLMs) like the marvel that is GPT-4, we're looking at a whole new ball game. Now, it's like having the entire library at your fingertips, but it's not just about the books you can read. It's about the code you can write, the obscure functions, libraries, and even languages you might not even have heard of – all of these can be handed to you on a silver platter. Sure, there might be hiccups; the AI can get things wrong, but to me, the thought of missing out on something essential is a far graver concern. I'm all for cross-checking and verifying what the model suggests.
To put it into perspective, we now live in a world where someone can whip up a PhD level thesis over a weekend, references, all of it. 20 pages? 50 pages? 200 pages? 1000 pages? Yes, yes to all of that. A world where the innovation we'll witness this year alone will blow us away. A world where GPT-4 isn't just a tool, but a game-changer that's going to play a massive role in our ability to keep up with our own ever-expanding knowledge base. Now, isn't that an exciting future to envision? A future where our efficiency in assimilating knowledge increases into the double digits. So buckle up, friends, because the ride is only getting started.
If you aren’t spending 30-60min/day using these newer technologies you are wasting time, and time is precious. I’d much rather spend that doing something I enjoy that poking around the internet scratching the surface of what I need.
This entire blog from concept to writing and graphic took 26min
Next week it will take 14min, buckle up.