OpenAI's ChatGPT as the ultimate form of abstraction
By Dann Berg
Published or Updated on
DALL·E: “the personification of an artificial intelligence, as a personal assistant helping a person complete their work”
I’ve never before, in my life, witnessed the release of a new technology that I knew was going to fundamentally change the world. The impact will be as large, if not larger than the Internet itself, and even the Internet took a couple decades to catch on and change everything.
I’m talking, of course, about artificial intelligence. Specifically OpenAI’s ChatGPT. In case you missed it, this month an AI chatbot based on OpenAI’s GPT-3.5 language model (an improved version of GPT-3) was released to the public, for free (for now).
We’ve heard rumors of these types of AI before (I even talked about the ex-Googler who quit because he argued the AI was actually sentient in a past newsletter) but never had a chance to play with one before.
It created a buzz the likes of which I’ve never seen before. I think Paul Graham says it best (emphasis mine):
The striking thing about the reaction to ChatGPT is not just the number of people who are blown away by it, but who they are. These are not people who get excited by every shiny new thing. Clearly something big is happening.
It’s fun to play with, but has real-world practical uses as well. I’ve been using it since its release, and it’s already helped me finish writing an app that was only about 30% done after having worked on it for months. It’s also provided a fantastic assist with writing some complex excel formulas.
Many people are looking at this technology and thinking that this AI is so good that it will soon replace us. I strongly disagree.
At its core, this form of AI is really just a radical new form of abstraction.
The abstraction that I’m talking about here is a concept that will be familiar to developers, while less concrete to everyone else. Abstraction is best thought of as a ladder. And the higher up the ladder you go, the more abstract the idea; the further down, the more concrete.
A coffee machine is an abstraction layer. All you need to know is how to use a coffee machine. You don’t need to know how the beans were grown, harvested, roasted, and ground. You don’t need to know how to make the coffee machine yourself. All that “concrete” information is abstracted away in the form of the steps needed to brew a cup of coffee with said machine.
ChatGPT is an abstraction layer for all of human information. It’s not the first abstraction layer for all of human information — that honor goes to Google search. But ChatGPT makes Google Search feel like growing your own coffee beans.
With Google search, you first need to know how to search, then you have to suss through all the results to try and find the answer you want. It’s a specialized skillset.
With ChatGPT, all that vanishes. You can ask it general or highly-specific questions, in plain English, and get useful results, in plain English. If you need more info, you just ask — it remembers the conversation and can modify its answer. It’s the ultimate abstraction layer.
It’s not perfect. It’s still not as good as experts. For these people, working with ChatGPT in their field of expertise won’t have much of an impact. But most people aren’t experts, and should see a tangible productivity boost when partnering with ChatGPT.
There are still a lot of open questions about where this technology will take us over the next decade. I think we’ll start to see this type of technology permeate everywhere in our lives, in the same way the Internet became ubiquitous.
ChatGPT works best when thought of as a collaboration partner that is always motivated, ready-to-help, and contains a universe of information.
Here are my predictions for the future with AI:
- A few people will lose jobs, but that happens with any technology. How many travel agents still exist?
- The gap between people who have access to this technology, and people who don’t, will be extreme
- The “haves” will each have their own version of ChatGPT, that’s trained specifically to them
- Privacy is going to be more of a public talking point than ever
- In ten years, the movie Her with Joaquin Phoenix will feel like a quaint period piece
An atomic bomb just exploded, and the world will never be the same.
- Perhaps It Is A Bad Thing That The World’s Leading AI Companies Cannot Control Their AIs By Scott Alexander
- OpenAI CEO Sam Altman | AI for the Next Era Interview (YouTube)
- Superintelligence: Paths, Dangers, Strategies by Nick Bostrom (I’m reading this now)