Exploring ideas with GPT-3

While it's not perfect, I've been struck by the connections GPT-3 draws between words and ideas in almost every domain. It's a new way to have a conversation of sorts with a computer, and get a peek into how an AI understands patterns in the world.

When I asked GPT-3 (Q&A preset) to explain 'unorthodoxed design' here're some of its responses:

Design is fundamentally about creating ideas that innovate in the spirit of wild dreams.

It's usually smarter to start with the obvious idea and then work your way around to something new and uncommon.

I highly recommend trying out the @openai playground, it's free, publicly available now, and easy to play with. You can get free access at: openai.com/api

I ran the top half of my essay, Humans will always do more than survive, through GPT-3 and tried to get it to write me an alternate conclusion; here's what I got.

We have a tendency to overcomplicate things, but the truth is that dance, music, art and theater are all ways for us to escape from reality, and enjoy ourselves in the process. I hope that people never become so advanced or obsessed with their advancements that they forget why they got here in the first place...so we can all go out and dance until the sun comes up.

— 🤖 GPT-3

One other paragraph in this newsletter (outside this section) was written by GPT-3, can you guess which?

Why you should stop using the word 'interesting'

It's a diplomatic word, used when there's no other word available. A substitute for thinking. It means nothing. Because the word 'interesting' doesn't say anything about what's interesting. It's the verbal equivalent to a shoulder shrug, and it usually means either that a speaker doesn't have anything substantive to say, or that they just aren't saying anything at all. Why are you 'interested' in something? Does it mean that you want to know more? Or does it mean that you already know what's going on, but you think it might be fun to look at anyway?

In place of the word 'interesting', say why you think it is interesting instead.

Speaking of Futureland

I want to share the various tools, mental models and ideas I come across that are valuable in some way. Sometimes these could get a little nerdy, but I mean well.

Futureland has been one of the most delightful pieces of software & community combinations I've come across this year. It's quickly become one of my favourite apps. Try for yourself at futureland.tv.

From the archive

An excerpt from Polymath Letters - 05/07/2020

Even when the human is about to die, his illusions, mind, hopes, and opinions don't die. One patient was afraid of the pain to come, and one patient was happy remembering his good old times -- both are entangled in the games of the mind. Even in the last moments of our life, the mind doesn't let us just live; it keeps us busy in its games. The past and the future, both are not real; they exist only inside your head. The only moment that truly exists is the one you are living right now. In remembering the past and imagining the future, the present moment passes by.

Reading: Good to Great – A management book on transitioning from good to great companies. 1/ Finding your Hedgehog concept and why great companies have one 2/ Put the best people on the biggest opportunities, not the biggest problems. 3/ Confront brutal facts head-on.

Watching: How Airlines Quietly Became Banks — each airline runs its own loyalty program as a separate company, and these programs consistently bring in billions more than the airlines themselves. Also, The Art of Code, a talk by Dylan Beattie: there's the code that makes our banks and hospitals, as well as the code we use to browse the web and create spreadsheets. And then there's code that just exists because somebody wanted to write it. Maybe it solves a problem. Or maybe it doesn't do anything at all, some software comes into being just because somebody wanted to write it and make their friends smile, laugh, or dance.