Sunday Scaries (7/19/20)
Hey everyone,
I spent this week perched on top of a windswept mountain near Pray, Montana. This is the part of the country where paved roads are an exception and the animals have fences to keep the humans out.
My home here is a single-story cabin designed to make the most of the evening’s dying light. Sunsets here are an intellectual affair. Long after the sun has dipped below the horizon, light continues to paint the peaks of Dexter Mountain above my porch. The result is a long period of twilight which seems to stretch for hours.
I’ve been spending my time working and marveling at the results of GPT-3, a form of artificial intelligence that can respond to requests in conversational English. This is the first time that I feel like I’m interacting with real intelligence, instead of a machine.
Something big is happening right in front of us.
This week’s subscribers only posts:
Programming 3.0 discusses an opportunity for the next wave of software, built on top of general-purpose AI models like GPT-3.
Why Should Name Our Models is a perspective on why giving advanced models simple names like Bert or Pete can help us improve our dialogue around AI.
What’s New This Week
GPT-3 Fiction Newsletter. I’ve been collecting interesting examples of things that people are doing with machine learning in a newsletter called “GPT Stories”. Since I was granted access to the beta version of GPT it’s honestly hard to concentrate on anything else. This post contains a poem about Elon Musk in the form of a Dr. Seuss rhyme, a job cover letter written for Proctor & Gamble, answers to medical questions, and more.
More GPT Stuff 👇
Here is a conversation I had earlier today with the machine. Everything in bold is from me, the rest is computer generated:
If you want to talk with GPT yourself, check out LearnFromAnyone.
(please note, this website is currently down for review by the OpenAI team)
Photo of the Week
Photo: Cabin near Pray, MT
In This Week’s Edition:
A new theory for the Fermi Paradox (a.k.a. why we haven’t found aliens).
The resignation letter of Bari Weiss (formerly of the NYT) describing what she calls “self-censorship” and “orthodoxy” from the paper. Worth reading.
Why we keep setting moving targets for artificial general intelligence and reasons to start thinking about it now.
A New Theory on Why We Haven’t Found Aliens Yet
The Fermi Paradox is my favorite rabbit hole on the internet. Put simply, it describes the unusual situation we find ourselves in: if there was intelligent life in the universe, we would expect to see some sign of it. Since we have found no sign, there must not be intelligent life.
Is this because life is unlikely (how narcissistic of us!), or because there is something that prevents life from becoming multi-planetary?
This article outlines a new solution to this problem: there is life, but it is hibernating until things cool down a bit.
There is something strange happening inside of The New York Times. Read this letter closely and you’ll hear the voices of other journalist who share Bari’s sentiments, but without the means to go independent.
Twitter is not on the masthead of The New York Times. But Twitter has become its ultimate editor.
For these young writers and editors, there is one consolation. As places like The Times and other once-great journalistic institutions betray their standards and lose sight of their principles, Americans still hunger for news that is accurate, opinions that are vital, and debate that is sincere. I hear from these people every day.
There’s No Fire Alarm for Artificial General Intelligence
Technological changes tend to happen slowly, then all at once. We tend to look for linear changes, but fail to realize that non-linear shifts in the world are just as common.
This essay is worth reading, if only to make you reexamine your own biases. If there is smoke in the room you’re in, but no one is reacting, will you have the independence of thought to yell “fire!”
I’m not so sure we’re there yet.
Some key points:
One: As Stuart Russell observed, if you get radio signals from space and spot a spaceship there with your telescopes and you know the aliens are landing in thirty years, you still start thinking about that today.
Two: History shows that for the general public, and even for scientists not in a key inner circle, and even for scientists in that key circle, it is very often the case that key technological developments still seem decades away, five years before they show up.
Three: Progress is driven by peak knowledge, not average knowledge.
Four: The future uses different tools, and can therefore easily do things that are very hard now, or do with difficulty things that are impossible now.
Thanks for reading,
Phil
Sunday Scaries is a newsletter that answers simple questions with surprising answers. The author of this publication is currently living from his car and traveling across the United States. You can subscribe by clicking the link below. 👇