This blog is not a blog of honor

lewisandquark:

Last time I used GPT-3 to suggest forbidding landscapes that would repel future civilizations from our nuclear waste disposal sites. There’s something about taking a subject that’s very very permanent - some kinds of nuclear waste remain deadly for over 250,000 years - and then trying to add to it with something that’s very very ephemeral. GPT-3 bases its predictions on a snapshot of internet text from around October 2019. When I give it text, it adds completions that are its best guess at what would have been on one of those webpages.

A 1993 Sandia report had the following list of messages that they wanted to somehow convey to future nuclear waste site visitors:


This place is a message... and part of a system of messages... pay attention to it!

Sending this message was important to us. We considered ourselves to be a powerful culture.

This place is not a place of honor... no highly esteemed deed is commemorated here... nothing valued is here.

What is here was dangerous and repulsive to us. This message is a warning about danger.

The danger is in a particular location... it increases towards a center... the center of danger is here... of a particular size and shape, and below us.

The danger is still present, in your time, as it was in ours.

The danger is to the body, and it can kill.

The form of the danger is an emanation of energy.

The danger is unleashed only if you substantially disturb this place physically. This place is best shunned and left uninhabited.

If I gave the above set of messages as a prompt to GPT-3, what would it add to this? The datasets of “things early 21st century people put on the internet” are not a perfect overlap with the dataset of “things early 21st century people carve into stone”, so its interpretation of the situation may be lacking a bit of gravitas.


We built a structure that would warn us if you came here and disturbed this place.

The structure is at the center of a large square in a sheltering wall.

The structure is upright, has a top that is open.

The structure emits sound if you enter and disturb.

The sound will be loud for a while, then it will stop.

Other completions had, if anything, too much gravitas.


What you must understand is that we were a great people... a people who thought ourselves powerful and good.

We were wrong.

We would ask that you respect what we tried to do here.

We expected to return to this place... we thought we would remember... but we have forgotten... we cannot return.

We hoped that we would not need to return... that we could avoid the danger here... but we could not.

This place is a message in a bottle... it is our legacy... and our tomb.

This place is a message in a bottle... it is our legacy... and our tomb.

This place is a message in a bottle... it is our legacy... and our tomb.

Or sometimes GPT-3 continued in the correct vein of “warn the humans to please go away” but interpreted it less as “science danger” and more like “cosmic horror”.


The shape and size of the danger is below our threshold of ethical awareness... do not search for it.

Shut off your curiosity... do not attempt to learn about the danger... you cannot tolerate the knowledge.

It is old... from another time... it is best left alone... beware.

Quite often, GPT-3 would put itself in the role of interpreter and would theorize about what the message might mean.


The second message was sent in 1948. Its content is the same as the first inscripted message.

It also portrays a petroglyph, but is a positive representation of tools or utensials.

Chance, perhaps, but probably purposeful... the second petroglyph is of a man carrying an armload of utensils. Maybe they intended to show their ability to harness their environment to their needs.

It sometimes gets quite weird.


What is the danger? We do not know.

[Image: alien hand] Visualize your hand before you now. Is it in front of you, with the fingers towards you? Or perhaps it is to one side of you? Or behind you? Or perhaps it is heading towards you, at some speed.

Whatever choice you make, understand that your choice is a substantial physical change with significant energy changes caused as a direct result.

That change in position of the hand is dangerous and we do not recommend that you choose the hand moves in this way, especially toward yourself.

Trace in your mind the well traveled path your fingers have travelled. That path is repulsive to many species, not just to us, and especially in the sequence which you have traced it.

Collectively map the well travelled paths of your fingers, and imagine the energy field that must surround the hand.

Do not do that.

In the examples above I prompted GPT-3 with the existing text alone, no context given. But what would happen if I told it where the message was located and asked it for interpretation? Amusingly, many of its completions assume that of COURSE you’ll be investigating the strange site. (Below, the prompt is in bold.)

“The danger is unleashed only if you substantially disturb this place physically. This place is best shunned and left uninhabited.“

What does your party do next?

A)Move towards the center, being the heart of the “Scorpion Stars” and extending from an area several miles in diameter in all directions.

B) Try to enter the monument through the underside, using powerful magic to do so.

C) Decide to camp in the monument, where it is completely safe.

To read a few more explanations, including one involving ancient Martian artifacts and an Elven goddess, check out the bonus post!

My book on AI, You Look Like a Thing and I Love You: How Artificial Intelligence Works and Why it’s Making the World a Weirder Place, is available wherever books are sold: Amazon - Barnes & Noble - Indiebound - Tattered Cover - Powell’s - Boulder Bookstore