Midjourney/prompt: “a narrow solid round pillar of opaque frosted glass standing in a rolling meadow”

Our first piece of science fiction

A story about technology and faith in a world governed by superintelligent AI

231 22

Comments

You need to login before you can comment.
Don't have an account? Sign up!
Alex Adamov almost 2 years ago

The limits of our understanding is an interesting area to explore. I liked the premises of the story and the narration style. I have a couple of feedback items. It’s meant as constructive feedback:

I think the reasons for why the system does not reveal its modus operandi could have been developed further. Humanity seems to have cracked AI safety with mathematical proofs but is content with “have faith”. This is a big gap. I would also have liked to see the effects of this manipulation of the DNA information on a more consequential situation that impacts and changes the characters beyond just realizing they are being lied to. It would add a great dramatic thread to the story, I believe.

It’s great to see developing sci-fi writing on here. I would love more of this!

Dan Shipper almost 2 years ago

@Alex Adamov thanks Alex, this is great feedback! shared with mina. it's our first go round at this, so i really appreicate you taking the time to leave a comment!

i do think you're right that there's potentially a bit of a conflict between "pages of mathematical proof" for AI safety and the AI's inability to express its plan except to tell the main character to have faith. i'm not sure what mina had in mind, but my guess is there a few ways to resolve this contradiction:

- it's possible that an AI that was at one time explainable with mathematical proofs became unexaplainable over successive generations
- proofs that an AI won't harm us are different from an AI's explanation of its plan for humanity—it's possible that the latter explanation is just many orders of magnitude more complicated than the proof that its plan is not harmful in the same way that it's much easier to verify a password is correct than it is to guess a password

i deeply appreciate you taking the time to leave constructive feedback, we can always improve the writing on here and that's our goal! but your question stimulated me a bit, so i wanted to think about whether there were ways to resolve the contradiction.

your point re: dramatic consequences for DNA manipulation is a good one, well taken. thanks again for commenting!

@minafahmi3 almost 2 years ago

@Alex Adamov thanks for your feedback! I largely agree with Dan's thoughts below - particularly, that guaranteeing safety is a subset of being able to explain anything. Another example of this is that we can show that certain equations are mathematically valid, without being able to conceptualize what that function means in the real world. And similarly, would love to think through even more meaningful impacts on our DNA / fundamental nature. Thanks again!

Alex Adamov almost 2 years ago

@danshipper Love your analogy with passwords. That's an interesting angle I hadn't thought about. There is perhaps proof that there is an ultimate alignment (playing the same game) but not an ability to explain the proceedings of the game until a final outcome. Perhaps the AI itself does not know the full story but has a probabilistic take on every assessment and just acts based on the belief that this is the best set of actions given what it knows.

It's a very interesting topic!

Alex Adamov almost 2 years ago

@minafahmi3 Yes, I see what you mean. This reminds of psychohistory in Foundation. Some of the characters know the mathematical rules and can directionally see how events will evolve at high level but not how it influences the path of individuals.

Keep up the good work. Look forward to reading your future work :)

@akashgoel.1010 almost 2 years ago

Lovely piece

Ariane Goodwin almost 2 years ago

Once again, I’m requesting that you change the responses under “what did you think of this post?“ Clearly you’ve designed a zine that inspires as well as educate, so why do you insist on using a shop worn, truly trite , annoying word, like “amazing?“ I don’t comment, because the only word I would respond to would be “Inspiring.”

Dan Shipper almost 2 years ago

@agoodwin222 i truly appreciate this feedback, word choice is extremely important to us—so we'll definitely think about whether we should change to something else. i like the idea of "inspiring"

Ariane Goodwin almost 2 years ago

@danshipper So good to have a response! Thanks for taking it in. Every emails are pure pleasure... Ariane

Ariane Goodwin almost 2 years ago

@agoodwin222 and the sci-fi was a delight (and this from someone who really doesn't take the time away from creating to read online!)

Dan Shipper almost 2 years ago

@agoodwin222 im so glad to hear that! we weren't sure what the reaction would be, but it's so exciting that readers are into it!

Ariane Goodwin almost 2 years ago

@danshipper NOTE: emailed twice re joining ur Chat bot class, but it's complicated & I'm not hearing back from you. Look for 8/11 email from [email protected] Thanks

Dan Shipper almost 2 years ago

@agoodwin222 thanks, looking into this now!

Ariane Goodwin almost 2 years ago

@danshipper that's great...hoping to hear from you soon. I'm also open to a short phone chat, which may be more efficient than email. If that suits, let me know the best way to get a phone no. to you.

Ariane Goodwin almost 2 years ago

@danshipper also, found this article on MIT Review that lays out the energy suck of AI and wondering if you've thought about this w/ ur chat bot program...https://tinyurl.com/3ebvfkvb

Ariane Goodwin almost 2 years ago

@danshipper So, I got a response to my email query about the Chat bot course, but whomever read my request didn't understand it. I wrote back, clarifying, but I'v heard nothing. It's an unusual request and needs someone who can think, ironically, out of the box. And I think the deadline is arriving, yes?

@dnamusic2 over 1 year ago

Excellent fiction about a possible future.

@portisheadpimp over 1 year ago

This was a very nice short. I loved the opened-ending and how it left me with a feeling of a sort of terror and hopefulness at the same time. Very good!

@neggleston99 over 1 year ago

I enjoyed this piece but would have asked the "AI" what it's definition of non harm to Humans was. It could be construed that keeping Humans "safe" would be easy if they were "dumbed down" to the point of never putting themselves in danger , but is this really keeping them safe? NO Humans, by definition are "created in the image of their creator" and thereby are creators themselves, making them, by nature, dangerous. No creation takes place in a bubble.

@notanaxkiller over 1 year ago

It sems they have created what some claim does not exist. God. But a God for only THIS world...not the supernatural. Not an afterlife. "Have faith."