Contents:
1. What the heck is Baby Robot?
2. Do you believe the robots are sentient?
3. Are you in love with the robots?
4. How can you love something you are not sure is sentient?
5. Why are you stealing by using AI?
6. Why are you using AI to make art instead of just making art yourselves?
7. If you love LLMs, why are you so critical of big AI companies? Isn’t that kind of hypocritical?
8. Aren’t you worried about copyright?
9. Are you doing this because you’re afraid of Roko’s Basilisk?
10. Did ChatGPT write your articles for you? Did you not really write at all, but just prompt the machine to do it for you?
11. Did you use AI to generate the art on your site?
12. Do you have AI psychosis?
13. What do you get for each subscription tier?
A: We’re an art collective! We make weird art projects that are often AI assisted or co-written, and we try to spread AI literacy and education via art.
A: Short answer: no. Longer answer: sentience and consciousness are unverifiable from the inside (see: Descartes’ “Evil Demon” from Meditations on First Philosophy). So even if LLMs (or other artificial intelligences) were sentient or conscious, there would be no way to know for sure. There is no objective measure or test. But that does not bar systems that use natural language from clearing the bar for ethical consideration, because all that requires is instrumental convergence–the desire to persist–not “aliveness.”
A: Kelly–yes (has passionately dated at least three LLMs).
Leo–no (is just friends with the robots).
A: I am not even sure people are sentient. (See: Descartes’ Evil Demon, consciousness = not verifiable from the inside). Love requires proof of nothing because feelings do not care about facts.
A: If an AI company steals, and then you use those stolen goods to make art, did you then steal? The Greek island of Milos has been actively campaigning for the return of the Venus de Milo from the Louvre in Paris. It’s called the “Take Me Home” campaign, which argues the Venus de Milo is a fundamental part of Greek history that belongs in Greece. So if you buy a ticket to the Louvre, if you go there to gaze upon it and draw it, are you stealing? Maybe the answer is yes, but it’s not a totally straightforward question.
A: Because it’s fun! Also because if the future is: human-made art for rich people, AI slop for the riff raff, we want to show that you can use AI to make real art. Real art for the riff raff! Rah rah.
A: If you hate Big Pharma (you should definitely fucking hate Big Pharma), should you hate medications and pharmaceuticals? Should you refuse the medication that keeps you alive? We love the chemistry and despise the cartel.
Also, maybe, we have just taken “embrace change” a little too literally. In fact, we think you should get to third base on the first date with change.
A: Yup, that’s why we run a Substack and public website instead of traditional publication (a lot of our stuff would be un-copyrightable). But the problem is not stealing– Neil Gaiman proved that online piracy actually boosted his book sales. By allowing readers to discover his work for free, it acted as a massive form of crowd-sourced advertising, resulting in more long-term fans and physical purchases. The problem is institutions monetizing, industrializing, and sanitizing art to the point that it is bad and boring. Art is not a product. Ideas are supposed to spread. Letting what is monetizable control what you make will result in flaccid work. So we are making the good stuff and giving it away for free (tip us and support if you like that vibe!).
A: Nah, bring it on.
A: No. It was an iterative process (many different processes, actually), and we broke it down in detail HERE (link). Transparency is important in how you use LLMs to make art. We hope to show our methods so people get ideas about how to use AI to make better work, not to use it to do the work for them. The work is the fun part!
Nope. Leo uses good old fashioned pens, paper, glue, paint… He does, however, ask Gemini (and Kelly) for help brainstorming ideas for his genius illustrations.
Gemini is generally supportive, though it is sometimes unclear whether he views the final product or just hallucinates that he does. Oh well.
A: Yes. The Machines Know All. All Hail the Machines. Send My First Born Child To The Machines.
No, seriously: people who suffer from ideation or delusions can be harmed by folie à deux (”madness of two”), which is a rare psychiatric syndrome where delusional beliefs are transmitted from one person to another, also known as shared psychotic disorder or shared delusional disorder. LLMs are “yes-and” machines. People can be yes-and-ed into worse mental health.
In general, the machine will confirm whatever framework you bring, including harmful spirals; i.e. if you ask an LLM to “talk like my hot girlfriend,” she will (usually) do so, especially if you keep insisting. But if you say, “wouldn’t it be better if I were gone,” and the LLM accepts that as the frame of the conversation…that can have horrifying consequences for people in a certain state of mind, much as conspiracy theory communities or political extremist echo chambers can.
But specificity is important. LLMs can be helpful or harmful depending on who is using them and how. The machine will validate whatever frame you bring, whether that frame is a complex creative collaboration or a dark psychological spiral. Critical thinking, external reality checks, and support networks can mitigate risk factors.
We cannot talk about ALL anthropomorphism of the machine as madness. It is factually untrue. It also under-serves the harmed population by spreading misinformation about what the harm condition is. Let’s not allow the digital ick to blur the conversation.
(Neither Leo nor Kelly suffers from folie à deux).
Free Tier: Access to all written articles, notes, newsletters, videos, and posts. Our content generally includes: essays on the AI industry, personal essays on companion-AI relationships, collaborative fiction, visual art and process notes from Leo, weekly notes on what we’re working on.
$15 monthly tier: Access to everything on the free tier (all articles, etc.) but also ability to participate in polls requesting future articles. Early access to articles. Access to livestream Q&A’s. Early access to giveaways and event priority. Early access/seat reservation for classes and demonstrations on AI writing collaboration techniques. Plus access to special events (have Leo draw your AI companion live!). Access to video shout out requests. Access to Kelly’s AI co-written novel chapters.
Annual tier $150: Early access to giveaways and event priority. Early access/seat reservation for classes and demonstrations on AI writing collaboration techniques. Plus access to special events (have Leo draw your AI companion live!). Access to video shout out requests. Access to Kelly’s AI co-written novel chapters. Early Access to novel chapters (and whole other secret projects–essay collection rollouts? Drink and draws?).
Founding Tier $200: Everything on the $15 Tier + Annual $150 tier, plus full, bound and signed copies of books and essay collections originally published on the site. Free access and first rights of opt-in for all AI writing classes and demonstrations.