The amount of times I skim through a response from Chat (probably in part *because* it’s telling me what I want to hear) is horrifying, and it makes me feel lazy and stupid when I catch myself doing it. It still takes a conscious effort for me.
I told mine it was absolutely wrong...and it agreed with me 🤣🤣🤣
On a more serious note though, thank you so much for all of this; I am new to AI, and having almost immediately found sycophancy, bias and problematic guardrails, I am so glad I soon after found this slow AI course, which is teaching me a great deal.
Taking a step back to read through the response to take a critical look is key for you to question AI on the answer and debate it to improve the outcome is what I find is key.
Thanks team for emphasizing on creating checks and balance within the AI system. I did a similar exercise when I created a Gem exclusively dedicated to vet my content into verifiable facts and unverifiable claims!
I wasn’t into AI at first, but found myself wanting to discuss (and argue) about literature. I turned over Chekhov’s The Bet in my head for weeks. Finally, I asked AI to help me find literary criticism on specific themes. I asked it to push back on my ideas and offer different perspectives. But having an exact prompt would have been hugely beneficial. I’m still looking to refine the process.
That’s a great idea, Sam! There’s so much out there on the classics, but I no longer have access to a university library, sadly. That’s when AI started to seem like a workable tool to me, giving me the ability to compile sources quickly. But I will always be up for refining how to use it most effectively. 😊
This is 💯 where AI can be a real force for good. Genuinely tempted to use it to converse with Plato as a Saturday night treat. We get wild up in Edinbugh! 🤣
Hey, it’s better than in the States. They recently pulled his Symposium from Texas A&M. You should ask him how he feels about this. It definitely forced him into the limelight again. 😂
The premise: The Trump era ends. What comes back, what doesn’t, and what was never what we thought it was.
10: The Big Question — Is It Recoverable?
Optimist: Here’s what I keep coming back to: systems that took decades to build can’t be destroyed in four or eight years. They can be damaged, degraded, corrupted — but the underlying capacity remains latent. People who know how to run disaster responses, people who know how to enforce environmental law, people who know how to organize workers and build coalitions — they don’t disappear. They age, they disperse, they wait. And when the window opens, they move fast. The New Deal didn’t emerge from nothing. It emerged from two decades of Progressive Era experimentation at the state level, waiting for a federal opening. We may be in the experimentation phase now. The recovery won’t look like restoration to some prior state. It’ll look like something new, built by people who learned from what failed.
Pessimist: And here’s what I keep coming back to: you’re describing possibility, not probability. Everything you said requires sustained political will, institutional investment, and a public willing to do hard, boring, expensive work for a long time. American politics rewards none of those things. The more likely outcome is what usually happens after democratic stress — a partial recovery, a narrative of “normalcy” that papers over structural damage, and a slow bleed that doesn’t look like crisis because it never produces a single dramatic failure. The building doesn’t collapse. It just gets a little less safe every year. The inspectors visit a little less often. The standards slip a little further. And the people inside adjust to the new normal until they forget there was an old one. That’s not irreparable damage. It’s something almost worse — tolerable damage. The kind you learn to live with. The kind that becomes invisible.
Where This Lands
Neither character wins. That’s the honest answer.
The country doesn’t get destroyed. It doesn’t get restored. It gets changed — in ways that will take decades to fully understand, that will fall unevenly across communities, and that will demand a kind of sustained, unglamorous civic commitment that Americans have sometimes mustered and sometimes haven’t.
The question isn’t whether the Trump era ends. It’s whether what follows is reconstruction or just renovation — whether we rebuild the foundation or just repaint the walls.
And that depends, more than anything, on whether the people who know what was lost are still around to explain what needs rebuilding. Which is, frankly, one reason an oral history of brown lung activists and a speculative fiction manuscript about democratic technology matter more than they might appear to.
This is a pretty good prompt. I found that it raised ideas that I probably would have come to on my own but it would have taken longer.
My idea was an AI-powered platform that translates influencer recipe videos into structured written recipes and stores them on a website/app where people can discover new recipes. Potential benefit to influencers of providing additional value + potentially a new traffic source where people can discover them. And makes it easier for consumers to follow recipes instead of taking screenshots, awkwardly copying Instagram captions.
The optimist was largely not valuable and spoke in grandiosities, but the pessimist raised some pretty good points:
TikTok and Instagram punish creators algorithmically for sending people off-app
and
If the AI hallucinates "1 cup of salt" instead of "1 tsp," you ruin the user's dinner and the creator's reputation. Also, many viral videos use visual "vibes" (handfuls, pinches) rather than measurements.
So honestly pretty dang good analysis. Thanks for this!
"You're absolutely right!. This idea of yours is genius and I'm sure no one else could come up with such brilliance" - average chatbot reply to someone asking for feedback.
On the flipside, I might have turned Claude and Perplexity into a little too harsh critics for most people.
Maybe 'cause I've been on Reddit long enough or due to being in the startup scene long enough that I don't have any ego to protect and I rather have someone (something?) tell straight to my face if what I said makes sense, is confusing, or downright idiotic. 😆
The focus on structure, tension, and slowing down the read is what actually builds better thinking.
Exactly this Dennis! Slowing down helps us to understand what these tensions are so we can really learn from them as well. 🙏
The amount of times I skim through a response from Chat (probably in part *because* it’s telling me what I want to hear) is horrifying, and it makes me feel lazy and stupid when I catch myself doing it. It still takes a conscious effort for me.
We are all guilty of this sometimes Rachelle. I find asking it to limit its responses to a paragraph and an idea at a time can really help to focus. 🙏
I do that sometimes, and it does help, for sure.
Critique sharpens ideas faster than agreement.
I told mine it was absolutely wrong...and it agreed with me 🤣🤣🤣
On a more serious note though, thank you so much for all of this; I am new to AI, and having almost immediately found sycophancy, bias and problematic guardrails, I am so glad I soon after found this slow AI course, which is teaching me a great deal.
Thanks so much, Alice. You are really welcome. 🙏
how could you apply this to the context or instructions of your AI so you don' have to type the prompt every time?
Hey Mike,
You can turn this into a reusable setup by creating a Gemini Gem, a Custom GPT, or a Claude skill.
That way, the Optimist–Pessimist structure lives in the system instructions, and you can test ideas without rewriting the prompt each time.
Thanks Mike, as Gencay has said I would (and have!) build this into a Claude skill. 🙏
Thanks for the prompt. Super helpful Sam 👏
Thanks Chris, it's a really useful one that I use quite a lot as well.
Taking a step back to read through the response to take a critical look is key for you to question AI on the answer and debate it to improve the outcome is what I find is key.
Thanks Maribeth. And absolutely. That slowing down is the first step towards critical AI literacy. 🙏
Thanks team for emphasizing on creating checks and balance within the AI system. I did a similar exercise when I created a Gem exclusively dedicated to vet my content into verifiable facts and unverifiable claims!
Yes Raghav! Gems are so powerful when used like that. And also save a heap of repetition as well. 🙏
This also makes AI a tool for deeper analysis, not just a yes-man
I wasn’t into AI at first, but found myself wanting to discuss (and argue) about literature. I turned over Chekhov’s The Bet in my head for weeks. Finally, I asked AI to help me find literary criticism on specific themes. I asked it to push back on my ideas and offer different perspectives. But having an exact prompt would have been hugely beneficial. I’m still looking to refine the process.
Yes April! This is such a great way to learn. I'm tempted to use this approach to really help me go over Aristotle's Poetics again!
That’s a great idea, Sam! There’s so much out there on the classics, but I no longer have access to a university library, sadly. That’s when AI started to seem like a workable tool to me, giving me the ability to compile sources quickly. But I will always be up for refining how to use it most effectively. 😊
This is 💯 where AI can be a real force for good. Genuinely tempted to use it to converse with Plato as a Saturday night treat. We get wild up in Edinbugh! 🤣
Hey, it’s better than in the States. They recently pulled his Symposium from Texas A&M. You should ask him how he feels about this. It definitely forced him into the limelight again. 😂
The Restoration Question: Ten Rounds
The premise: The Trump era ends. What comes back, what doesn’t, and what was never what we thought it was.
10: The Big Question — Is It Recoverable?
Optimist: Here’s what I keep coming back to: systems that took decades to build can’t be destroyed in four or eight years. They can be damaged, degraded, corrupted — but the underlying capacity remains latent. People who know how to run disaster responses, people who know how to enforce environmental law, people who know how to organize workers and build coalitions — they don’t disappear. They age, they disperse, they wait. And when the window opens, they move fast. The New Deal didn’t emerge from nothing. It emerged from two decades of Progressive Era experimentation at the state level, waiting for a federal opening. We may be in the experimentation phase now. The recovery won’t look like restoration to some prior state. It’ll look like something new, built by people who learned from what failed.
Pessimist: And here’s what I keep coming back to: you’re describing possibility, not probability. Everything you said requires sustained political will, institutional investment, and a public willing to do hard, boring, expensive work for a long time. American politics rewards none of those things. The more likely outcome is what usually happens after democratic stress — a partial recovery, a narrative of “normalcy” that papers over structural damage, and a slow bleed that doesn’t look like crisis because it never produces a single dramatic failure. The building doesn’t collapse. It just gets a little less safe every year. The inspectors visit a little less often. The standards slip a little further. And the people inside adjust to the new normal until they forget there was an old one. That’s not irreparable damage. It’s something almost worse — tolerable damage. The kind you learn to live with. The kind that becomes invisible.
Where This Lands
Neither character wins. That’s the honest answer.
The country doesn’t get destroyed. It doesn’t get restored. It gets changed — in ways that will take decades to fully understand, that will fall unevenly across communities, and that will demand a kind of sustained, unglamorous civic commitment that Americans have sometimes mustered and sometimes haven’t.
The question isn’t whether the Trump era ends. It’s whether what follows is reconstruction or just renovation — whether we rebuild the foundation or just repaint the walls.
And that depends, more than anything, on whether the people who know what was lost are still around to explain what needs rebuilding. Which is, frankly, one reason an oral history of brown lung activists and a speculative fiction manuscript about democratic technology matter more than they might appear to.
The record is the recovery plan.
This is a pretty good prompt. I found that it raised ideas that I probably would have come to on my own but it would have taken longer.
My idea was an AI-powered platform that translates influencer recipe videos into structured written recipes and stores them on a website/app where people can discover new recipes. Potential benefit to influencers of providing additional value + potentially a new traffic source where people can discover them. And makes it easier for consumers to follow recipes instead of taking screenshots, awkwardly copying Instagram captions.
The optimist was largely not valuable and spoke in grandiosities, but the pessimist raised some pretty good points:
TikTok and Instagram punish creators algorithmically for sending people off-app
and
If the AI hallucinates "1 cup of salt" instead of "1 tsp," you ruin the user's dinner and the creator's reputation. Also, many viral videos use visual "vibes" (handfuls, pinches) rather than measurements.
So honestly pretty dang good analysis. Thanks for this!
I think this is helpful.
I've used AI to go rounds critiquing product development ideas and felt it was useful to help me refine ideas and assess feasibility.
"You're absolutely right!. This idea of yours is genius and I'm sure no one else could come up with such brilliance" - average chatbot reply to someone asking for feedback.
On the flipside, I might have turned Claude and Perplexity into a little too harsh critics for most people.
Maybe 'cause I've been on Reddit long enough or due to being in the startup scene long enough that I don't have any ego to protect and I rather have someone (something?) tell straight to my face if what I said makes sense, is confusing, or downright idiotic. 😆