Simulating history with ChatGPT ⊗ We should all be Luddites now ⊗ Taking business personally

No.279 — AI NPC teammates ⊗ Adventures in narrative time ⊗ Hopeful futures in science and storytelling ⊗ AI Index Report 2023

Simulating history with ChatGPT ⊗ We should all be Luddites now ⊗ Taking business personally
3D reconstruction of Tenochtitlan, the capital of the Aztec Empire.

Simulating history with ChatGPT

Last week we had an article about how AI is helping historians better understand our past, this week we have a history teacher, Benjamin Breen, who uses ChatGPT to create simulations of medieval situations and uses the results and interactions to teach his students.

The author might have rose tinted glasses when he says that the humanities will have a resurgence because LLMs are “inherently textual.” It is likely just a temporary state and even if it isn’t, it’s easy to imagine that being good at writing prompts will only be a valuable for a short while, as the AIs’ understanding of how different people express themselves will evolve.

That being said, although I believe his “textual” claim is misguided, it might be the case that humanities education makes for a good toolset to create better prompts, critique replies, adjust, and validate that the results are accurate and not hallucinations. I’m undecided, but it’s definitely an intriguing proposition to keep an eye on.

I’m drawing more attention to that humanities aspect of his piece because it might be more broadly relevant to which skillset is useful in using AI, but you should also have a read for his fantastic examples of how he asks for simulations and how he uses this tool to enhance classes and understanding, instead of just worrying about cheating and trying to block it. A growing number of people have been referring to SALAMIs as an endless number of interns. Breen basically asks each each student to create their own teaching assistant, and uses the imperfections of the assistants as features of the teaching. Love it!

I am under no illusions that these simulations are accurate: they are littered with confidently-stated falsehoods and hallucinations. Sometimes, though, hallucinations can be a feature, not a bug. […]
Student engagement in the spring quarter, when I began these trials, was unlike anything I’ve seen. The first time I tested the idea out informally (asking students to simulate their home town via an up-ended general purpose simulation prompt) I realized that we had gone 5 minutes past the end of class without anyone noticing! […]
The simulation idea, which I called “History Lens” because it provides a distorted perspective on the past.

We should all be Luddites now

I’ve shared a lot of articles about Luddites and will share a lot more I’m sure. Here Dave Karpf effusively reviews Bran Merchant’s Blood in the Machine: The Origins of the Rebellion against Big Tech, and it encapsulates exactly why I’m always coming back to this topic.

“[The] Luddite rebellion was, first and foremost, about labour power. The Luddites were not reflexively anti-technology. They were skilled artisans who had a history of incorporating new technologies into their profession. The specific technology they opposed (the power loom) was poised to wreck their industry and replace them with factories filled with child laborers, who would flood the market with lower-priced, lower-quality goods.”

Ring a bell? People asking the though questions about technology repeatedly get called Luddites, it’s meant as an insult, it should be a compliment and reminder of a great lesson in history. One that ultimately failed but still showed that technology yielded by capitalism with no restraints can lead us astray. Slowing things down doesn’t mean blocking progress, it means insuring more thoughtfulness and fairness.

Even when the word itself is not on our lips, the ethos it embodies, breathed into us from our earliest engagements with technology, always seems to be there: that to oppose technology is to oppose the future, and prosperity; that questioning technology can only point us backward. […]
The Luddites, in other words, were struggling with many of the same questions that we face today. Who should reap the rewards of technological innovations? Whose interests should government represent, promote, and protect? What sort of society do we want to build?

Taking business personally

Over the last few months and years, a lot has been written about dreams of a more decentralized web, more open, less strident, less surveilled, not as controlled by big tech, etc. A lot of is theoretical and it sometimes feels like real examples are few and far between or at least at very small scale. Are.na is at relatively small scale but it’s real, it’s been going for 12 years, and it’s everything the dreams above want to see. Meghna Rao interviews Charles Broskoski, the co-founder and CEO of Are.na who is refreshingly down to earth and very much on mission. That mission is what we need more of around the internet.

This is what Are.na believes in: Superfast is not the only pace for growth. It is possible to sustain and grow a digital business without throwing venture capital (VC) money at it. You can still build something lasting on the internet that looks the way you want it to. […]
Having a strong point of view about this type of thing is a meaningful differentiator. A lot of other platforms are diluted; their ethics do not translate to their product choices.

More →These two article are what (re)prompted me to want to be more fluid in formatting, they are so idiosyncratically their authors’ that I don’t have a summary or much of an opinion for now, they are just articles I want to share because it’s fun to read them and to go through their thinking and favourite topics. Matt Webb built his first AI NPC teammates and he shares what he learned while Venkatesh Rao goes over some of his adventures in narrative time.

Asides

Your Futures Thinking Observatory