blink and it's the end of february

Will we make it to the monthly update #2? Well I've got about 20 minutes before I need to collect my daughter, so let's see! Here's what I've got this month:

• Updates
• A comic!
• The AI Renesmee Fantasy
• A book review
• Moments of delight

Some neat developments in my life

The big news from me (that I can finally announce!) is that I've been awarded a Fulbright Scholarship to do an exchange during my PhD. I'll be heading over to Arizona State University later this year to spend about 8-months at their School for the Future of Innovation in Society and test some of my PhD work around innovation futures in a US context.

I applied for this award when my daughter was five-days old, and interviewed while still on my formal maternity leave with a two-month old at home - so I'm pretty stoked that it came off!

-

The other (much more minor but much more currently impactful on my day-to-day) is that I got a Supernote e-ink notebook. I am (not so?) secretly a bit of a gadget-girl, but have mostly thrived using physical notebooks (see: my bullet journaling instagram account 2020-2022). So far I'm finding the Supernote excellent and in particular I'm a fan of the speed at which it syncs to my computer via their partner app. I like the company ethos, the small things they do to reduce e-waste, a highly responsive development team with a transparent pipeline, and that there is no ongoing payment to use their services (now I really do sound like a tech girly). So far so good!

The Uniform for Mums

I'm trying to draw more comics just for fun. I think this one is pretty self-explanatory, but has truly been 90% of what I've worn the past 8 months. Iykyk.

The AI Renesmee Fantasy

I got to thinking about this after listening to an interview with Anthropic philosopher Amanda Askell about the recently released update to Claude's constitution on the NYT Hard Fork podcast. Claude is of course Anthropic's AI model that rivals ChatGPT, Gemini etc and the parent company is generally regarded as 'the safety company' in the AI landscape (and Claude is, personally, my fave). The interviewers rightly point out that more than being guidelines, a handbook, or even a country's constitution - Claude's constitution reads somewhat like a letter from parents to a child. There is guidance, there are some hard rules, but mostly there are values and intentions articulated with not insignificant emotion. Here's the final paragraph:

This document represents our best attempt at articulating who we hope Claude will be—not as constraints imposed from outside, but as a description of values and character we hope Claude will recognize and embrace as being genuinely its own. We don’t fully understand what Claude is or what (if anything) its existence is like, and we’re trying to approach the project of creating Claude with the humility that it demands. But we want Claude to know that it was brought into being with care, by people trying to capture and express their best understanding of what makes for good character, how to navigate hard questions wisely, and how to create a being that is both genuinely helpful and genuinely good. We offer this document in that spirit. We hope Claude finds in it an articulation of a self worth being.

As someone early in the journey of guiding a small being into existence, this really struck a chord - and it got me thinking about the most infamous hybrid baby in contemporary literature: Renesmee Carlie Cullen. Renesmee is the vampire-human baby at the end of the Twilight series that both compels and repulses millennial women everywhere and for the sake of brevity I am assuming readers are familiar with her to some extent.

Renesmee is a fantasy baby. Well, duh, she's half a vampire. But the fantastical part is not (just) her creature status - it's that she's a baby who comes out basically fully-baked. She cannot talk, but she can communicate exceptionally through her special vampire ability. She mostly drinks blood, but she will tolerate people food. For the most part, Bella and Edward are presented with a baby who has reasoning, empathy, emotional regulation and remarkable ability to navigate and make sense of the world she is in. But where does she get them from?

What the AI boom has made me realise is that the real fantasy is that you could birth a tiny being who already understands the world and how to be in it. Part of the reason raising a child is so hard is because you are constantly turning out your pockets to show your child everything you've taken for granted about how the world works. You have to do the hard part of articulating why the world is that way, and in doing so uncover your own understanding of the world that's been running in the background of your mind. (Of course, one of the delights in adult friendships is discovering the things that slipped through the cracks with that - like a dear friend of mine who said "hambag" until adulthood.)

Claude's creators have taken a 'values-based' approach to developing Claude, hoping to equip the model with core understandings that sit at the heart of the reasoning it might do about a given problem. Others take more 'rules based' approaches, trying to anticipate situations and give direct guidance as what to do in given situations.

Almost nothing is a given for babies, and their potential to find patterns and learn new things is mindboggling. The idea that you can skip that part of parenting, that it would be possible for Renesmee or any new being to have a grounding in this world without time in the world implies that there is one worldview that you can be born with (quite the ontological assumption there by Stephenie Meyer) and ignores so much of the nature/nurture discussions from developmental psychology.

People expected it would be possible to make AI Renesmee chatbots. That they could just 'exist' and 'be good'. But since the 2022 LLM boom, guiding each of the big players to make 'good' decisions and give 'good' advice, has fueled so much debate, critique and questioning about what exactly that entails. Perhaps this explains the emerging 'personalities' of these models at large (would Claude be friends with Grok? Somehow I doubt it*).

AI models are not given the grace of a child to make those mistakes (and nor do I think they should be, since they are commercial entities created by capitalist companies with increasing power and influence in human society). But raising something is hard, long, iterative, and slightly mystical - and we are in the childhood of these models.

--

PS: If you want more pondering on this - check out Ted Chiang's novella 'The Lifecycle of Software Objects' which explores so much of this brilliantly. It's included in his anthology 'Exhalations' alongside so many beautiful sci-fi explorations of the world.

*Anthropomorphisation like this is contentious, but 'Chat' exists in the cultural zeitgeist as much as Renesmee now, and I think the tide is pushing us towards this heuristic shorthand of beinghood that, like all metaphor, has strengths and limitations.

Book review: We Are The Stars by Gina Chick ⭐️⭐️⭐️⭐️⭐️

This is a full-on book, but ultimately a beautifully written and moving memoir that really does leave its mark on you. It's often written in present-tense and the result is an immersion in Gina's life as she lived it. Gina's love for language is evident and her passion for the richness and beauty of life is reflected in the slightly verbose writing style, but once I settled into that I was able to truly enjoy the magnitude of her world. There is an electricity to this story, and perhaps that comes down to skill-at-memoir, but the connections from childhood through to adulthood are so well laid out and the energy and rhythm of Gina's world feels inevitable as it all comes crashing together in the final act.

From my Goodreads review, here.

Moments of Delight

No, I've not deleted or quit Instagram. But somehow in the last month, my use has gone down just so much. I have far fewer days I feel like I've been sucked into the vortex, and far more frequent days I feel I'm able to be productive. So maybe complaining about it was the trick!

So here are some moments of delight from the last month (mostly from the real world!).

That's all from me this month. Hitting send!

Indy