Fortune Telling and Speculative Design

  • #engineering
  • #fragments
  • #web

, 9 min read

https://blog.chaoticgood.computer/content/annotations/highlights/maggieappleton.com/Fortune-Telling-and-Speculative-Design

This, along with her recent Boundary Objects presentation, have got to be two of the greatest slide decks I’ll ever see.

From maggieappleton.com

Annotations

Quote

I don’t think design engineering means making fancy micro-interactions and posting them on Twitter, despite what some companies I won’t name may have led you to believe.

Oh, lordy. I’ve been out of the Twitter game long enough to’ve forgotten about seeing people showing off entirely impractical micro-interactions. Those types of post are the pre-LLM hype LinkedIn-core engagement bait. I feel like anybody actually in the space of making things knows that those are in the same category of use as… say… flight propulsion on a car that will never actually make it to market.

I feel bad that this was my first note on the post, but I couldn’t let that slide.

Quote

My preferred definition is a designer deeply involved in the technical craft and material of software. You’re a designer, but you help make decisions about functionality and architecture in code.

Expressing how this type of creativity and freedom can exist in the general sphere of “engineering” is a very underrated take. In a general shift where writing everything line-by-line feels more accidental than essential, I feel like engineers are grasping for what fills in the creative vacuum of code. Thinking holistically like this is one of many different T-shapes that a post-LLM software engineer can begin to spec into.

Quote

Second, act as a bridge between design and engineering: a translation layer that helps create solutions neither discipline could reach on its own.

After her last presentation about Boundary Objects, a short skip through the Wikipedia rabbit hole leads to the idea of Trading Zones:

Quote

Exchanges across disciplinary boundaries can also be carried out with the help of an agent: namely, a person who is familiar enough with the language of two or more cultures to facilitate trade.

T-shaped people is not a new concept, but I think the new most-valuable shape is biomodal (a W, maybe): somebody who can translate between two disciplines, and see where the two can cross-over.

Quote

We need to engage

I tell people a lot: this is a very unwise time to tune out. Regardless of where the changes end up, I think it’s valuable to engage, productively, with changes. No matter where it lands, it feels good to both be prepared, as well as feel like you had some control in where you’ve landed.

Quote

The first was digital gardens. During the COVID lockdown in 2020, I wrote a very long essay about the idea. It was a different way of publishing on the web: blogs could feel perfectionistic, performative and chronological; gardens grew over time, linked ideas together and pushed back against the feed. That idea became a bit of a phenomenon, peaking around 2021 and 2022.

I know digital gardening predated her essay, but it was certainly a jumping-on point for many folks (myself included). However, I think the theory goes a bit further: it has become incredibly valuable to both a) get comfortable finding your own voice in your writing, as well as b) have a written-down record of your own thoughts, even if they aren’t the most polished thing in the world.

On having a record – and this is advice I give often: technical interviews have always been a bit silly, but LLMs really did expose just how silly they can be. I’ve noticed a general trend of even junior engineering interviews becoming “higher-order” (design this system) over your classic LeetCode problems.

It’s really difficult to learn how these higher-order structures work without just making things – over, and over, and over again – so keeping a running record of “here’s what I’ve been up to, here’s how it’s going, and here’s what I’ve learned” is a great way to demonstrate not just what you know, but your experience with coming to know it.

Quote

I predicted an explosion in personal software as models made more people capable of building things, and a rise in local developers who could make software for their communities. Even after the talk, people told me the models wouldn’t be good enough. Now it looks like that prediction is coming to pass.

I went to a really great presentation by Alex Voss last weekend called Can we use AI for better ends?. I think a good framework on which to think about the question is: what is possible now, that wasn’t before?

One of the pieces of data that I’ve personally seen, in my own work, is that there’s been a rise in both:

a) the number of US counties that have websites; and b) the number of US county websites that adhere to ADA guidelines for website accessibility.

It’s pretty obvious that some of these sites had a bit of a(i)ssistance for their frontends. None of their actual data was hallucinated, but, y’know… sometimes, you can tell. I wouldn’t say this is any different than the previous trend of, say, Squarespace-assisted web development, though.

Ultimately, the fact that the barrier of entry for creating something online can be a positive thing. More folks than ever, and more folks who need accessibility tools like screenreaders, can now access their community’s information and resources.

I’d call that a pretty solid upside.


After pulling Dr. Voss’s Ph.D. thesis (Corealisation: A Radical Respecification of the Working Division of Labour in Systems Development), the abstract and introduction hum at almost the exact frequency as the definition of “design engineer”, along with many other portions of this presentation.

Quote

“Not invented here” syndrome

I can’t believe this is the first I’m hearing of the formalized “not invented here”.

This has almost certainly gotten worse in the last 12 months, with the rise of decompilation/“clean room” re-ideation. I’ve seen it a lot on an interpersonal level, even: the whispers of “why would I adapt to this tool, when I could make a bespoke one myself?”

There is almost certainly a new “generate” option to the old buy versus build debate that has dialed up NIH instincts to 11.

Quote

When you come up with a theory about the future, ask where it sits. Is it possible – allowed by the laws of physics, even if unlikely? Is it plausible? Is it probable? And then there’s preferable: a future we would hope to see. Some futures might be possible or probable without being preferable, but they’re still worth exploring.

A bit of this peaks out in Planning with Agents: the preferable, and plausible, future state that LLM tokens at a current frontier quality will drop, substantially, in cost.

Quote

First, we’ve been sleeping on non-LLM intelligence: decision models, classifiers and embeddings. These aren’t new, but we’ve been focused on frontier language models and ignored everything else.

Dr. Fei-Fei Li has an interesting section about this in her book, The Worlds I See, where she talks about a (pre-LLM) sensor and analysis setup that would help hospital staff know when they’ve washed their hands enough to reduce the risk of preventable infection.

She called it ambient intelligence: things around you being just the perfect bit smarter.

Versus, of course, the heavy Claude-isms.

Thinking... (2m)

Percolating... (4m)

That's a great idea!
Quote

Another exercise is to maximise and minimise. Our industry loves maximising: what if you had 100,000 agents? I don’t even have a hundred. What if you had a thousand agents that cost a pound a day? What’s possible then?

At the Alex Voss talk I mentioned earlier, somebody brought up an adage that’s been ringing around in my head for awhile.

The interesting question isn’t “what can I do with a team of geniuses?”

It’s “what can I do with an unlimited supply of morons?”

Quote

I asked ChatGPT to make old-fashioned adverts for Codex and Claude. I love the retro look and the promise of agents doing things like planning a day out, checking the trains and remembering what everyone needs.

I gotta say – this is probably the most tangible, actual benefit I see to AI-generated art. I can see myself doing it, but never, ever admitting it to a soul outside of my immediate vicinity.

(I’ll type my thoughts on the distinction of “slop” either at the end of, or after, finishing the annotations for No Silver Bullet. The TL;DR is that this isn’t slop because “create art” or “publish this” was not the essential part of what was being done, here. The essential part was “explore an idea out loud”, which was accomplished).

Around the Web

No discussion yet. Be the first to reply.

On Bluesky

No discussion yet. Be the first to reply.

On Hacker News

No discussion yet. Be the first to reply.