Thoughtwax
a new type of computer
Traditional computing could be boiled down to operations over numbers. Every app, microprocessor, website, and shitpost is ultimately reducible to binary digits, the star-stuff of digital information. Every output is entirely predictable. *slaps roof of traditional computing* Yessir, reliable foundations.
But AI’ll soon soften your cough. It’s no longer just operations over numbers; it’s operations over information. AI systems essentially eyeball some data, perform the robot version of a rorschach test, and then just freestyle it live. AI makes gloopy, fuzzy, inconsistent decisions.
Neither are without their charms. In fact, each can be both astonishingly capable and remarkably dumb in their own special ways.
What you really want is an elegant combination, the best of both worlds. So now an intriguing design puzzle is emerging: how do you take these two very different approaches to conceiving of what a computer even is, and alloy them into something new?
Maybe you try to tame the LLM: add guardrails and safety nets and suggestions and whatever else is needed to bring the beast to heel long enough to get some work done with the damn thing. Or, maybe you try to extend the existing paradigm by stuffing little AI trinkets into existing products. Lots of people are now excited about Agents, the goal-oriented, System 2 reasoning, action-taking, and generally better-looking descendent of crummy old chatbots.
So computers are about to get weird! There’s a ton to figure out. No doubt it will all make sense in retrospect. But it can also be hard, after a big shift like this, to viscerally remember what the before state was actually like. Did the resort match the promise of the brochure? What didn’t you know back then that seems obvious now?
For fow we’re still in the middle, with the before and after states still only an armspan away. So setting aside the bigger societal questions, here are some high level product-ey questions that I don’t currently have confident answers for, but I’m interested to see what happens:
- Will probabilistic LLMs become like determinisitc computers faster than deterministic computers can become like probabilistic LLMs?
- What new UI patterns will we have for AI-native applications? Will we see a “pull to refresh” moment for AI interfaces? (My esteemed colleague Des is maintaining a great thread of emerging patterns here.)
- How will this change how products are sold, and how will market share be divided? Is SaaS really dead? Will we move from subscription-based to usage-based pricing? Is the money at the model layer or the application layer?
- What level of the tech stack will Agents live at? OS-level (e.g. Siri), browser (Gemini in Chrome), application (ChatGPT), or product (Fin by Intercom)? All of them? Is there a Dunbar’s Number for Agents I’d want to use? Will they talk to each other?
- Will highly capable AI Agents become the ultimate general-purpose application? How much will it displace? Taken to the illogical extreme, might we end up with just one piece of software that does everything?
- Are we all screwed?
The Tools We Use
Sure, I could have posted this talk I gave at UX London last year sooner. But I decided to sit on it in case the global social, economic, and infrastructural system happened to go belly-up and people might want a little distraction.
I possibly went a little self-indulgent on this one, rambling through hammers, monkeys, 2001: A Space Odyssey, pencils tied to bricks, the Whole Earth Catalog, the Space Jam website’s HTML, and 1980s Toyota manufacturing all before arriving at a simple point: designers should iterate on how they work as much as they iterate on their work itself.
If that’s your type of thing you might enjoy it. If video or Irish accents are not your type of thing you can read the full transcript on the Intercom blog.
This Too Shall Pass
“This too shall pass” is often said to struggling new parents. Solid advice! You’re cooped up at home, minding a baby, unsure how to cope, no clue when things might be normal again. It’s a gentle way of advising someone: I know. But you’re going to have to dig in now. You will endure.
“The days are long, but the years are short” is the best parenting advice. In truth there’s a melancholy to being a parent: you’ll want to jump out the window at times, but when it’s all over you’ll wish for that time back again.
“When this is all over…” is something people have been saying a lot recently. When this is all over we’ll meet up more, we’ll have a new appreciation of nature, we’ll hug our parents, we’ll pay the nurses what they deserve, we’ll reinvent society, we’ll have the mother of all sessions.
When this does pass we won’t pine for it. But a couple of weeks in, I’m now trying to notice the little moments in all of this. Our little family cocooned away. We’re doing okay. This is going to be the only summer we’ll ever have with the kids being 6 and 3. There are small moments to be had. These days do feel long, but the year will be short.
TLDR: We’re in lockdown and my kids are giving me Stockholm Syndrome.
Design in Interesting Times
Just realised that I never posted my talk from the 2017 Inside Intercom World Tour. It’s about using system thinking to cope with rapid change, whether that’s inside a fast-growing startup, or a world that seems to be going to hell in a handbasket. I gave versions of the talk in New York, Tel Aviv, and finally in San Francisco, which is what’s recorded below. Had a ton of fun working with lovely folks from our Brand Studio on the visuals for this one.
Full transcript on the Intercom blog, or there’s an audio version if that’s your thing.
Stories and storytelling
In my post last week about Google Duplex, I mentioned Tristan Harris’ 2013 internal presentation at Google about the ethics of attention.
I looked for a public copy of it, which didn’t exist last week, but does now: it was just leaked/released. It’s an interesting read in itself. But I still often think about this actual presentation – the slide deck itself – as an example of GREAT storytelling. The whole thing is designed to be viewed, not presented. Each slide is a single thought; just an image and a few words. There are no speaker notes. It’s an essay designed to be viewed rather than read. I remember it spreading like wildfire through Google at the time and immediately sparking conversation. It’s a object lesson in the power of presenting your ideas or work narratively.
Another post last week was about the Dissect podcast and how new types of media lead to new forms of content.
Tristan’s slides feel like a desktop predecessor of the most obvious new mobile-first content type: the Story. Invented by Snapchat and mainstreamed by Instagram, Stories are the ne plus ultra of smartphone patterns: atomic units of vertical rectangle content navigated via the simplest possible interaction. Tap, tap, tap, one pellet of info at a time. But done well it can add up to a fully realised narrative told across a single day.
There’s a lineage here. Robin Sloan’s 2012 “Fish” tap essay feels like a more literary precursor to Stories. Both Fish and Tristan’s deck are writing, but the delivery is edited into discreet cue cards to create rhythm and emphasis. Twitter threads are part of this scene too, being a series of atomic thoughts strung together into an argument. Bad threads are bad because they are nothing more than longer essays arbitrarily chopped into 240-character chunks; get a blog, dude. But a well-written thread can use the limitations of the medium to create pacing and tell a story, one thought at a time, and work better than it ever would as prose.
DVD commentary for music
I’m currently deep in binge mode with Dissect, a music podcast that picks an album and spends an entire season analysing it, one episide per track. It’s a one-man show entirely created by host Cole Cuchna, and it’s serious music nerd-out territory.
The first season takes on Kendrick Lamar’s To Pimp a Butterfly, an album which I had already loved, but in retrospect had no real appreciation of the depth of theme and narrative throughout it. Kenrick’s album is genuis on a level that I hadn’t understood before listenting to Cuchna’s 12+ hour critical analysis of it. Which to me is not a bad return.
Dissect fits into what seems like a new category of “breaking down music” podcasts: Switched on Pop (previously recommended here) and Song Exploder do somewhat similar things with individual tracks.
It strikes me that this is a new genre that could really only ever exist in podcast form. More than ever music that deserves analysis like this has become an internal affair, listened to privately on headphones while working or commuting. Podcasts seem like a natural fit. They are well suited to audio exploration, obviously, but also match the intimate nature of how I think people dissect music themselves: largely as a private internal monologue picking apart tiny phrases or production touches. There are probably hundreds of brief moments in songs that I privately recognise as something that resonates, but it would feel weird to talk to someone about. Yet the one-on-one format of someone speaking directly into your ear with no time limit seems to create space for that, even if it does mean a 40 minute episode discussing a 4 minute song.
Looking back at the short period during which DVDs were actually a thing, I only miss bonus material: those little extras that were often included with a movie like a director’s commentary or behind-the-scenes featurette. For a certain type of person, me included, understanding how or why something was made only increases my appreciation of it; to me, that’s the main function of good criticism. The rise of YouTube video essays seems to have bridged the gap for movie criticism, but I can’t think of anything similar that had already existed in this space for music.
Which is why Dissect and other podcasts like it are exciting to me: the content is often great in itself, but it’s also cool that there are new types of media to be invented that are truly native to these new formats – in this case in-depth musical analysis. Who knows what further new art forms podcasts or Stories or messaging apps or whatever else will throw up.
Anyway, I reccomend Dissect, I guess is what I’m saying. Season 2 did Kanye’s epic My Beautiful Dark Twisted Fantasy, which I’m fully here for. Meanwhile Spotify gave Cuchna a full time job and Season 3 (subject still a mystery) is dropping next week. Good times for music nerdery! \o/