A one-off, special, never-to-be-repeated Acts Not Facts weeknote

18.25, Monday 6 Nov 2023

I re-read Robin Sloan’s The art of working in public (Snarkmarket, 2011) which riffs, in part, off my Week 315 post at BERG. I was big into weeknotes.

Work in public. Reveal nothing, says Sloan.

Well consider this a weeknote for my new-ish micro studio Acts Not Facts.

Just this once though.

It’s Monday morning. I’m standing outdoors on the train platform on the way to a workshop tapping this out with my thumbs.

Coming up this week

  1. PartyKit – mainline project. My summer residency was about exploring multiplayer, realtime apps (blogs posts and prototypes here). It’s a different project now. Loosely I would say doing the groundwork for product strategy and go-to-market, but via making things rather than getting lost in plans. Writing code, making Notion pages.
  2. Oh I need codenames. I need to speak to comms to check I can name this one. I’m helping to set up the AI experiments pipeline in a big org. The premise is that AI is a 10 year leap forward and we’re imagination bottlenecked: how do you explore and learn with a portfolio approach? On day 1 (a couple weeks ago) I sketched out what the next big bet could be – for my personal comfort I need to know that there’s at least one workable idea on the table. There’s an existing, great team, so I’m working with them 1 day/week. This week is the first of the workshops to start capturing ideas.
  3. Another one I’m not sure about naming. I’m on my way there now: a product strategy workshop for a hardware startup approaching product #2. We’ll look at roadmap, commercials, and brand. I like these broad, spilly problem spaces that take synthesis to arrive at a concise, memorable approach.
  4. On Wednesday I’m doing a talk at PA Consulting about my AI Clock, currently working its way towards a Kickstarter. Looking forward to talking about the prototyping and design process, and also sharing some meta thoughts about both opportunity spotting and large language model vibes.

I’m in a cafe opposite the Royal Courts, tapping on my laptop. This is my current pov. My idea was to bang out these words before the week proper kicked off but I don’t think that’s going to be possible.

Here’s the premise of Acts Not Facts, which is my new-ish product invention studio, but it’s actually just me:

  • As the homepage says the studio is interested in AI, multiplayer (i.e. group use), and embodiment (learning from physical space, cognition, etc, and physical computing). Not all three areas at once necessarily.
  • Why the name? For the last decade or so, the way to bring new products into the world was to think carefully and make PowerPoint decks and cover the walls in post-its. No longer. The landscape of possibilities is unknown so the appropriate approach is to roll your sleeves up. Things-which-are-made teach you about the technology, open up new thoughts, and (vitally!) let you work with people who aren’t as close to the technology as you but probably have better ideas.
  • The plan is to find projects that I can achieve with a small team hired by the studio. That means building a small portfolio on my own for creds. That’s the stage I’m in now.

I’m lucky that these first engagements align with the Venn diagram on the homepage.

Ok, workshop time, wish me luck.

Media last week

  • I run an unofficial archive of my favourite radio show, In Our Time, which is on the BBC and been going about 25 years. There are 1,000+ episodes on every topic under the sun. I prototyped then then shipped a new AI-powered search widget which you can use right now on braggoscope.com. Tap Search in the top links section. It’s neat because it uses “embeddings” – a kind of semantic space. So you can search for jupiter but you can also search for the biggest planet. My post about this did pretty well on LinkedIn. People like to see simple but effective uses of AI!
  • Venkatesh Rao published his new essayOozy Intelligence in Slow Time which builds on and says very positive things about a recent blog post: Matt Webb’s excellent reframing of AI as intelligence too cheap to meter… – I am v proud to be on vgr’s radar to say the least.
  • I spoke recently in Hamburg at NEXT23 about human-AI collaboration, and playful experiments, and after the talk recorded a 30 minute interview about product invention. It was a super fun conversation and has now been published: Exploring AI personalities and poetic clocks (YouTube).

Think of the studio as a funnel. Awareness becomes conversations becomes work.

I don’t know yet what the right shape for a project is – it takes a while to learn where my edge is and what resonates in the market. Ideally I would be consistent building awareness around that.

To begin with it is enough that people know I exist via an activity which is more efficient than 1:1 conversations. Hence tracking media.

Brief situation report: I am sitting outside on a low wall eating a cheese sandwich. The product strategy workshop went well I think. It is useful to be able to map out a company’s products and activities and compare that against all the other companies I have seen. It means I can say what is different or what feels missing. Anyway, the 2 hours of immersion resulted in 3 recommendations.

I’ll carry on writing when I’m on the train home this evening.

Also in my in-tray

I regard the AI Clock as a studio project. I’m pricing up a second manufacturing route before starting the crowdfunding campaign. An email of questions arrived this morning that I need to respond to. Then there’s a ton of small tasks to do with content for the campaign itself… and I need to line up media…

And remember I said that awareness leads to conversations? I have an email or two to respond to on that count too, but before I do that I need to figure out an ask… like, if the point of a meeting is to beget more meetings then how do I ensure that happens?

The scarce resource is always attention.

I have a stack of experiments I want to build. But it is never top priority to go through my list and draw/develop/combine/prioritise/next-steps then all.

Let alone those that could be businesses in their own right! Those are the ones I really need to carve out time to develop.

I ran across the idea recently that you operate a queue-based or stack-based priority list. With a queue, you work down from the top, crossing off tasks as you go, and add new ones to the bottom. Stack-based: you add and take from the top, optionally re-ordering the stack based on priority.

I have always been stack-based. (Which anyone who has ever emailed me will recognise. If the email is in the top 6 at the exact moment I am looking at my phone, I will reply immediately. If not it will be buried forever. Sorry.)

Which begs the question: given scarce attention, how do you get to the tasks which are lower down the stack but none-the-less vital? Answers on a postcard please.

The smart thing to do would be the take my wish list of experiments, and cherrypick for possible collabs. Sketch a project pitch every time I’m about to walk into a conversation with potential clients, so I have something in my back pocket.

Huh. Maybe I actually should do that.

Nothing profound. It’s dark now and, although I’m on the train, my thumbs are slow from typing in the cold on the platform. So time to wrap this and publish and maybe I’ll have time to reply to the manufacturing folks before I get home.

UPDATE: Email sent. Approaching my stop. Save. Commit. Deploy.

More posts tagged:

If you enjoyed this post, please consider sharing it by email or on social media. Here’s the link. Thanks, —Matt.