Saying goodbye to the UKI Ignite team.

HCID Conference

On September 18th I went to HCID with, and on the recommendation of, Stephen Wood. This is a design conference at City University of London and there were some great speakers:

• Tobias Revell at Arup presented their approach to Design Futures and how we can help stakeholders think on longer timelines and create empathy for non-human stakeholders.
• James Gadsby Peet talked about Equity by Design. A lot of what he talked about was starting with the client-agency team relationship, making sure the goals are shared, and building deep trust. I'm stealing his "On this team it's OK to..." frame as a simpler way of coming up with shared team agreements. - He also runs a community-owned pub, which I'm very intrigued by given there's an empty pub just round the corner from me.
• Sara Heitlinger talked about moving beyond a human-centred approach to cities and how she's used multi-species live action role play to help stakeholders role model the perspective of different stakeholders including nature. Similarly, John Tweddle at The Natural History Museum talked about his role in creating advocates for nature and helping people feel confident to speak up on its behalf.
• Tony Sampson traced the current state of User Experience design. There was some great critique in here including: - Reliance on "we asked the user" as a trump card over existing expertise. This made me think about The Unaccountability Machine (and how I really must actually read it). - In slight contradiction to that was the point that, in UX, choice is illusory. Designers guide people down a path using cleverly designed choice architecture. - How UX is 'spatially detached' from cultural context. Designers are still not spending enough time immersing themselves in the lives of the people they are designing for.
• Georgia Panagiotidou shared a really interesting experiment around energy communities where participants who were all sharing energy from solar panels had to negotiate energy use with each other. For example, if two people wanted to do washing at the same time, but the panels couldn't support that much energy usage, someone would have to do it a different time. There was a standout insight here for me. Apparently people wanted AI to do the negotiation for them to avoid the awkwardness. Again, The Unaccountability Machine seems relevant. The larger point of this talk was that lots of behaviour change interventions are aimed at individuals, but real change needs to target communities and help them work together.

Papercamp 3

Two conferences in a week! Papercamp was held on a Saturday(!) and brought together a bunch of designer types who nerded out about paper as a tool for a day. It was lovely to bump into Micha Nicheva on the way and hang out with her and husband Sean for the day. As well as getting to hang out in the St. Bride's Foundation's print room, do some printing, and hear about the newspaper presses of the 70s/80s, the standout talks for me were:

• Dan Catt talking about his experiments with AI and Handwriting. Dan has developed a personal AI who journals for him and writes notes to friends in Dan's handwriting. It's worth exploring his website. There are some really interesting implications for me around memory and beliefs, but maybe these questions were all answered when people used to have scribes/secretaries to take down their dictation?
• Stefanie Posavec took us on a journey through the history of paper as a tool for organising information and thinking with. John Willshire later took up the baton and honed the focus onto cards (you can see a version of the talk here). Loads of interesting stuff in both of these.
• Steven Watson took us through the development of magazines and shared a few of his favourite contemporary examples. My favourite was Dysfluent magazine where the editor created a typeface to represent stammering and transcribes interviews exactly as they were spoken, stammers and pauses and all.
• Matt Webb took us through his journey from tweeting a picture of a prototype AI poet-clock to manufacturing a product after a successful Kickstarter. I don't think anyone, anywhere, has given as much thought to what it means to use AI for commercial creativity.

The efficient decline of empathy

Given that both conferences were populated by designers, it might be unsurprising that the recurring theme for me was around empathy. How paper and handwriting can create it ("A real person wrote this!"), and how technology can erode it. From What Happened to Empathy?

Empathy is cultivated through interactions with people we don’t know well, those glimpses into other interior worlds. We have, over the past two decades—slowly and then quickly—“optimized” other people out of our lives. One app at a time, we’ve greatly reduced our need to casually engage with anyone we don’t know—or even to meaningfully engage with those we do.

There's a business lens here too. I happened to come across What the language of business reveals about the corporate soul which shows that "terms related to the human experience–emotions, creativity, and aesthetics are underrepresented in business language, relative to the language of our daily lives". This not only dehumanises employees, contributing to their lack of engagement, but the analysis also found that businesses whose language uses terms more associated with the human experience are more creative (more patents and citations) and achieve better ESG ratings.

There's often a fear that businesses will become less empathic with the adoption of AI, but there are a host of experiments showing that LLMs can actually be more empathetic, for example, scoring higher than human doctors on tests of bedside manner. AI is just a tool, but most businesses want to use it for greater efficiency gains, not deeper empathy for their customers. These goals don't have to be in conflict with each other. You can build empathy more efficiently. But questioning whether efficiency should be the sole business case for AI is something only humans can do.

Will AI kill SaaS?

I posted a while ago that I was starting to use Clay to get better at keeping in touch with people and stop using LinkedIn so much. It's certainly helped me with the first part but, at $15 a month for all the features, I just don't think it's worth it. But could I recreate some of those helpful features in Obsidian where I do most other things? It turns out with a little bit of code, and a bit of help from ChatGPT, I can. So now I have a searchable database of people directly in Obsidian and, because there are other plugins to make your Obsidian Vault accessible to LLMs, I can use those features to ask queries like, 'Who do I know who has worked at [company]'. There's an argument that AI is going to kill a lot of SaaS software and after spending not a lot of time on this experiment, I'm giving more weight to it.

The previous section (written later) casts a different light on my efforts to use software to help me get better at staying in touch with people. I'm alright with it because I am explicitly not trying to increase efficiency (i.e. I am not doing this to meet a goal of connecting with X many people a day). I'm doing it to help me create deeper connections with the people I meet.

Valencian Textures

Finally, here a few photos from a couple of recharging days in Valencia (sin niños!).