I appreciate it’s been a while. I just returned from an actual holiday that turned into a small series of rolling disasters, like crossing a room full of lego bricks on bare feet. Before that I was in Eindhoven as part of my visiting professorship at the university and at some point I should probably capture my thoughts properly and write that up. I started this while there, in fact. I’m also trying to really roll into PhD work at the moment, the threads are coming together and I can feel the flow but it’s just too difficult to pull substantial time together between all of these side-gigs, a full time (x1.2) job, a three-year old and trying to do things I enjoy like exercise.

Musn’t grumble. Now look, here are five things.

Five Things

1. The Existential Risk of #BatteryPolitics

It’s possible that by 2040, 9% of energy in some cities might be supplied by electric vehicles, which are effectively batteries on wheels. We could speculate further and add domestic home batteries into the mix and say that maybe even as much as 30% of energy in a city might by then be coming from individual private battery owners selling back to the grid from energy generated by their own renewables or bought at a previous time and price but stored in batteries. Basically, the locus of control over energy supply and demand starts to spread into people’s homes and cars.

The first implication of this is a massively diversified marketplace for energy. Where currently, a few dozen major actors might be active on a grid in a city this might be multiplied to tens of thousands, hundreds of thousands, maybe millions of individual citizens.
The thing is, those individual citizens are now in charge of delivering critical infrastructure.

And imagine a world in which those citizens just so happen to be political actors who are miffed about schools or conspiracy theories or whatever and maybe decide to withhold their own energy from the grid in protest. An implication of this distributed critical infrastructure network is that, similar to blocking road, withdrawing labour or boycotting products, energy is now in the popular arena. This is perhaps the downside of #breezepunk – that this dispersing of energy generation and storage to atomised individuals brings their own proclivities and preferences to bear. While we quite rightly might be concerned at large companies intervening in critical infrastructure because of their own political drivers, I haven’t seen much consideration of the risk of a multiplicity of private citizens doing the same.

So what? I genuinely think this is a pretty massive and undiscovered area of consideration with the idea of decentralised infrastructure put in the hands of citizens. We were working on a project about AI and Energy when this notion suddenly struck. Infrastructure is a political operation and the redistribution of this introduces the politics of those its introduced to. I might look into this or think more about this.

2. How do you balance AI skepticism vs hype?

I really don’t have a good sense of what the future holds but I’m more disinclined to believe the assertions and speculations of startup founders and venture capitalists saying how Ai is going to change everything (and personally stand to gain) than social scientists saying it won’t (and for whom either way, it will have little impact).

So what? I find myself torn between these two poles and that’s right, that’s critical thinking; being able to see wood and trees, fact and opinion but how does one avoid being maligned as a doomer. How do you know what’s worth paying attention to? I suppose the point is to stay on your toes, stay reflexive. ‘If someone says it could do something, ask why it isn’t.’

3. FOFB

This piece went out where some rich newspaper tycoon guy decided to build an ‘AI newsroom’ then immediately proceeded to make sexual advances on one of the avatars.

Between the sexual advances on chatbots and the ‘I’m also worried about the torment nexus and want to continue to work with people preventing the torment nexus however, in the name of science I have built the torment nexus’ it’s a great blog about all of the things that are wrong in this whole maelstrom of shit. Basically every upturned garden fork, he trod on, gleefully and then responded in comments to say it was all in the name of starting a conversation.

So what? Well you see, all that aside, what it highlights for me is the increasing trend of Fear of Falling Behind. You see it most in LinkedIn hucksterism, the idea that to stay relevant you must loudly proclaim your AI bonafides, the louder the better. It’s a less extreme version of Singler’s work on transhumanists; that underneath it all is a desperate anxiety of becoming irrelevant. It’s certainly a trend worth keeping an eye on. Loud and obnoxious professions of faith, whether in so-called AI or disastrous economic policy seem to be very de jour for 2025.

4. Empathy grind

I’ve had to, after putting it off for as long as possible, write a bit about empathy in AI systems for the PhD. I really was hoping I could either swerve or duck it. It’s an enormous Can of Worms which is super vague and hand-wavey and makes me feel a bit icky to write about. But I need to get from “designers use anthropomorphic cues to increase engagement and therefore profitability” (fine, very good evidence, based on hard stuff) to “people express feelings for machines and that changes their behaviour” and unfortunately there’s no way to join those two things without trudging through the vague treacle of ’empathy.’ It took about three days to form up 260 words that didn’t make me want to gag relying mostly on Turkle et al., (2011), Abercrombie et al., (2023), Claypool (2023) and of course Bender and Koller (2020) with a bit of Manzini et al., (2024) and Ibrahim et al., (2025). No I don’t want any recommendations, I never want to think about it again. I’m just telling you so that you never have to go down there.

So what? Just doing my public service. Actually, having written this, the one that is pretty grounded is Sundar and Kim (2011) who prove that anthropomorphic cues trigger heuristics. So, similar to when someone says ‘trust me, I’m a doctor’ you fall into a conversational register, mindlessly, when a chatbot says ‘I’m here to help’ you mindlessly do the same and treat it as a helpful anthro-thing even though you know it isn’t because the heuristic kicks in. Thanks Dave for the recommendation there. Again though, no more recommendations.

5. Stay Focussed

Which reminds me. In my desperation not to have to read or write those 260 words I spent a couple of hours refreshing news websites or browsing Reddit and Pinterest. I already have a screen time lock on my phone for the dangerous apps which I have found has done wonderfully to alter my behaviour (I was sick a few weeks back and laid up on the sofa was shocked to find out I’d spent four or five hours on Instagram as if it was nothing.) I got Stay Focussed for Chrome which essentially does the same thing for websites you list and has been pretty good. This largely works because I’m an ‘eat your greens first’ kind of guy and think ‘if I’ve only got ten minutes of Instagram time today then I’d better save it for later’ and invariably end up never using it (except for posting).

So what? I don’t know, maybe you’re profoundly disciplined, maybe you think all the stuff about social media having profound and damaging effects on the brain is wrong.

Shorts

• 4chan is gone. Just sort of happened.
• Rival Strategy’s new report on quantum tech and the arts to coincide with the 100th anniversary of quantum computing.
• I like Ionna Sotritou’s blog.
• The junior designer career path barely exists any more
• Cameron Tonkinwise interviewed by PGR
• Wes and Eva have finally started the Critical Climate Computing research group at UAL.
• Honor sent me this thing on community and resilience. I have a default cringe setting to ‘community;’ it’s all too often an aesthetic exercise where the mere notion of the word is invoked as a salve to all problems. But in this case the article is urgent and references some good science, including heart-throbs Wennigrow and Graber.
• This via Danah – things London has lost in the last ten years.
• The London Interdisciplinary School where I occasionally teach has a new MBA.

Recents

• I was back on the Near Future Laboratory podcast with Julian talking about the chapter we wrote together. I always enjoy these natters with Julian and you can listen to the end to hear real life footage of use trying to organise our calendars.
• Next week is Re Shape. Just look at that lineup. I’m quite serious about how good it’s going to be.

I’ve been bingeing pods so now music this week really. Sozzy. Ok let’s get back to work. Love you, speak next week.