Thursday, 3 July in Portland, Oregon, which is the day before the day that the main events of Independence Day (1996) occur.
\nJoin myself and Matt Jukes for my inaugural How People Work, Live! on Friday 10 July at 9am PT.
\nHow People Work Live1 is based on my How People Work workshop, where I get together with interesting people to talk about what it takes to work successfully in teams. We’ll talk about influence, relationships, communication, strategy -- basically most of the software development stuff in this newsletter -- and because it’s me probably also some tortured Star Trek references.
\nMatt Jukes is a veteran digital leader and product coach with more than twenty five years of experience leading digital and product teams for public service organisations like the Office for National Statistics, Essex County Council, Jisc, the BBC, the Department for Business and Trade, the Ministry of Housing, Communities and Local Government and the Government Digital Service.
\nFriday, 9am to 10:30am PT, on Zoom.
\nUnlike the Hallway Track series, these events will be recorded and are open attendance. Here’s the registration page again.
\nTwo things today.
\nYes it’s a Random Post On Social Media, but my attention was caught this morning by this post by Len on Bluesky:
\n\n\nNintendo is going to be the company that survives two hardware crashes. If I had to call it in the air, their competition in the 2030s is going to look like Huawei. Chinese electronics companies currently doing other stuff deciding to do what Sony did in the 90s.2
\n
(the post is accompanied by a graphic table showing 2026 Major [Videogame] Console Price Increases, ordered by the dollar amount of increase (ascending), starting with the Nintendo Switch 2 at $50 and ending with the Steam Deck OLED at $300).
\nSo the whole preoccupation with AI has broken the technology industry in that manufacturing capacity has been (rationally!) switched over to the components that will provide the most profit margin, which in this case is a specific kind of high performance memory used in AI training and inference. (Training is, well, the training; inference is the process where a model creates output based on input)
\nNintendo famously have a strategy of not chasing after technological supremacy in the way that historically Sony and Microsoft have. The shorthand here is that if you get Sony and Microsoft’s latest console, you’ll get the best videogame experience for your buck, where “best” has historically been along the lines of stuff like “best graphics” or “highest fidelity graphics”.
\nNintendo don’t do that -- ever since the Wii at least they’ve deliberately pursued a strategy of using older technologies and figuring out how to make them interesting in terms of interaction, which is how they get the title of Videogame Console That Has Broken The Most TVs Through Playing Tennis Or Some Other Thing.
\nI could be boring and say this is some example of zigging instead of zagging or skating to where a puck is going to be or deciding you’re playing a different game from everyone else, but if I did that I’d be writing an airport book for an earlier time.
\nIt’s not that Nintendo went around explicitly trying to formulate a strategy that would survive a test of “oh no, there’s this horrific AI boom and now everything’s gotten more expensive”, because Nintendo aren’t immune to the effects, it’s not just about a reliance on high-end semiconductors but also just a lack of capacity. But! I’d say Nintendo certainly been affected less!
\nAll of this is happening against a background of Microsoft still not figuring out what its games strategy is other than “maybe we had more money than everyone else?”
\nThis one is JauntyWunderKind’s fault, who wrote this on Bluesky:
\n\n\ni wonder how much ai psychosis is that people have never ever ever had help or labor they can access before (✋) and are just totally unaccustomed to, unused to, overwhelmed by such a world. by having help. man, it feels great.3
\n
I am not kidding here: this totally reminds me of Sarah Connor’s line from Terminator 2. If memory serves, this is a voiceover of Sarah’s around the end of the 2nd act of the film, where she’s watching John mess around with Arnold the T-800 in the desert before they go off to kill Sam Altman4:
\n\n\nWatching John with the machine, it was suddenly so clear. The terminator would never stop. It would never leave him, and it would never hurt him, never shout at him, or get drunk and hit him, or say it was too busy to spend time with him. It would always be there. And it would die to protect him. Of all the would-be fathers who came and went over the years, this thing, this machine, was the only one who measured up. In an insane world, it was the sanest choice.
\n
(Unlike contemporary LLMs, there’s not that much evidence that Arnold the Terminator would encourage John to die by suicide, even though there is a bit of valorization of suicide when Arnold the Terminator does his infamous thumbs-up scene (no relation to Ryan Gosling))
\nThe bit that I latched onto here is Jaunty’s point that people have never ever ever had help or labor they can access before.
\nLLM chatbots like ChatGPT and the Claude models are... I don’t know, aligned to be nice? By default, they generate text that reads as patient and non-judgmental. The thing that makes your mirror neurons for this-is-a-human fire like mad will wait, and does not mind if you do not know or ask questions. I mean yes you need to set aside the whole “frequently provides inaccurate information” and “is dangerously sycophantic” but, you know, apart from all those other things the Romans have done for us, it’s kind of horrifying how one thing chatbot LLMs have is infinite patience and, well, a kind of kindness?
\nAndrew Losowsky had an excellent riff on this that I think makes emotional valence(?) of chatbot LLMs clearer:
\n\n\nI’ve said something similar to editors in newsrooms. “If you tell it it’s wrong, it immediately apologizes, acknowledges the errors, and tries to do better. Are you surprised they prefer it to you?”5
\n
... and then Ryan Randall connected the concepts to AI-wives because the whole infinitely patient thing is also so close to subservient, uh, servant who is unquestioning.
\nIt feels like a horrible indictment of the world in how quickly and how easily a pendulum swings towards being fooled by a text-generating machine that looks like it’s being patient and understanding and non-judgmental. But then again, so was Dr. Sbaitso and Eliza. Maybe it says something to exactly how much of a social species we are and how much we yearn (or are biologically biased towards?) social connection that we have such an almost instinctive reaction to such tone of voice. (That said, all of this is from a western perspective. Maybe this is significantly different in other cultures! It would be a super interesting soc-anth/socio-linguistics research paper, to me at least?)
\nBut no. Instead we have a world where you can get patience and non-judgment on-demand, in your pocket, from anywhere, and best not think about where it came from, or how it’s funded, or what the quid-pro-quo is for its existence and availability. Who could’ve known that it’s so important for humans to not feel stupid or not be worried that they’d be laughed at or otherwise for just not knowing something? I mean, what does that say about us as a species, again?
\nThe joke of course is that Skynet was totally not aligned with human values and it was only future John Connor who was able to take a T-800 and un-align it so that it would be nice to contemporary John. And that’s even forgetting the fact that Arnold was sent back in hardware-locked inference-only mode -- there’s that entire scene where John has to more or less flip a DIP-switch on Arnold’s GPU to turn on training mode.
\nAnyway. A thing-that-listens.
\nThat’s it for this episode. You can always say hi, you know. I love emails where people say hi, and then I reply.
\nHow are you doing? It’s complicated!
\nBest,
\nDan
\nHow People Work Live is, I have to admit, a potentially confusing string of words. How People Work-Live? ↩
\n(4) Post by @len.bsky.social — Bluesky (archive.is), Len, Bluesky, 30 June 2026 ↩
\nJauntyWunderKind on Bluesky, on 28 June 2026 ↩
\nActually Miles Dyson ↩
\nAndrew Losowsky on Bluesky, on 3 July 2026 ↩
\n