Apple Games icon

Apple is reportedly working on an Apple Games app. But we’ve all played this game before, and I’m not sure it warrants an extra life. So over at Stuff, my column this week is ‘What I think the Apple Games app needs to work – and why it won’t’.

Arc browser is dead. A blog post outlines the company behind it is pivoting to new product Dia. Two things. First, I’ll be sorry to see mobile app Arc Search mothballed. It was an excellent example of the potential in AI-assisted search, serving up magazine-like synopses with full source links. Secondly, when you’ve built a niche product that commands a loyal audience and pull the rug out from underneath them, what makes you think they’ll be back for more?

I asked ChatGPT to be Jimmy Carr. This in response to his arrogant bullshit about AI being better than art and design. Doubtless, he’d argue it was a joke. But, hey, people might argue that about Carr. Anyway, my thread suggested his quippy one-liners can be adequately replaced by ChatGPT. (Or can they? Probably not! But I’m not a comedian and so don’t fully know what good looks like in the context of new puns. Which is the entire bloody point. If you don’t have humans in the loop with relevant knowledge in the subject matter at hand, GenAI is mostly some level of bad.)

AI discourse needs nuance. Following on from the above, Steve Klabnik wrote about the pros and cons of AI/LLMs and how he’s disappointed in the discourse. On Bluesky, Prof. Christina Pagel suggested LLMs have benefits and that public comms arguing against them mostly comes from AI-hating journos.

I’d say there are intertwined issues here. LLM creators oversell. Consumers and managers assume LLMs can do everything and replace humans when they ultimately need expertise at each end of the process – and for those people to know what good looks like. Much of the pushback I’m seeing about GenAI/LLMs is actually from educators rather than tech hacks like me, because students are using this tech as an alternative to thinking. But also, many creators are staring aghast at LLM output making it to public eyes and asking: why? (Answer: if an LLM can churn out, say, some text that’s better than what the prompt author could write themselves, it looks good to them. And right now, that seems to be what many folks are going with. Which means a massive WOMP WOMP for writers like me who actually care about quality output.)

What we need is more nuanced discussion. GenAI and LLMs can be beneficial. They can also be terrible. Often, they’re somewhere in between. But you get way more eyeballs online when you argue something is the Best Thing Ever™ or evil incarnate. As ever: sigh.

However, Nick Clegg discourse needs no nuance. Following on from the above and the ‘above above’, Nick Clegg can go fuck himself. As can the current Labour government when it comes to AI policy. Constant whining (the latest being from Clegg, but various Labour figures have said the same) that the AI industry cannot survive if it has to ask permission to use the content it ingests and recycles (or even be transparent about what content is used) is irksome beyond belief. If your business model depends on mass copyright infringement, maybe you don’t have a business model.

Again, I feel the need to point out that the UK’s copyright laws are so strict that it in almost all circumstances remains illegal for a consumer to rip a CD they’ve bought to their own computer, for entirely personal use. Yet every single thing you’ve ever created should be freely available for a handful of AI companies, because their business model depends on that? Pfft. Anyone else up for starting a company whose business model depends on giving away Nick Clegg’s stupid book for free? Something tells me he wouldn’t be so supportive of that.

Bluesky is dying. Apparently. There was a lot of THE SKY IS FALLING this past week, with people posting stats that show the site trending towards approx minus fifty billion users by next Friday. Reality: things there seem… fine? There still seem to be many folks happily nattering away, rather than slinking back to Musk’s bosom. But author Jarrett Walker’s post caught my eye, basically being an argument that he should stay on X because 40% of his audience wouldn’t follow him if he quit. My take: if I were still posting in the Nazi bar and half my followers said they’d scoot over to somewhere else, I’d take that as a win. But then I don’t care much about numbers. 

Just as well, TBH, given my actual numbers. Then again, I’ve always thought it’s about who is reading rather than how many people. This blog’s stats will never set the world on fire, and yet my writing here is the reason I’m writing for Stuff. (An ex-editor was – maybe still is – a reader.) Over on Bluesky and Mastodon, I’m not posting there in an attempt to amass an audience that can fill Wembley Stadium. I’m just parping out random thoughts and having nice chats with lovely folks. That’s enough for me. (Unless, you know, you’re an editor who’d like to employ me. Because I wouldn’t say no to that right now.)

The Phoenix Comic reached 700 issues. A phenomenal achievement in a tough market. If you’re in the UK and have/know a kid in the 7–12 range (or fancy reading fun comics yourself), there’s a zero-risk trial that nets you six issues for a quid. Barg.