If you like hardcore, you might know there’s a review of the 1998 album The Shape of Punk to Come on the fan site RateYourMusic that says: “Saw this in a shop with a ‘this rocks’ sticker, so I took it to the counter, slammed it down, and shouted, ‘how much?’. When the dude answered, I said, ‘not the price dillweed; how much does it rock?’” It goes on; you get the idea (search ozzystylez on RYM). It is long, sweary, and funny – because it skewers music snobbery, laughs at itself, and captures the spirit of the record. It is known by fans of the band, and on certain subreddits.

There is a point, here. Often the best criticism is not from paid critics. A couple of the stories below are better summarised elsewhere, and don’t need adding to, including about this Nvidia/OpenAI circular-trading fallout, where Nvidia said its $100 billion deal with OpenAI in September was never really a deal, and OpenAI responded that it doesn’t rate Nvidia chips anyway. “Imaginary money impacts imaginary products to create imaginary services for imaginary customers,” said one commentator, wondering about the impact to “imaginary economies”.

The linked article in The Guardian about the same story quotes Forrester Research, that “you don’t know what’s going to happen – and so you let other people put numbers out there and let that drive the hype”. Another commentator responded: “Just a jaw-dropping admission that it’s all been a shady exercise in subterfuge.” LinkedIn is filled with good discussions like this at the moment – during moments, like now, when big-tech news gets madder, and invites a direct human response, rather than when your feed is filled with AI slop. 

There was some good back-and-forth last week as well between AI cheerleaders and AI killjoys after Anthropic chief Dario Amodei said at the launch of new AI agents that AI agents like his will create mass unemployment at software companies – and SaaS stocks plummeted, software engineers lashed out, and most analysts backed the money-side. Today’s story about Starlink’s gold medal in Ookla’s satellite sprints also brought some interesting debate – sort of. “Fun fact: Starlink has lower latency than fiber for intercontinental distances,” said someone else on LinkedIn.

The logic goes that “light travels faster through a vacuum, hopping between satellites before returning to Earth”. San Francisco to Mumbai in 250ms on terrestrial fiber, versus 120ms on Starlink – is the sum. To which Roderick Beck, always good, scolded: “Grasping at straws. Fiber’s advantages over LEO satellites are enormous in terms of capacity, reliability, and cost per bit.” Which perhaps summarises the difference between the state of tech we are in, as presented by engineers, and the shape of tech to come, as presented by billionaire hype merchants.