Tuesday 27 December, 2022

2 x 8 Star Street

Quote of the Day

”Cricket is full of theorists who can ruin your game in no time.”

  • Ian Botham

Musical alternative to the morning’s radio news

Diana Krall | Narrow Daylight


Selection prompted by the slowly-lengthening daylight hours.

Long Read of the Day

Tech Journalism Doesn’t Know What to Do With Mastodon

Perceptive rant by George Dillard.

I was prompted to write this post when I ran across the following post on Mastodon by Annalee Newitz:

Gotta love tech journalists who describe Mastodon as “that impossible-to-use website.” First of all, it’s an app. C’mon. Second of all, aren’t these the same people who write breathless explainers about the wonder of cryptocurrencies, which are not only impossible to understand but literally built from bullshit?

Like Newitz, I’m an increasingly enthusiastic adopter of Mastodon, and, like them, I’ve been kind of confused by the press coverage around the platform. The media seems to be regarding Mastodon as a bizarre curiosity, something that the general public couldn’t possibly grasp. Sure, the guys with a Linux server in their basement might geek out on it, but this thing isn’t for the masses…

The problem — as Dillard acutely observes — is that Mastodon doesn’t fit the standard tech narratives.

He’s right. Do read it.

Reflections on Generative ‘AI’: #1

(A thoughtstream on a current obsession.)

2023 looks like being more like 1993 than any other year in recent history. In Spring of that year Marc Andreessen and Eric Bina released Mosaic, the first modern Web browser and suddenly the non-technical world understood what this strange ‘Internet’ thing was for.

We’ve now reached a similar inflection point with something called ‘AI’, which is really the tech industry’s name for an arcane technology called machine-learning. Until now, most people hadn’t a clue what it was about, or indeed what it was for (except perhaps automating jobs). But in 2022 a new variant of this ‘AI’ arrived. It’s called ‘Generative AI’ — machine-learning systems that can ‘generate’ plausible artefacts. Midjourney, for example, can create interesting and/or amusing graphics in response to text prompts like “Draw a picture of J.K Rowling as an astronaut”.

Up to now, most people have regarded them as interesting toys (though graphics artists fear them as threats to their jobs). But the first killer-app of Generative AI has just arrived in the form of ChatGPT, a system that can often (though not always) generate plausible text in response to a prompt. It’s become wildly popular almost overnight — going from zero to a million users in five days. Why? Because everyone can intuitively get that it can do something that they feel is useful but personally find difficult to do themselves. Which means that — finally — they understand what this ‘AI’ thing is for.

Well, maybe they do. But my guess is that they don’t understand what it might mean. Hence this series…

More tomorrow.

My commonplace booklet

What we don’t say in Silicon Valley anymore

From the blog of Om Malik (Whom God Preserve)…

In my time writing about Silicon Valley, it has gone from being a place of naive curiosity to a place where posturing is everything. And the reason we have this state of affairs is that, with extreme success, the denizens of the valley have ostracized these four phrases from their vocabulary:

  • I’m sorry
  • I don’t know
  • I was wrong
  • I need help.

This Blog is also available as a daily email. If you think that might suit you better, why not subscribe? One email a day, Monday through Friday, delivered to your inbox. It’s free, and you can always unsubscribe if you conclude your inbox is full enough already!