Press enter to see results or esc to cancel.

The100: Who’s rich, model collapse and ‘quite good’ 

“Did the Beatles ever mention television in their songs?”

In an article that’s both worrying and funny, Nicholas Carr shares Gemini’s problems:

“First, with the confidence characteristic of chatbots, Gemini declares that, yes, indeed, the words “television” and “TV” appear in Beatles lyrics. It offers “Yellow Submarine” and “Here Comes the Sun” as prominent examples. That’s a baldfaced lie. Neither of those songs has either word in its lyrics. Neither has anything whatsoever to do with TV […] Google is now giving precedence in its search results to a chatbot that it knows is unreliable — that it knows spreads lies. That strikes me as being deeply unethical.”

Model Collapse

Related, what if AI starts to forget reality? While guest editing Strands of Genius, Pip Bingemann said:

“if future models train on AI-generated content instead of real human data, they start to “forget” reality. […] It’s like making a copy of a copy—each new version gets blurrier. Over time, AI models trained on AI-generated data will misrepresent reality, affecting everything from search results to creative writing. The solution? Keep people in the creative equation and use AI in a way to generate NEW and NOVEL ideas, not more of the same average.”

(Editor’s note: Enough of the AI chat now, Alistair.)

Getting close

I read with interest Khary Campbell’s piece ‘Your consumers are not in your conference room’. Unsurprisingly, I have some additional thoughts…

If you’re trying to understand people better, in-home visits are more effective than one-to-one video interviews. You get the full context. A broader picture of their life.

However, in-home visits can still have their problems. If some randomer came to your house, what are the chances that you’d tidy up, bin all out of date food and, most importantly, be completely aware that a stranger is watching you?

Either way, travel budgets aren’t growing. So one-to-one video interviews have become popular. But, what about that context? You don’t get behaviour, you get a conversation. It’s cheaper, but inferior (IMO).

If only there was a way to see behaviour, in a less intrusive way, and without anyone having to go anywhere (ahem 👀).

In the words of the Grail Knight from Indiana Jones ‘choose wisely’. Know what each method can and can’t do.

“Nobody thinks they’re rich”

Now then. This article on who’s rich is a belter. The data looks at the different scenarios in which someone is considered rich, what it takes to be rich and whether they consider themselves to be rich:

“For those with a household income of under £10,000 a year, the median to be rich was £75k; then, for the 11 earnings brackets between £10,000 a year and £80,000 a year, the median to be rich was steady at £100k – before moving up to £150k a year for those earning between £80k – £100k and £100k – £120k, and then, finally, those with a household income of over £120k a year had a median to be rich of £200k a year.”

Story is king

Steve Jobs made the case for getting the story right before doing anything else. He’s talking about films, but it applies to storytelling in any form:

“Walt Disney solved the problem decades ago and the way he solved it was to edit films before making them: you get your story team together and you do storyboards […] No amount of technology will turn a bad story into a good story.”

And finally…

Drew Struzan is a legendary movie poster illustrator and his story is fascinating.

A guide to what the Brits mean when they say certain things. I wonder how LLMs cope with this? (Editor’s note: STOP IT!)

Some great ideas in the comments for ways to replace the awkwardness of going round the room and introducing yourself. I particularly like the ‘tell us your name and something about your name’.