‚A large-language model is not a brain. Give it probability, and let it play. Don’t ask it for originality. That’s antithetical to its principles.‘

A LLM is not a brain. Give it probability, and let it play. Don't ask it for originality. That's antithetical to its principles
A LLM is not a brain. Give it probability, and let it play. Don't ask it for originality. That's antithetical to its principles

User generated quality content – especially in times of AI – is truly much more important than ever. That’s why SCRIBERS[HUB] provides top-tier freelance writers and journalists for any field, any subject and any media. In the Q&A series ‚Content-Buzzer‘ Sabine Fäth, founder of SCRIBERS[HUB], asks communication experts the following questions …

Matt Potter Author, journalist & broadcaster

What does content mean to you?

First of all, content is a terrible word. It’s become a placeholder for ’stuff‘ in so many contexts, that we’re in danger of forgetting what it really is. That it’s the chance to tell a decisive story – and one that could well change a life, a mind, or a day. It’s the chance to share a vital detail; to connect with fellow minds out there – minds completely remote from yours – and shape the way they think, react, or see something.
And if we think about it, that’s quite extraordinary. If you were to tell me that by putting certain words in a certain order sitting at home or in an office somewhere, you could produce a physiological reaction – crying, laughing, say – or change the neural pathways, behaviours, perceptions or opinions of someone over the other side of the world, it would sound like strange magic.
But that’s what content does. That’s its USP, if you like. And while a lot of the world spends money on ad campaigns, it’s content that gets into people’s hearts and minds, wins their attention for a while, and genuinely shapes hearts, minds, and behavioural patterns.

What was the worst piece of content you have ever come across?

Oh, this is one of those questions that demands categories, like the Oscars! I mean there is so much bad stuff out there – the AI/autogenerated gunk, the creepy ‚because-we-care‘ advertising-minded content that gets out there; the stuff that looks like someone created it by hammering on the keyboard with their head; the patronising corporate brand stuff. My god. But there can be only one winner, so…

It was when I was editing a wealth magazine some years ago. A very highly paid and distinguished writer who seemed to have exclusive access to some incredibly rarefied circles sent me a piece, right on the last and tightest minute of the deadline, which he’d been promising and delaying for weeks. It was supposed to be a series of interviews with people who had bought islands for themselves. In the whole 3,000 words of weird non-sequitur and jumbled sentences was not a single interview – although I couldn’t, for the longest time, tell whether there was any interview copy inside it or not, so bad was the prose. I’ll always remember the feeling of doom and creeping dread as my eyes took in opening line’s misspelled, incomprehensible word-salad:
‚But really, who could more of today’s gratin than Mick Jagger?‘
Spoiler: It didn’t improve. What’s more? This distinguished writer may not have had the connections he claimed. Bad writing should always be a warning. The only question is: What is it trying to warn you of?

How can AI help create good content that connects people?

How can a calculator make good decisions? How can a synthesiser make the best album you ever heard? How can a plough create a good farm?

Human beings have always had a really perplexing tendency to create a new technology and then ask if they should be competing with it, or if it can control us – instead of asking the right question, which over and over again has been: Instead of competing with this, how can we use it?

AI is no different – despite what the two lobbies on either side would have you believe. Between ‚This is AI and it is on the verge of becoming sentient and can do anything you can do!‘ and ‚This is useless but it is also Terminator 2 and do not touch it!‘ (again, arguments we have seen rehearsed since at least ancient Mesopotamia), there is an incredibly useful tool that we’re in danger of losing sight of if we let ourselves get distracted by all the yelling.
That tool can take live data streams and implant them into your storytelling. What data journalists dreamed of doing just a couple of years ago in static terms can now become a rolling, dynamic reality.
That tool can reformat and curate backlogs of content, reviving archival work and updating legacy content – that brands will have spent money on, and see as a loss – so that it keeps earning for them. Spend is redeemed. Archives are no longer dead storage, but thrilling, livable content ecosystems.
That tool can take inert matter and make it intuit customer intent and preference, and begin to turn broadcast into dialogue.

The key is to remember that it’s a large-language model (LLM) and not a brain. Give it probability, and let it play. Don’t ask it for originality. That’s antithetical to its principles. It would be like asking a Casio synthesiser to plough a field. Or a plough to make hit records.

What mixture of talents and skills should a good content team offer today?

You have to mix the ‚What if‘ with the ‚here’s how‘. Everyone thinks a content team should be execution-orientated. Get people in, chase SEO. But so what? Someone’s got used to writing meaningless garbage in a way that its words index correctly. The key to a successful content team is to have navigators and strategic thinkers out there asking, ‚What are the ideas we should be giving people? What are the new forms we want to introduce and why?‘ Not just, ‚What are the things people are already creating too much content on, and how do we add to the pile.‘

Complete the sentence: Content is always …

… just one step away from being the most important encounter with a new thought, idea, feeling or way of seeing that someone will have today. So let’s make it count, right?

 

 

Teile gerne diesen Beitrag:

Facebook
Twitter
LinkedIn
WhatsApp
Email
Print