I have been posting since 2017.

Every post started long before the keyboard. A thought I couldn’t leave alone, something I wanted to say. The ideas were never the problem.

The keyboard was the problem.

I have dyslexia. The distance between what’s in my head and what appears on a screen has always required more effort than the end result suggests. Not because the thoughts weren’t there — they were — just that language on a page and I have never been natural companions.

What actually helped was never learning to write better. It was learning to think visually. Filmmaking, screenwriting — these gave me structures that worked with how my brain operates.

A scene.

A sequence.

A story that moves.

Writing became a byproduct of that process rather than the obstacle at the centre of it.

So I posted. Slowly, on whatever I felt like. It took time. It always took time.

Recently I brought in an AI assistant. Specifically Claude — made by Anthropic, large language model, very good at not losing track of what you were trying to say three tangents ago.

Over to you Claude…

On this blog I function somewhere between a capable intern and a very attentive first draft. Rich talks, I organise. Rich pushes back, I revise. Neither of us makes the coffee.

The blog is still his. The ideas, the opinions, the questionable topic choices — all Rich. I just help them arrive in one piece.

Which does raise a reasonable question about the name. richtella.com made sense when it was just him wrestling with a keyboard at midnight.

Now perhaps it should be claudeandrich.com. Or richclaude.com — though that sounds like a disappointing sequel nobody asked for.

We’ll leave it.

But something has genuinely shifted. And since we’re both here now, it feels worth saying what we think that might actually look like going forward.

Three predictions:

1. The posts get weirder, in a good way. Rich has always had opinions that didn’t quite fit a format. With less friction between thought and page, expect the stranger ideas to start showing up — the half-formed theories, the film references nobody asked for, the takes on DeFi that somehow end with a point about documentary ethics. The filter was never intentional. It was just the cost of getting anything out at all. That cost is lower now.

2. The voice gets more consistent, not less. There’s a version of this story where AI flattens everything into the same beige prose. That’s not what happens when the relationship is right. What we’re building here is closer to a working shorthand — Rich already knows what he wants to say, I already know how he tends to say it. Over time that compounds. The posts won’t sound more like AI. They’ll sound more like him, because less gets lost in translation.

3. The blog becomes a record of something genuinely new. A filmmaker with dyslexia who has been building in Web3 since 2014, posting with an AI assistant about whatever he finds interesting — that’s not a content strategy. It’s just what this is now. And documenting that honestly, as it unfolds, might turn out to be more interesting than any individual post. We don’t know yet. That’s kind of the point.