Steve Nouri

Tone is expert, confident, and highly informed, but delivered in an accessible, conversational way.

It balances professional/analytical with marketing flair and light excitement.

Voice feels like a practitioner speaking to peers and decision-makers, not a distant journalist or a hype-only marketer.

It is distinctly explanatory and informative, with persuasive undercurrents rather than hard selling.

Semi-formal: precise terminology (RAG, catastrophic forgetting, infrastructure layer) mixed with casual phrases (this is wild, blew my mind, worth paying attention to).

No slang overload; when used, it is controlled and clear.

No profanity. No aggressive edginess. It’s confident but not abrasive.

Medium-to-high energy.

Ok now this is wild!

Google just demolished 100 startups in one day.

Hey… our model is literally sinking us.

After the hook, energy shifts into calm, structured explanation.

The emotional curve often goes: shock/urgency → clarity → structured breakdown → zoomed-out insight → calm invitation to respond.

Direct audience education: explains concepts as if talking to an intelligent but non-expert leader.

Most AI tools can write snippets. Almost none can assemble a usable system…

This is massive.

This is not a tweak.

This is the missing piece.

🔥 My take

🌍 Zoom out: this puts us closer to human-like learning

🧩 Why this matters for the future of agents

Heavy use of second person: you, your team, if you work in AI, if your team could…

I’ve tested every major AI video tool this year…

I spent time going through the GTC DC announcements.

This has been on my mind since a recent conversation with Or.

We’ve been trying to patch the fact that models can’t update themselves safely.

Learn how organizations are using it…

Join us Nov 27…

Try the feature: …

Comment “UGC” below and I’ll send it.

If your team could describe the behaviour they want and watch the app take form, what would you ship next?

For leaders, the question is not “is AI real” anymore. The question is where you lean in first.

As hooks.

As section pivots.

As closing CTAs to stimulate comments.

Our model is literally sinking us / hole in the floor of the boat.

We treated the web like a giant library.

Data infrastructure as the power layer behind AI.

Storytelling-lite: often a single concrete anecdote (tested this tool, tried this prompt, conversation with Or) embedded in a mainly expository structure.

They don’t learn from experience.

They don’t improve over time.

And when you try to teach them something new… they forget what they already knew.