What's the most important question right now?

I’ll focus on the editorial strategy for any model that relies on large audience, where I believe AI is the central question. And within that, the real question is: what do we make that remains valuable when answers are free and content is abundant?

AI has changed the game in two directions at once. First, platforms are increasingly offering AI-generated answers that satisfy intent without the click. Second, the supply of “good enough” written content has exploded, as new players use AI to produce volume at scale – sometimes that’s responsibly augmented, sometimes it’s irresponsibly fully automated.

That matters because the traditional publisher moats we’ve relied on – originality, authority, voice – are still present, but they don’t automatically translate into demand. Editors have always known, slightly painfully, that audiences often reward immediacy and drama more reliably than depth and complexity.

Subscription-first models are one logical response, because they reduce dependency on search and social. But they also force a different kind of editorial clarity: what is so distinct, useful, or identity-relevant that someone will pay for it and build a habit around it?

In practice, I’ve seen the uncomfortable reality of that at several of the titles I’ve worked on. We could land exceptional exclusives and in-depth interviews, and still not get the conversion or engagement uplift we expected. Meanwhile, certain sub-niches – tightly defined beats with clear utility and community – produced much stronger subscription growth than “important journalism” in the abstract. Audiences don’t always engage with the content that we want them to want.

For editors and publishers optimised around advertising and reach, this shift in incentives is hard. It requires a willingness to pursue new content shapes, new audience contracts, and new success metrics. When the levers have traditionally been output, audience through search and advertising yield, those are really hard transitions to make.

Are we taking AI seriously enough?

Certainly not at the operational level. Reuters Institute research in late 2025 found that around four in ten journalists said AI hadn’t been integrated into their newsroom’s processes at all, and where it had, it was often described as limited.

Even setting aside AI generating any copy, much of publishing is still weighed down by manual production work: CMS formatting, headlines and cross-heads, metadata, internal linking, image selection, and distribution across various channels. That’s exactly the kind of mundane work AI excels at. In our own workflow at Velora, we’ve automated the entire path from raw copy to a finished page – including structure, metadata, links, image selection and social sharing templates. It’s striking how few media businesses seem to be treating writer workflow automation as a major priority.

When the conversation shifts from workflow to writing, the debate often becomes moral rather than practical: focussed on keeping journalism “sacred,” keeping it human. The principle is completely correct – responsibility and agency can’t be outsourced to machines

But it’s also true that many steps around writing don’t threaten authorship at all. AI can help with tasks like tracking multiple news sources, initial research support, transcription, source summarising and quote, source and fact verification when using the right tools in the correct framework.

The industry is starting to move in that direction in places. In January 2026, News Corp announced a partnership with Symbolic.ai to roll out an AI platform across its newsrooms to augment research and writing.

My view is that the biggest AI risk is that in a time of inconceivably fast technological progression, publishers keep running 2020 workflows in a 2026 where cost and distribution have completely changed.

Any following recommendations?

Alberta Tech on TikTok. She’s a software developer and she is ridiculously funny. As an amateur vibe coder, I find her experienced engineer takes are equal parts humbling and hilarious. She’s great on AI in the real world: enterprise adoption, integration pain, and the pure comedy of the AI hype machine.