We were seeing the same story being covered because the news pipeline always chose the most significant stories. So we had the Iran war covered 10 days in a row. Not ideal. Here's what changed, and why.
The most important story isn't always the right story for today. The old pipeline had no memory. Every morning it searched for the biggest national story, and every development in a long-running story looked like a fresh biggest story. The repetition check couldn't help, because a new development is legitimately different from yesterday's. We ended up with the same war ten mornings running, and nothing looking for the stories we weren't covering at all.
Knowing what story to lead with requires context. Each edition now keeps track of its storylines, so when news comes in it's either a development in a story we've already covered or a new storyline altogether. That distinction is what tells us something has been covered before, even when there's a fresh development. It also lets us judge how significant the development is against what we've already published, instead of treating every update as a new biggest story. With that in place, the pipeline gathers stories across politics, the economy, courts, science and everyday life, remembers what it has already covered, and holds an editorial meeting before writing.
What you hear
It leads with one story told properly, quick updates on any other stories that had developments, and introduces a new story you probably wouldn't have gone looking for. In practice that gives every briefing the same shape:
- One lead story, told properly.
- Quick updates on any other stories that had developments, including a one-line "no change" where that's the honest answer, so you're never lost dropping in mid-week.
- One story you probably wouldn't have gone looking for.
Tap any of them to dive deeper and Fylom generates a full episode on that story.
Where to find it
To find these, go to Discover and browse the news section, or listen on the web:
Twelve new voices
A male and a female voice for Australian, New Zealand, South African, British, Canadian and American accents:
| Accent | Voices |
|---|---|
| Australian | Lachlan, Isla |
| New Zealand | Hamish, Ruby |
| South African | Daniel, Nadia |
| British | Tom, Harriet |
| Canadian | Owen, Claire |
| American | Grant, Callie |
Local news is read in its own accent, and you can pick a voice for any episode from the voice selector. Each one has a short preview, so you can hear it before you commit an episode to it.
A tip: name the format
Fylom picks the format from what you ask for. If you want a particular one, say so:
- "deep dive on how close we actually are to AGI"
- "headlines digest on whats happening in tech this week"
- "how to guide for training for my first half marathon, im starting from zero"
It also does narrative stories, trend analysis and two-sides-of-a-debate episodes. The format shows on the plan before anything is generated, so if it picked the wrong one you can change the wording and try again.
Reply to any Fylom email, or write to support@fylom.com. Your messages shape what gets built next.