Nobody stars the news anymore

For a while, most mornings started the same way: six browser tabs of games press, three platform blogs, and whichever trailer channel had posted something overnight, trying to work out what had actually happened since yesterday and whether Playtopia’s readers needed to hear about it. A good third of it was the same PlayStation Blog post wearing three different headlines. By the time anything got chosen, written and put through Postiz, it was mid-morning and the story had already made the rounds twice.

So I built a collector instead. Every hour, a cron job reads a fixed list of official feeds and press sources, works out what’s actually new and what’s just the same story told twice, scores it, and from the morning lays the day’s worth of it out as a numbered card in the calendar. Plain code, no AI anywhere near it yet.

What clears the gate

Why it is here. Not everything a feed hands back deserves a line in the roundup. A forum post about a homebrew release and a PlayStation Blog post about a major sequel shouldn’t carry the same weight just because both came back from the same hourly fetch.

How it works. Each source carries a fixed weight: forty for Epic’s free-games feed, thirty-five for the platform holders’ own blogs, down to twelve for a forum RSS feed that mixes news with homebrew releases. Press outlets sit under the official feeds on purpose: a single games-press story about something official already covered sits just under the thirty-point cutoff on its own, and only clears it if a second outlet picked it up too, or it’s a free game, a sale or a trailer in its own right. The same story landing from two feeds gets deduped into one item rather than showing up twice, with the second source carried through as “also reported by” instead of a line of its own.

The card that used to need a hand on it

Why it is here. A scored list isn’t the same as a published roundup. Something still had to look at each day’s items, drop the ones not worth Playtopia’s time, mark anything big enough to deserve a post of its own, and say go.

How it works, or did. From when this was built, the card was a watch nobody but me stood. Every morning I’d open it, delete what wasn’t worth a reader’s time, put a star against anything that earned its own Facebook post on top of the roundup, and tag the card Write. An agent run within the hour would write up what was left: one credited article grouping the day’s free games, deals, headlines and trailers, a Facebook post linking it, and a separate post for each star.

That’s the part that’s gone. The card still arrives every morning, but it arrives already tagged Write, and nothing on it gets trimmed before it’s turned into a roundup. The star is still a field in the data, still something the tool would honour if a story carried one. On every card I can see since the switch, the star list comes back empty.

What takes the place of judgement

Why it is here. If nothing upstream is deciding what’s worth covering anymore, the only job left between the collector and the publish button is getting every fact right, because there’s no second look coming after it.

How it works. Each story in the roundup gets written from its own source’s page, not the feed summary, and ends in a line that names where it came from and links it, every time, including every outlet that also carried it. A date, a price or a platform the source doesn’t actually say doesn’t get softened into something vague, it gets left out and flagged instead. If a date slips through wrong now, nothing catches it before the public page does. You’re killing me, Smalls, is the fair review for that one.

Verified: the picks card now arrives pre-tagged Write, the run never trims it, and the roundup gets scheduled the moment it’s written, straight from the config flag and the run’s own instructions. Assumed: that the star genuinely hasn’t been needed, not just that nobody’s got around to using it. The field’s still live in the code either way, so pressing it tomorrow would still do something.

Fair winds.