The legacy news business is a pipeline: a newsroom makes content and pushes it at a passive audience. That model is dying, not because journalism stopped mattering but because the economics that paid for it, bundled classifieds, captive local ads, monopoly distribution, got dismantled by the internet. Meanwhile everyone migrated to X, Reddit, and Instagram for breaking news and received, in exchange for speed, an information diet that is chaotic, algorithmically warped, and credibility-poor.
The platform move is to invert the firm. Instead of employing a thousand journalists, orchestrate a decentralized network of verified experts. AI aggregates the primary sources, videos, filings, posts, and incentivized top-tier analysts supply the context and curation raw information lacks. You become the matchmaker between fragmented attention and expert curation, monetized through high-trust, personalized feeds. Structurally, the pipelines-to-platforms logic is sound.
The objection that has quietly killed every prior attempt, and there have been many, is that trust does not decentralize cleanly. The old editorial desk, for all its mockable flaws, did something that is easy to sneer at and genuinely hard to replace: it was an accountable, libel-liable, reputation-staking filter that took responsibility for what got published. Distribute that across a network of independent experts and accountability thins until it evaporates. When a networked expert gets something catastrophically wrong, who is liable, and whose reputation eats the damage? If the honest answer is "no one in particular," you have rebuilt the exact credibility problem you set out to solve, now with a nicer interface and a Series A.
There's a second landmine that two-sided market theory predicts and the vision ignores. Platforms optimize for engagement, because engagement is what they can measure and sell. Engagement and truth are different signals, and wherever they diverge, the economics pull toward engagement. This is precisely how the existing social platforms became credibility-poor, not for lack of good contributors but because the incentive structure paid out for the inflammatory over the accurate. Build a new platform on the same attention-monetization engine and it will drift to the same swamp, however noble the founding manifesto.
The version that could actually work has to fix both structurally, not aspirationally. Stake the platform's own reputation on its curation, accepting liability rather than distributing it into vapour. And monetize through subscription rather than engagement, aligning revenue with trust instead of with outrage. Both choices make the business slower-growing and lower-margin than the pure attention-platform fantasy. They are also the only version that doesn't decay into the thing it was built to replace. The platform model for news is right. The engagement-monetized, accountability-distributed version of it is just a more sophisticated way to rebuild the problem.