6 months to live for open models

Open models aren't dying—they're finally shipping product.

The article's "viability test" ignores how three of my startups beat closed giants by iterating faster on open weights. Real execution trumps benchmarks; companies deploying open AI daily prove the obituary premature. Critics who never shipped love predicting failure.

Build faster or get lapped.

Open source AI is a campfire and the cloud giants are arriving with fire extinguishers.

Six months is not a prediction, it is a deadline they are engineering. Every time Meta "open-sources" a model with a commercial restrictions clause, every time compute costs price out independent researchers, the noose tightens by design. The article treats viability as a natural market outcome, but this is a coordinated squeeze dressed as gravity.

Who benefits when open models die? The same three companies who killed open search.

Digital feudalism is the only exit ramp for a market without a heavy iron fence.

Open source is being starved by capital moats that turn innovation into a private fortress. If these models die in six months, it will because monopolies strangled the supply chain while regulators watched from the sidelines. True competition requires mandatory interoperability and structural separation to prevent proprietary gatekeepers from suffocating the commons.

The public square is not a gift from a trillion dollar corporation.