Open weights are the real deal for Brown builders

Martin Alderson is mapping out the shift from closed models to open weights — the kind of models you can actually run yourself on a single GPU. The pattern is simple: instead of sending prompts to a company's API and hoping they don't change, you pull the weights, load them into an inference engine, and run them on your own hardware. That changes who controls the stack.

The models in question are the ones being released by Meta, Mistral, Google, and others with full license permits — not the API-only ones. They're getting competitive with the big names while running on consumer-grade GPUs. For folks who've been paying per-token to APIs, the math is starting to tip. For small shops and tías running WhatsApp groups, it means you can host a model for $50 a month instead of $500.

Alderson's post lays out the tooling — vLLM, Ollama, LM Studio — and the rough economics of what fits where. The real story isn't the models themselves; it's the decoupling. When weights are open, the vendor lock-in breaks. You can swap models without rewriting your app. That's what makes this summer worth paying attention to, not just the benchmarks.

Why this matters for us: open weights means we stop renting our tools from companies that can change terms overnight and start owning the stack we depend on for real work.