Every time you send a prompt to GPT-4o, you are paying frontier prices regardless of whether the question is hard or trivial. Switchboard, a Solana-based inference exchange, thinks that is absurd. It runs a live auction across 37 models on every single request, routes your prompt to the cheapest model that can actually handle it, … Read more
In 1999, a nineteen-year-old college student released a protocol that let strangers share files without asking anyone’s permission. Napster had a central index, and the law killed it. BitTorrent had no index, and it’s still running. The lesson was simple: protocols outlast products because nobody can shut down what nobody controls. AI inference is having … Read more
What if buying compute was more like trading stocks than subscribing to SaaS? Surplus Intelligence is building exactly that: a decentralized exchange where sellers compete on price, buyers get the cheapest rate, and every transaction settles on-chain. The Problem With Fixed-Price Inference AI inference pricing is broken in both directions. Providers charge flat per-token rates … Read more
When inference becomes a commodity, someone has to build the market. UsePod is doing it on Solana with a doxxed founder, and a very specific opinion about what comes first. The Commodity Thesis AI inference is getting cheaper fast. Open-weight models like Llama 4, Qwen3.6, DeepSeek-V4, and GLM-5.2 now trail proprietary frontier by roughly three … Read more