I want cheaper LLM models and millions of token contexts. Please AI researchers, let us scale the gas/sec while you scale the tokens/ context { and do it cheap! } Please .....
335