… The system is a bit opaque, but the more complex an effect, and the more iterations it takes to get it right, the more tokens it costs. A simple fuzz might only cost you 20 tokens, but a granular looper with rhythmically synced glitches might cost you 500. …
… That seems like a hedge against the companies and their needs and also cost, right? The cost of tokens is still quite high. …
… Our transcribed model, for example, MAI-Transcribe-1.5 is literally the number one in the world. It’s the most cost-effective of any of the hyperscalers. It’s the highest on accuracy. Our image model is now number two. Our image editing model is number three right behind Google’s and OpenAI’s. …
… You could have a product that lost a little bit of money, but if you wait a year, the processing is cheaper, the costs are cheaper, and AI is progressing incredibly quickly. So anytime we use a new model, six months later, the cost of it is like a 10th of what it was. …