Agents are causing token usage and costs to explode because a single task can require many model inference calls. That's bringing back the old build versus buy question, but in a new form.
Why listen
It goes beyond the title with direct discussion of like, that's, right, including: At the same time, open-weight models are getting good enough to handle more and more of that work.
Key takeaways
01Agents are causing token usage and costs to explode because a single task can require many model inference calls
02At the same time, open-weight models are getting good enough to handle more and more of that work
03That's bringing back the old build versus buy question, but in a new form
Best for
research-minded practitioners comparing model behavior