As releases slow down, it's time to think about what we got this year and where we are going. o3's search, agent vs model progress, and scaling's settling.
Yup I agree. o3-pro can reduce formatting errors but is pretty minor. Deep Research I only use in creative ways like "read all my blog posts on interconnects.ai and do XYZ"
Agree not enough talk about how good o3 is at search. Does at least hours work of serious work in mins.
I don’t use the deep research mode of o3 often as even the regular seems to be researchy enough. Have you found compelling uses for it?
Similarly o3-pro is often overkill. o3 regular just hits the sweet spot and surprised it’s not generating more buzz.
Yup I agree. o3-pro can reduce formatting errors but is pretty minor. Deep Research I only use in creative ways like "read all my blog posts on interconnects.ai and do XYZ"