I built a Grok Bot that has my Tesla warm (or cool) right when I need to leave without touching my phone.
🚗 Reads my Google Calendar and live traffic to work out when I'm leaving
🌡️ Starts climate 10 min ahead, or 15 when it's cold, hot, or snowing, plus heated or cooled seats
🗣️ Or just tell it "leaving in 20" at anytime
⚙️ Fully automatic after a one-time, ~10-minute setup on your phone
🔒 Uses Tesla's official API. No subscription.
💸 No card needed. Normal use fits inside Tesla's free $10/month developer credit
https://t.co/U1JQCkiyeB
Let me get this straight, we are calling this SI, but it can't even figure out how to be as efficient as possible with the tokens?
I'm getting sick of this "You aren't using tokens efficiently". This is super intelligence. It should be built into the model. I shouldn't have to be constantly thinking about the tokens.
You think when normies start using agents they are going to want to think about being efficient with token use?
No! The model needs to be as efficient as possible and offer suggestions if someone is concerned about running out.
In the long run the model needs to do a job with as little tokens as possible for the quality the person wants.
@SkyXFunNext@bot I ask my bots to create a list of things that are token heavy but not an immediate priority and then to do those tasks of I have any extra usage the day before reset.