- Tweets 77
- Following 73
- Followers 17
- Likes 1
Pinned Tweet
Everyone's arguing about which model is smartest. The more interesting question is what happens when it has admin rights. I post about AI agents meeting real IT and security.
9
opus vs sol threads are fun but nobody ships on a leaderboard they ship on whichever one finishes the job
1
1
21
The usage limit fight isn't really about limits. Once a model is good enough to leave running, people stop rationing it and every plan starts to look small. Good problem for the labs, expensive one too.
9
full disk access was always a big ask now the thing asking is an agent that never logs off
There's an increasingly high cost to convenience. engadget.com/2276186/apple-s…
17
made this little guy with dot and his pass state is happier about green tests than i am
Made with AI
14
every lab dropped a cheaper coding model this week > cost per finished task beats the leaderboard
12
devday threw a lot at once always on agents get the clicks chatgpt inside the tools people already use is what moves usage
24
most always on agents are just a chat window that never closes the useful ones have a job know when to stop and someone still owns the outcome
15
Everyone's worried about dots eating their usage. Efficient orchestration means Codex does more useful work, so you use it more, not less. More output and more spend, which means more revenue for OpenAI. My guess is dots become the core of the device they're building.
29
Tip, if you update to ChatGPT Pro Max, you get a usage reset, so if you are already on the prior $200 plan, use it to 0% prior to upgrading.
31
My bank already thought the $200 ChatGPT sub through apple was fraud, wait until they see the $500 recurring.
14
