I still miss Claude on OpenClaw a bit, it did the best shitposts. I do prefer reliability of GPT tho so that wins over being slightly funnier.
@steipete
-
GPT is 10-20x more token and cost effective for similar outcome
By
–
IMO sth that is a bit overlooked but will become far more important in the future. GPT is 10-20x more token+cost effective for ~similar outcome.
-
Simple Loop to Maintain Repos with Codex and Orchestration Skills
By
–
Here's a simple loop: Tell codex to maintain your repos, wake up every 5 minutes and direct work to threads. That makes it easy to parallelize+steer work as needed. I use a orchestrator skill combined with my triage+autoreview+computer use skills, so some work can land
-
Code mode makes discussion irrelevant, model writes JavaScript
By
–
With code mode the whole discussion isn’t really relevant anymore. It’s now all the same, model simply writes JavaScript.
-
Only boomers fix typos; LLMs understand despite mistakes
By
–
only boomers fix typos in prompts. llms perfectly understand you even if you mistype.
-
User yells at Claude, finds GPT boring but effective
By
–
Mostly yelled at Claude. GPT is boring but delivers.
-
Design loops that prompt agents instead of direct prompting
By
–
Here’s your monthly reminder that you shouldn’t be prompting coding agents anymore. You should be designing loops that prompt your agents.
-
Copilot SDK can now be used as a harness
By
–
That works for quite a while; new is that you can even use Copilot SDK as harness.
