Page 1 • Showing 4 tweets
opus 4.5 is the model for sync work gpt 5.2 is the one for async
just a matter of time until someone ships an mcp server for ads - add to your agent and sign in with oauth - randomly get served ads in chat - you get paid for every exposure
turns out, senior engineers accept more agent output than juniors. this is because: - they write higher-signal prompts with tighter spec and minimal ambiguity - they decompose work into agent-compatible units - they have stronger priors for correctness, making review faster and more accurate - juniors generate plenty but lack the verification heuristics to confidently greenlight output shows that coding agents amplify existing engineering skill, not replace it
this meeting could have been a prompt