@0xSero I understand your frustration, but I have the exact opposite experience with Claude in general😅 the only thing I have to correct once in a while for some reason, is telling it again to use the tmux session it’s supposed to use
@danveloper I suppose you haven’t read the thread then :-) I have had a ton of success either it, see my comment. Another thing it did, I told it to make a reservation at a restaurant, it was fully booked, but it then located the phone number of the chef,and he got me a table :-)
@helloiamleonie It’s managing my home as a voice-butler: controlling, heat,lights, air quality through venting,knows when I’m home to schedule cleaning with my robots,plan solar charging/battery mgmt, does shopping lists,infer meal plans from photos of my fridge(and shopping lists).a lot more!
@0xSero Great you found out😅A tip maybe, if you’re not already doing it - name a tmux session and have Claude code use that (he will use tmux send-keys and capture pane), it’s like pair programming/developing with Claude :-)
@herrmann1001@rfidiot Nice! Looking forward to an update. I just had my team of AI agents autonomously suggest that skill, they just needed me to buy the hardware😅
@bridgemindai Try telling Alex Finn about it 😅 jokes aside - I agree with you. Reason for local models could be privacy (personal or business). If you can iterate on a local model with expensive hardware, it could get useful. Have you tried running the biggest Qwen model yet?
Soooo could we soon have the option to disable output indention in Claude Code? @ClaudeDevs pretty please! multiple bug reports are filed - I stopped counting after finding about 10... :) https://t.co/nXJPkXumXf for one. TIA!!