| ▲ | eddd-ddde 5 hours ago | |
Your macros are not implementing features, fixing bugs, and writing test cases. | ||
| ▲ | orev 17 minutes ago | parent | next [-] | |
Which is fine. Burn your tokens (with all the carbon emissions they generate in data centers) doing more involved tasks like that. Don’t waste it on silly, easy things like simple macros, search/replace, etc. AI doesn’t need to be a hammer used for everything. | ||
| ▲ | cryo32 4 hours ago | parent | prev | next [-] | |
I literally just watched someone do a text replace on a conference call to show me how good Claude was today. | ||
| ▲ | SirFatty 3 hours ago | parent | prev | next [-] | |
Said with such confidence... | ||
| ▲ | runarberg 3 hours ago | parent | prev [-] | |
They are. And have been since the 1970s, and will continue to do so for another 50 years. Particularly the fixing bugs parts. So often is one bug manifested across 200 lines in 50 different files, and to fix it you need the same 5 exact keystrokes across those 200 lines. vim (or in my case Emacs) macros will allow me to find the fix in the first of these lines where I also define the macro, and then apply in mere seconds across the other 200 lines. As for writing test cases, copy-paste has worked fine since unit tests became popular in the 1990s, and will continue to work fine for the next 30 years. editor macros are not too dissimilar to copy-paste-edit. | ||