I code daily with AI - real programming tasks, professional, real work, read customers, I use below 3:
- codex 5.5 medium - best results less hand holding medium speed
- opus 4.8 max - mediocre with hand holding medium speed
- glm 5.2 max - mediocre with hand holding and super slow
- composer 2.5 - mediocre with hand holding and super fast
I use all, since i run mulitple coding in parallel.
disclosure - I use rexide which we created for all these agents to run in parallel with good visibility and feedback.
Coming from java spring and knowing some python looking around - I decided I would django for side projects and leave java and spring for my day job. I think it's more suitable for side projects.
IntelliJ is much slower than any other editor including Zed and VsCode it's much slower to open and navigate, much slower to work with, much slower, it's so slow! but the code completion, refactoring, code navigation, and debugging features and endless other smart features are incredible. For me, that extra intelligence and code awareness boost translates to way faster development overall, even if the IDE itself takes a bit longer to load or work with or consumes huge amount of memory. Sometimes the smarts outweigh the raw speed.
I hope the technical debt does not explode on me one day so i'll ignore it.
On last 10 years 250,000 missiles were prepared in the north to attack israel, 10,000 in gaza, Iraq, Syria, Yemen all are collecting weapons to attack israel + not to talk about Iran nuclearization. Iran clearly states it's going to destroy israel. So when you say hope this does not escalete how do you deal with all these missiles and soon to be nuclear missiles? just let israel vanish from the map?
I fully understand Israel is not our concern, but should we leave it to have a second holocaust?
Tactical 10x engineers - leave lot of mess and tech debt. Eventually this becomes 0.1X engineer.
0.5X strategic engineer will make the whole surrounding 2X engineers.
Call me old fashioned I just read the Holy Bible before opening intellij at morning for 2-5 minutes. And at the end of the day close intellij the 15min of Talmud.
Hey do you know if it's possible to use chat gpt while training it on proprietary data without sending any of this proprietary sensitive data over the network to openapi?