It's my daily driver, it's fast affordable and with a bit of guidance gets the job done.
I only reach for Claud when i need to plan something big or want to have a sparring partner to fire of some ideas.
I think what a lot of people don't realize is that you don't need a fronteer model for 80% of coding tasks. Composer 2.5 is often more than good enough, less token hungry and way faster
When you normalise for time and money, Composer 2.5 is way, way, way, way better than anything else out there. Yes it requires more babysitting, but that's a good thing.
This resonates a lot with me. I often use AI for the plan and let it propose multiple possible implementations, I often have to point out the glaring easier / logical solution.
When implementing its often a lot of misses with a few golden hits. The other day it used flex for a table layout while our app uses tables everywhere sigh.
Another typical one is that it tends to prefere frontend aggregation and looping of data instead of letting the database and backend deal with it.
I find Cursor/Composer is really good at mimicking existing code when writing new code. And it will often do so without being asked, but I try to always explicitly mention an existing bit of code for it to read as inspiration (e.g. "use the TPS 2.0 report as a style guide").
There's plenty of reasons to set a unique identifier before database save, or to want a unique identifier that doesn't have a 1-to-1 relationship with your object.
For example, in the idempotent kafka consumer pattern we set a unique ID in the header of every kafka message at the time of message publishing. We then have our consumers do a quick check of the ID against their data store to see if they have processed the message before or not. This way there is no impact if a consumer sees the same message twice. This allows us more flexibility during rebalancing events or replaying old offsets.
Interesting. I'm in EU and see these constantly but usually in the afternoon so it bothers me less as I'm already wrapping up, but my US coworkers are getting hit much worse.
I also bet my money on Azure. Someone who allegedly worked there recently posted an article here on the numerous problems with Azure. Sadly I didn’t bookmark it.
I can only say: learn how to use reduce and you never loop twice through a list of items or objects.
To me reduce is very easy to reason about and makes it super easy to properly filter, combine, extract values without ending with filters on filters on maps and maps
Doing everything in one reduce step sounds to me like the opposite of "easy to reason about". Reduce is a powerful tool that everyone should know, but you don't always want to wield the most powerful tool, especially if you're after intermediate values like the OP is.
1) obvious green field project
2) well defined spec which will definitely be in the training data
3) an end result which lands you 90% from the finish
Now comes the hard part, the last 10%. Still not impressed here. Since fixing issues in the end was impossible without introducing bugs I have doubts about quality
I'm glad they do call it out in the end. That's fair
We went from barely able to ask these things to write a function to writes a compiler that actually kind of works in under a year. But sure, keep moving the goal posts!
Many people are convinced that we’re all going to die next year after these things achieve sentience. Can’t wait to see the goalpost shifting when AI 2027 doesn’t pan out.
IMO the reset should fire regardless what happens after it.