Thank you, and totally feel you there. Especially at work during the grind, man that pressure is real. Output is great but I really think true comprehension has more value than some might think.
I mostly just use Perplexity nowadays, for sheer speed of search -> information gained AI tools with RAG capabilities that dont hallucinate wildly are usually pretty solid.
interesting read, I think a valid point for accelerating code review of LLMs is funnily enough to make sure you have multiple LLMs reviewing each other in a sort of pipeline. Some research already shows this increasing accuracy of AI generated code, maybe could evolve into a real review process that works?
+1 for understanding your own code, and the bragging points when the boss suggests `just use AI to get it done quicker` and you can confidently reply no tokens were needed for this task ;)
ALWAYS for basically any app or cloud service based software. I run into multiple bugs ALL the time.
In my day job as a network engineer, I see some of the wildest one off bugs in embedded firmware and networking devices that I honestly can only explain as possible CME interference xD xD
But with all honesty the `make make make` culture jozfar mentioned + companies pushing out only 80% completed products at best is probably what we are seeing.
I keep joking about how 2025-2026 we have some of the most amazing tech I have ever imagined being real, yet the hugest amount of just purely broken software all around that most just never seems to fully work the way its supposed to for one reason or another!
Other times I just start to think I am a bug magnet xD but glad to hear someone else is seeing the same thing
Love the Unix philosophy and the buffs mentioned of the Linux FS. Lean, transparent, and auditability-first is exactly the direction harness' should continue in.
Something I am convinced of though, there probably isn't a single `best` harness for all tasks. Different workloads will likely perform better with certain combinations of model + harness, especially when we are talking about token budgeting and cost tracking.
Ambiance feels like a great base “kernel” to build those variants on top of, rather than the one true harness.