Most bugs I see these days are not from bad AI generated code. They are from developers who don't fully understand what the AI wrote. Should we slow down and read the code more, or is this just the new normal?
I use AI to write code every day now. But I still spend more time reviewing it than I used to spend writing it myself. Is this actually faster, or are we just moving the work around?
Guess what?
I have entered the Top 1% Club of experts on Topmate.
This accomplishment is as much yours as it is mine. Your support has been pivotal throughout my journey.
Let's continue creating the impact on topmate.
Client scope creep, ranked:
1. Can you just add one more thing?
2. This should be quick for you, right?
3. We didn't change anything, it just stopped working
Unpopular opinion: most AI chatbot MVPs don't need a vector database.
A well-prompted GPT-4 call with context stuffed in the system prompt will outperform a half-baked RAG setup 90% of the time.
Save the infra for when you actually need it.
Two years off Twitter.
Spent it actually building things instead of tweeting about building things.
Back now. Let's see if anything changed. (It hasn't.)
@cleverqazi@PvsNPcode Same here @cleverqazi . I learned a lot from you in my early programming days. If someone can fix the Twitter algorithm problem, it’s probably you.
I almost shipped console.log("here") to production three times in one month.
That's what led to LogRadar. I didn't want to build a VS Code extension. I just didn't want to do that embarrassing thing again.
Link is in the first comment👇