Ask HN: Will agentic AI grow to handle technology leadership responsibilities?

Posted by gengstrand 23 hours ago

I think it is now pretty clear what impact agentic AI has on coders. I am curious how technology leaders (e.g., software architects, staff engineers, senior developers, team leads) are changing how they work and what they are responsible for in organizations that have fully committed to agentic AI. Now that it is the AI that is writing the code...

Do we still need humans to manage tech debt?

Do we still need humans to review the AI output?

Do we still need humans to care about non-functional requirements such as scalability and fault tolerance?

Do we still need a human developer on call?

If you said “yes” to any of the above, then do you believe that it is only a matter of time until your answer would change to “no,” or do you believe that it would always be “yes, humans are needed.” Please elaborate on why.

Comments

Comment by idontwantthis 10 hours ago

I'm seriously trying, but I haven't seen it be all that helpful for complicated stuff. I had a bug where we were running out of database connections and cursor wanted me to create a wrapper class for the connection pool that would count connections and block if there were too many. Obviously that was dumb and I didn't do it. Turns out I needed to actually understand the problem and then I was able to fix it without too much trouble. On the other hand, I made a whole website in 2 hours without touching any of the code directly, but it was really simple and had no performance or security concerns to worry about.