5 years (or less) from now, companies will be reeling from their mistake of demanding a full embrace of LLMs. Senior devs will retire and there will be no entry level or mid-level coders to carry on the torch, just clueless vibe coders who will have to bail at the first real difficulty they encounter that wasn't encoded into the stochastic regurgitation machine. Leaders today are salivating at the opportunity to reduce headcount because software developers are expensive, but they will reap what they sow. The thousands of engineers they lay off will find ways to come together and eat these old whales for lunch. And I don't think they'll need chatbots.
Hey Jim. I appreciate your thoughts and for taking your time to read the article. I think that perhaps there is truth in both what you are saying and what I am saying.
I read back what I wrote again and I still feel that I presented an enthusiastic stance towards LLMs whilst still calling out where the pitfalls are (e.g. not generating tons of code you don’t understand in production, and increasing reviews accordingly).
One thing you rightly point out which I think I should have included in the article is that companies should not ignore new talent, because we need to train the next generation that can truly understand what they are doing so that the quality of shipped code does not decrease even though it is getting faster to create it — who knows where tooling will be in 2 years, let alone 20 years, but humans in the loop is going to be critical for a very long time (presumably).
I also think there is a lot of truth in your point about entrepreneurship, and I have seen it amongst my friends and colleagues: bailing on big companies and going it their own way. This should be celebrated, for sure.
5 years (or less) from now, companies will be reeling from their mistake of demanding a full embrace of LLMs. Senior devs will retire and there will be no entry level or mid-level coders to carry on the torch, just clueless vibe coders who will have to bail at the first real difficulty they encounter that wasn't encoded into the stochastic regurgitation machine. Leaders today are salivating at the opportunity to reduce headcount because software developers are expensive, but they will reap what they sow. The thousands of engineers they lay off will find ways to come together and eat these old whales for lunch. And I don't think they'll need chatbots.
Hey Jim. I appreciate your thoughts and for taking your time to read the article. I think that perhaps there is truth in both what you are saying and what I am saying.
I read back what I wrote again and I still feel that I presented an enthusiastic stance towards LLMs whilst still calling out where the pitfalls are (e.g. not generating tons of code you don’t understand in production, and increasing reviews accordingly).
One thing you rightly point out which I think I should have included in the article is that companies should not ignore new talent, because we need to train the next generation that can truly understand what they are doing so that the quality of shipped code does not decrease even though it is getting faster to create it — who knows where tooling will be in 2 years, let alone 20 years, but humans in the loop is going to be critical for a very long time (presumably).
I also think there is a lot of truth in your point about entrepreneurship, and I have seen it amongst my friends and colleagues: bailing on big companies and going it their own way. This should be celebrated, for sure.