Thanks a lot Addy for publishing this post this week. Your post brings hope to software engineers, perfect for this time of the year ☺️.
Do you have any recommendations for improving on domain expertise for frontend?
I also saw posts saying frontend engineers won’t be needed 🤔 and will evolve to full stack or product engineers. I guess full stack will be needed mostly for startups as always and since there might be more startups thanks to AI it might be more demanded than before. What do you think? Should we still specialize in depth domain knowledge of only one software engineering role?
I have had this dilemma on over relying on AI but reading this helps clarifies everything and gives valuable knowledge in leveraging AI in my career effectively.
Maybe I'm just a boomer but I would advocate for juniors to spend some time with their AI assistants switched off in order for them to get to develop some confidence in their own learned and innate abilities.
We end up agreeing on most everything (and I was untainted by your post, which adds believability), but there are a few small interesting differences:
* I didn't consider the operations upskill (#4). That's a good one, and I think it, in particular, underscores how hard it will be for AI Agents to replace humans--in order for a company to fire an engineer in favor of an AI agent, they will need to be able to operate and debug and fix production problems.
* Domain expertise. I agree broadly, but I'm more convinced by your list of product domains. Backend, frontend, and mobile engineering seem sufficiently generic, and with enough training data, that AI will make fine recommendations. I suspect you'd have to go quite deep in these domains before establishing a big VORP (value over replacement programmer)
* Big +1 on communication (#5). And note that this is also valuable for communicating context to AIs.
* Agree that being good at Integration and Boundaries will be critical--nice framing. I contrast this with "design patterns" which are more context-free solutions in a box that AIs can probably master with us only having medium-level knowledge. I imagine they'll be giving us menus of design patterns to choose from.
Hi Addy, if you don't mind, I would love to share an Arabic version of your great piece with my community and keep the reference to your original post of course.
Thanks a lot Addy for publishing this post this week. Your post brings hope to software engineers, perfect for this time of the year ☺️.
Do you have any recommendations for improving on domain expertise for frontend?
I also saw posts saying frontend engineers won’t be needed 🤔 and will evolve to full stack or product engineers. I guess full stack will be needed mostly for startups as always and since there might be more startups thanks to AI it might be more demanded than before. What do you think? Should we still specialize in depth domain knowledge of only one software engineering role?
This is indeed helpful and timely.
I have had this dilemma on over relying on AI but reading this helps clarifies everything and gives valuable knowledge in leveraging AI in my career effectively.
Thanks for this
Maybe I'm just a boomer but I would advocate for juniors to spend some time with their AI assistants switched off in order for them to get to develop some confidence in their own learned and innate abilities.
This is such an incredible written post! Bravo man, bravo 👏🏾
I'm diffing your post with https://drewhoskins.substack.com/p/the-most-promising-ways-to-upskill to help myself figure out what I might have missed.
We end up agreeing on most everything (and I was untainted by your post, which adds believability), but there are a few small interesting differences:
* I didn't consider the operations upskill (#4). That's a good one, and I think it, in particular, underscores how hard it will be for AI Agents to replace humans--in order for a company to fire an engineer in favor of an AI agent, they will need to be able to operate and debug and fix production problems.
* Domain expertise. I agree broadly, but I'm more convinced by your list of product domains. Backend, frontend, and mobile engineering seem sufficiently generic, and with enough training data, that AI will make fine recommendations. I suspect you'd have to go quite deep in these domains before establishing a big VORP (value over replacement programmer)
* Big +1 on communication (#5). And note that this is also valuable for communicating context to AIs.
* Agree that being good at Integration and Boundaries will be critical--nice framing. I contrast this with "design patterns" which are more context-free solutions in a box that AIs can probably master with us only having medium-level knowledge. I imagine they'll be giving us menus of design patterns to choose from.
Hi Addy, if you don't mind, I would love to share an Arabic version of your great piece with my community and keep the reference to your original post of course.
I appreciate your feedback!
Please feel free to go for it!
Many thanks!
Here is the Arabic version of your piece: https://medium.com/smart-minds/future-proofing-software-engineering-career-d178ca4c384d?sk=1b39ec0a292056b00e3e1148d5b5377d
Inspired by your content, I also came up with this English article "A Day in the Life of a Software Engineer in 2025": https://rakiabensassi.medium.com/gen-ai-software-engineer-87adfdbce0b6?sk=a08600efc96cd33a2bfd3ed5b4a0df0b