About
Hi, my name is Jeongwoo! I like tech and am interested in privacy, safety, and ethics.
Experience
Education
Other similar profiles
Explore more posts
-
Sriram Ramkrishna
You know I was just thinking about these code based interviews where they make you do all these coding tasks so they can watch you code and see how you think. It's just sort of silly to me. Consider asking these kind of questions instead: * What compromise did you have to do in a code base that led to a bad outcome and what did you do about it? * What compromise did you do that had a startling result that you didn't anticipate and how did it change your approach if at all? * What code refactor are you most proud of? What made you rethink that? * What piece of code kept you up at night? * What does good code look like to you? These interview questions help both the interviewer and the interviewee bond. Ultimately, it's not about lines of code you can write but how well everyone works together as a team and self discovery. If it is about lines of code, that's not a company I would cherish working at.
28 Comments -
Todd Fisher
Thanks #UtahJS organizers for an enjoyable event the other week and for having me present my thoughts on moving fast as a software engineer! It's been interesting to see how AI has influenced our ability to deliver value quickly. One thing that has been top of mind for me recently is seeing how far people can take things like swarms, CC agents and AI flavored Spec Driven Development (https://lnkd.in/gyM6QdwE) to move even faster. I definitely still see a lot of value in the iterative ai assisted development approach (of alternating between focused AI prompts and reading/adjusting code) with tools like Cursor, but this latest evolution of tools is taking us to the next level. Of course the biggest downside of AI generated code is the amount of AI slop code that we as engineers have to debug and fix. Sure, as we become more proficient with these tools the amount of bad code will also be reduced, in part by us doing better prompting, planning and guiding, but there will likely always be some amount of suboptimal code generated by AI we will have to address in a more manual way. Through this whole process of adopting AI assisted coding, I keep coming back to the thought that "code is a liability, crap code is a vulnerability." Stay safe out there my friends as you push your AI tools to their limits!
1 Comment
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content