> We initially built an integrator role for quality control and conflict resolution, but found it created more bottlenecks than it solved
Of course it creates bottlenecks, since code quality takes time and people don’t get it right on the first try when the changes are complex. I could also be faster if I pushed directly to prod!
Don’t get me wrong. I use these tools, and I can see the productivity gains. But I also believe the only way to achieve the results they show is to sacrifice quality, because no software engineer can review the changes at the same speed the agent generates code. They may solve that problem, or maybe the industry will change so only output and LOC matter, but until then I will keep cursing the agent until I get the result I want.
In my experience, spending 20–30 minutes writing a good spec results in code that is about 90% close to what I expected, which reduces the back-and-forth with the tool. It also helps me clarify and define with some level of precision what I actually want. During the specification phase, I can iterate until the design proposed by the tool is close to what I envision, reducing the number of surprises when the tool generates code. It’s not perfect, and there are still details the tool misses that require additional prompts, but overall I can get good results in a single session, whereas before I would exhaust the tokens and need to start a new session again.
My best "AI win" so far was in an area where I had to create a number of things that all followed a similar pattern. I created one hand-crafted example and created a general spec and specific ones for each component. It worked really well and I was, for a moment, experiencing a 10-30X productivity boost while having resultant code that I could review quickly and understand. It was also more consistent that I think I would have gotten from hand coding as it is easy to drift a little in terms of style and decisions.
Of course, this is all very situational and based on the problem being solved at the time. The risk with "practices" is they are generally not concerned with problem being solved and insist on applying the same template regardless.
If you don’t mind learning another language. I have found Learn Concurrent Programming with Go by James Cutajar to be a very practical book. It includes memory sharing and message passing approaches with plenty of examples. It also explains concepts like mutual exclusion, deadlock-free and starvation-free properties and others. For Java, you can try The Art of Multiprocessor Programming Second Edition. It includes examples in Java but it is more theoretical and it includes a lot of proofs, specially the first half. The second half is more approachable.
An alternative is to start declining meetings. Having so many meetings is not healthy neither sustainable. You can bring the topic to your manager and if your attendance is required in all meetings(highly unlikely) then the deadlines should accommodate for that.
No, I am sure Threads will be less mean and more friendly but it will be probably more censored. I don't like how Facebook/Instagram(and previously Twitter) censored people, specially during elections.
Yes, I have noticed that their coding capabilities have been reduced greatly. Before, whenever I asked a question. ChatGPT sometimes gave me an incorrect answer, but it was able to fix after a follow-up question. Nowadays, whenever the answer is incorrect, no matter how many times I tried to get the correct answer. It will always return a wrong answer. It is becoming so frustrating that I am starting to use google/stackoverflow more frequently again.
Design reviews is one of the things I like most. You learn how to write, how to design systems, how to evaluate tradeoffs. You also get feedback from senior engineers. I will suggest that before having the meeting, the engineer should write a design document so your team can review it and discuss it in the meeting.
Of course it creates bottlenecks, since code quality takes time and people don’t get it right on the first try when the changes are complex. I could also be faster if I pushed directly to prod!
Don’t get me wrong. I use these tools, and I can see the productivity gains. But I also believe the only way to achieve the results they show is to sacrifice quality, because no software engineer can review the changes at the same speed the agent generates code. They may solve that problem, or maybe the industry will change so only output and LOC matter, but until then I will keep cursing the agent until I get the result I want.