In the end, our leadership changed what we were building so often that all of the UI work was scrapped long before we shipped. We ended up launching a janky console, quickly assembled by SDEs who were racing against deadlines. We skipped virtually all operational readiness work to meet the launch deadline. After claiming the launch win, the director, two managers, and the pm promptly left for other orgs.

Per TFA, this thing is literally taking screenshots of what is on the employee's screen. At work my screen sometimes had things such as: performance data on other employees, my own PII from HR systems, PII from customers, password managers, etc. It's also logging keystrokes. How many times do you type passwords a day.

Collecting that kind of information on purpose is truly wild. Imagine the security safeguards you would need just to prevent it from leaking. Wait what, they're explicitly collecting it to train LLMs with it? God help us all.

I inherited some vibe-coded scripts that dealt with AWS services like Bedrock and S3. These scripts needed create various AWS SDK clients. These clients needed to know which account/region to use.

Had this been well designed, there would have some function/module responsible for deciding what account/region to use. This decision point might be complex: it might consider things like environment variables, configuration files, and command line arguments. It might need to impose some precedence among these options. Whatever these details, the decision would be authoritative once made. The rest of the code base should have expected a clear decision and just do what was decided.

Instead, the coding assistant added optional account/region arguments in many submodules. These arguments were nullable. When left unspecified, "convenient" logic did its own environment lookups and similar. The result was many "works on my machine" failures because command-line arguments affected only certain portions of the program, environment variables others, config files still others.

This is grim stuff. It's a ton of code that should not exist, spreading the decision all over the code.

This is a beautiful articulation of a major pet peeve when using these coding tools. One of my first review steps is just looking for all the extra optional arguments it's added instead of designing something good.

Also, I am a novice when it comes to programming with sound, and today I have been dorking with a simple limiter. ChatGPT knows way more than me about what I am doing. It has taught me a ton. And as magical and wonderful as it is, it is incredibly tedious to try to work with it to come up with real specifications of interesting properties.

Instead of banging my head against a theorem prover that won't say QED, I get a confident sounding stream of words that I often don't even understand. I often don't even have the language to tell it what I am imagining. When I do understand, it's a lot of typing to explain my understanding. And so often, as a teacher, it just is utterly failing to effectively communicate to me why I am wrong.

At the end of all of this, I think specification is really hard, intellectually creative and challenging work. An LLM cannot do the work for you. Even to be guided down the right path, you will need perseverance and motivation.

I mentioned all of this to my boss, who is great btw, and was told there is just no fighting this. After several months of trying to make the best of it, I’m done. I’m planning to leave after my next vest.

It’s really a shame. I liked a lot about the job. So many good people have been forced out.