KG.codes

How would you react to this?

At work, a non-technical person asked me a deep technical question.

I’m familiar with the topic, so I gave them a deep technical answer from memory.

They then go straight to ChatGPT and ask it, right in front of my face (I guess not fully believing me).

ChatGPT gave the same answers, and they finally said, “Oh, okay”.

Not going to lie - while waiting on ChatGPT to answer, I got kind of nervous about the accuracy of technical information that for years I had been pretty sure about.

This experience gave me a reality check on the years worth of technical information I’ve memorized..

The reality is.. today, nearly anyone today can quickly ask/validate a very specific technical question with A.I., versus the minutes/hours of aggregating Google search results like before.

Yes, it’s not always accurate, but Googling never was either.

I know there might be experienced technical people who are resistant to adopting A.I. I encourage you to reconsider as the baseline of knowledge and productivity in every adjacent field is changing fast.

On the other end, I think this makes our unique experiences, processes, practices, and principles in our given fields more important than ever now.

What do you think about this?

Are ya’ll noticing AI tools are kinda buggy, or no?

I have, and there’s growing opinions on Threads/X about it.

But there is a “good” reason in my opinion: these popular AI companies are in a race and our purposely choosing to “move fast and break things”.

The strategic tradeoff is releasing features (and fixes) daily versus having a less buggy product.

Thoughts?

K.I.S.S. = Keep It Simple Stupid

I got asked what are some key qualities that make a successful software engineer. I think this advice relates to many careers, in and outside of tech.
KEEP THINGS SIMPLE.

As we get better and more senior, we think we should be coming up with more complex, “clever”, intricate solutions.
But a lot of times what we should actually be doing is improving our skills in order to make the solutions we build more simple.
Often the most clever solution is the simplest; you find out a why to do it easier with the least level-of-effort and impact.

A lot of times when I’ve worked with heavily architected things, at some point, over time, we get to a point where everyone working with it is now saying how “over-architected” this thing is.
Over-architected solutions can be harder to understand, which makes them harder to change/maintain, which makes them less valuable / more troublesome, which creates headaches for you and your team.

Sometimes there is no way around a complex solution/architecture.
Just always keep in mind whether the complexity you are adding is truly required by the solution, or if you are just adding because it feels “clever”,

thoughts?

How are people making AI work for them while they sleep?

Here’s a simple example from something I do:

1. use GitHub Issues to list features and bugs I want done.

2. Then trigger AI to run on a frequent schedule, pick an issue, and create a pull request with the fix.

When I’m ready to review, it’s already pushed code, left a comment, and now I just review and merge.

You can schedule this with cron jobs, ChatGPT, Claude, or tools like Zapier and Make.

And you can manage tasks with GitHub, Jira, Linear, Notion, or Trello.

The key is giving AI a task list and a schedule, then it can work without you.

At that point it’s just about how many tasks you have and how much you have to spend on AI (how frequently you trigger it).

Thoughts/suggestions? Are you doing anything like this?