I fear AI will lead to inevitable decline in software quality. A lot of companies are going to try to replace humans with AI. And if the end result doesn't elicit backlash, more will follow. It may not fly in some domains, but a lot of people are already used to using buggy software.

I think a lot of AI replacing engineers talk is just cover for layoffs or lack of business growth. If AI ever gets to that point, it's not going to just replace software engineers in isolation. If it was capable of replacing engineers, what's to stop it from replacing management as well? And why would you need HR when all your employees are AI?

I asked it to make some changes to the code it wrote. But it kept pumping out the same code with more and more comments to justify itself. After the third attempt I realized I could have done it myself in less time.

Claude doesn't seem to have a built in search tool but I tried this with a MCP server to search google and it gives similar results.

Nim is fast, powerful and has a lightweight syntax. I used it for a lot for hobby projects. I wrote program in Nim that started bringing in some money. But soon as feature requests from customers started coming in, I had to rewrite it.

The tooling for Nim and library ecosystem is just not there. It's so much more productive to work with Python, .NET or JVM. I decided to rewrite it in Kotlin because JVM gave me similar performance.

I am curious and wondering if you could elaborate on what you ended up making and why picked nim over lets say python, was the core usage something that was performance based or did it really fit the nim feature set better than most languages?

Would you say nim is good for prototyping or would you say it was a waste of time since you ended up having to port it for tooling?

I don't think this was worth the long article. Sounds a lot like the SBF of Turkey.

They have to say that to make the number go up. All these companies have heavily invested in AI.

AI is just another tool whose output depends on the skill of the user. You can't put people without domain knowledge in front of a LLM and get good results. It's very good at producing output that's looks good enough to convince non experts that it can do the job.

If you are connected to a network without an Internet connection, it just becomes unusable. Internet connection is somewhat unreliable in my area, and I had an internet outage that lasted for days during the COVID lockdown. I feared it was a malware infection causing the slow down. I switched over to Linux not long after.

I have used this set up in the past. With GPU, NVMe and USB passthrough it feels pretty much native.

There are definitely some caveats. You can't just passthrough any GPU. Post High Sierra AMD has the best support. Select Nvidia GPUs like the GT 710 still work, but it doesn't do hardware video decoding. I could never get sleep to work either.

With Apple moving towards ARM in the future, I thought of giving Linux another shot. It's not as polished but so far has been adequate.