As mentioned in Toy Projects I have a Beelink Mini S12 Pro (N100, 16GiB of RAM. 500GB NVMe SSD).

I’m managing it using AI. To give root permissions to an agent felt crazy at first. Now I don’t want to go back.

The setup

Networking

Tailscale for VPN. Installed everywhere: on the server, personal laptop and phone. It’s extremely easy to get it up and running. Like, 5 minutes and it’s all done.

It’s incredible how much can you do with AI. I’m having so much fun lately, that it’s unreal; the speed, the progress… it is crazy.

I’ve been working on small projects. Some of them took an afternoon. Other took a little bit longer.

10 years ago it was not the case. Those projects would take weeks to completion. I would often not finish them.

With the use of AI, I feel 10x. I finish them without problems or friction. I release them with ease. Modifications and adjustments take no time. It’s fun.

Lets make some assumptions about the future:

• Agent’s contexts will be irrelevant. No “dumb valley”. No 1M tokens max.
• Token generation will be virtually free (100€/month/dev for unlimited tokens).
• Token generation will be virtually instant. Time for mid-effort tasks: 1 second max. Time for high-effort task: 20-30 seconds max.
• Current bottlenecks will be solved: code reviews, agentic code non-maintainability, security issues, deployment, error detection/automatic fixing, …

The following are my predictions. This will be a live post that will update as I have more ideas. In 2 year’s time (2028-10-26), I’ll check which ones are correct.

Curiosity is the feeling of something pulling your attention. Leaning towards that pull will satisfy your curiosity.

Learning while being curious about the subject is a great way to make progress. It feels effortless. You don’t know how long the curiosity is going to last, tho.

If you have curiosity, don’t wast it. Let it pull you in.

Obsession is different.

Obsession is the feeling of not being able to help yourself but to do whatever you’re obsessed about. You must do it. To not to do it, you must fight it.

AI has changed how we code. How we learn. How we set-up our environments. I’m exploring the new landscape, and so far I’m liking it a lot.

I’ve noticed I deviate from the original task. Taking on tangents. I do it all the time.

For example, starting this very blog. I want it to be versioned with git, of course. Which hosting provider do I want to use? Ask agent to investigate self-hosting solutions. Setup a promising one in my home-server. Reverse proxy rule for nginx. New DNS rules on CloudFlare. SSL? Lets ask Claude to do it as well. What about a CI/CD pipeline, to deploy the blog? Done in 15 minutes. Overall it took 2h to setup.

Companies are pushing AI everywhere.

Into departments: product, design, tech, customer service, ops, fraud detection….

Into they organization processess. How teams interact with each other. How people interact with each other.

Now everything it’s all AI-mediated.

Voxium, writes on twitter (highlight is mine):

I am done with this shit. It is over. The state of engineering right now is horrible. It has been half a month since I started a new role at a big company. Nobody knows anything here. The specs, code, tests, PRDs, tickets, resolution of those tickets, reports, etc., everything is made by Claude Code. Nobody on my team likes this. They are being forced to ship as much as they can. I have heard multiple times from higher management that pushing code is not a bottleneck, so why are we slow? People are working 12 to 13 hours a day just to press enter. Nobody is reading anything. Humans in corporate are doing nothing on their own. Everyone, literally everyone, from an L1 to an L7 engineer here is doing the same thing. Talk to Claude. There is no sense of victory. Nobody is resolving bugs. In reality, nobody is thinking anymore. Everything is done by LLMs. It is so soul-sucking. I would not mind it, to be honest, if we were at least given the time to check out the code and see what is going where. But no, the goal is to just ship. No matter what happens.

I’m seeing a shift on how companies work nowadays.

5 years ago, Qonto was using the Toyota Management System, which is people-centered, with continuous improvement in their processes. Finished working products were valued. We had project-planning phases, where it was decided what to build, why, and at what cost. We were doing things right the first time. We were focusing in one project at a time, until completion. We were keeping technical scalability down. We took the time.

AI produces code fast. The code generated is non-deterministic. Code complexity tends to increase.

Any new project started using AI, looks good. Until it doesn’t. It decays very fast without human steering.

Providing contextual style-guides, or rules as Markdown doesn’t work. You expect deterministic outputs from a non-deterministic process.

AI will either edit your current program, or extend it.

On edit, it will often reuse the same style and standards in the codebase its modifying.