aledevv

ps: Unfortunately I agree with you.

I particularly agree with this statement.

I don't know why manual work has been so denigrated over the last century. We believed that office labor was more important and healthier than manual labor. I don't think so.

As a developer, sitting all day typing in a stuffy office, without natural light, without sun, without air, is certainly no healthier than being outdoors, connecting with nature and other people. We come from nature and are made to be active, outdoors, and in the sunlight.

Today, with AI, many white-collar jobs are being called into question, and perhaps we can go back to loving certain traditional jobs.

What's under the hood, the people who use the product, don't care.

Customers, and ultimately companies as well, only care that the product works, is maintainable over the long term, and is bug-free.

Cathedrals in the desert are useless, and over-engineering only complicates things when there's no need yet.

I've also seen several successful projects that were actually quite weak behind the scenes, but they were simple and functional.

I don't fully agree this optimistic view. Unfortunately, for now, coding agents produce code that, if not further optimized upon "human" request, often generates more complexity than necessary.

It's true that this requires more computational effort for the agents themselves to debug or modify it, but it's also true that the computational cost is negligible compared to the benefit of having features working quickly.

In other words: agents quickly generate hyper-complex and unoptimized code. And the speed of delivery provides more immediate benefits than the costs resulting from bad code.

On the other hand, it's also true that the "careful eye" of an experienced developer can optimize and improve the output in a few simple iterations.

So overall (and unfortunately) the "bad code", if it immediately works, can wins against (or with) a good code.

This makes it even more crucial to have the most seamless ability possible to implement reverse and restore previous States.

The risk of the Agents actions becoming irreversible at the system level must be minimized.

I wonder how much all this can impact (and certainly will impact) the Real World, which will be increasingly robotized and automated: public services, finance, hospitals, schools, public administrations, military sectors (!), etc.

With AI, we have an exponential level of productivity. But what is being produced? 90%: garbage.

The problem is that what is being produced is essentially "garbage" generated by models trained on garbage. Quality knowledge is increasingly submerged and suffocated by spam and low-quality content.

The real challenge of the future will be filtering and cleaning up, on each level.

> "..but maybe it's a good thing that most of us don't allow this technology to reframe our thoughts."

No, you're not the only one experiencing this: I too had the same concerns as you: with every new thought, every new creation, I had to ask the AI's opinion, as if I were no longer able to judge, to decide, without consulting the AI (...just to be safe, you never know...).

The only way to regain your creative ability is to write down your thoughts yourself, read, reread, rewrite, correct, express your opinion...

What AI can't do is convey emotions.

AI and ML must be used to build a Better world, not to destroy it. Let's use it in a Humanly intelligent way.

My scripts typically also serve as technical documentation for specific features and they are a sort of "unique source of truth".

If you want to change the LLM, my advice is Kimi K2 that's much more chip and very skilled for coding.

(you are right: Opus is a very token-consuming even in my experience, but often it is over-skilled and you don't need it)

Most of the time, it's enough to build in a simple, clean, and lightweight way. Just like in the old days. Your server's resources will also thank you. Furthermore, the simplicity of web pages is also rewarded by search engines.

If it were up to me, I'd build sites exclusively in .md format :)

This one, on the other hand, is cleaner and lets you find what you're looking for quickly. And, last but not least, is much lighter.

It’s only right that the knowledge base remains in the public domain.

The real problem is that we must take a stand to establish proper guardrails for a technology that could become terribly disruptive and catastrophic.

If I were the creator of an LLM, I would forbid the use of 'my' technology for horrific purposes like warfare.

It’s fine to 'steal' knowledge to train models, but only if the goal is to build a better world.

Correct me if I'm wrong, but this means that many people are using LLMs despite claiming not to.

It's the first symptom of a dependency mechanism.

If this happens in this context, who knows what happens in normal work or school environments?

(P.S.: The use of watermarks in PDFs to detect LLM usage is very interesting, even though the LLM might ignore hidden instructions.)

Their presence has already radically transformed the orbital environment.

There are so many that in 2025 alone they performed around 300,000 collision avoidance maneuvers.

In short: on the one hand, they're convenient for us because of their fantastic internet connection, but on the other, they're generating truly unprecedented artificial traffic in space.

All this worries me a little.