(2025-09-08) As long as you are consuming, you are not creating
---------------------------------------------------------------
As work has started taking most of my focus time, I'm in a bit of a pause
regarding all my hobbies (except learning how to play Riichi Mahjong, of 
course). This, however, isn't going to extend for long, but at the same time 
it allowed me to think about some things I just didn't have capacity to 
think about. While watching yet another video on YT, I suddenly had a 
thought that bothers me to this day: "am I doing the right thing by watching 
YT instead of building something of my own right now, at this very moment?" 
Indeed, I came to a conclusion that no matter what you're doing, your 
creativity process stops once you start consuming other content.

This, to some extent, also applies to tools you use in your creative work.
You inevitably consume them even while you're creating. One thing that can't 
be doubted though is the creation still is attributed to you, not the tool. 
For instance, lexipwn wasn't created by Vim, Go and Fyne; it was created by 
me using those three tools. i still was consuming them but also putting in 
my own input for producing some valuable output. The tools I use are quite 
advanced but I still am in control of the process. The real problems start 
when your own input becomes almost nothing compared to what the tools do for 
you. But an even bigger problem is... that's what the users had always been 
wanting.

We may despise all those "AI art" and "vibe coding" trends as long as we want
(and rightfully so), but the reality is, that's how people had been 
imagining the future of interaction with computers almost since the 
beginning of the information age. Not a single command entered, not a single 
key pressed, ideally. And in pursuit of this utopian goal, this is not the 
first or even the fifth time either when the masses have given up a bit of 
control and freedom for convenience. Transition from simple and 
understandable architectures to x86, from procedural and stack-based 
programming to OOP, from simple graphical APIs to OOP-driven GUIs, from IRC 
and forums to "social networks" and anti-privacy messengers, from on-premise 
servers and self-hosting to cloud computing... Does it surprise anyone that 
now some people already do something as stupid as relying upon LLMs as a 
replacement for search engines, let alone coding assistants? Yes, those 
people probably don't even realize that they turn into consumers from 
creators more and more and more, but they still think they are in control, 
and nothing can be done about it.

I don't want to be that guy but sometimes I wish for a global technological
disaster to happen, just because they won't understand the truth in any 
other way. They keep installing Faildows on mission-critical systems (as if 
Stuxnet, WannaCry and Crowdstrike weren't enough for anyone with a 
functioning brain to stop doing that), they keep giving up their private 
data to megacorps (and, by proxy, the data of their unsuspecting customers 
in case they represent businesses), and now they also trust (cloud-based!) 
LLMs to do sensitive jobs. IMO they deserve to suffer all the consequences 
of such an event (the probability of which is increasing year by year), but 
the problem is, totally innocent people will also be affected. The only way 
of safeguarding them is to make them more literate in topics they don't want 
to know about in the first place, so I don't know how to deal with this 
besides spreading out the basic counterpropaganda like:

* don't use any cloud-based LLM (especially ChatGPT) for anything serious,
* don't use anything by M$ or Meta or Adobe at all,
* stay away from iCrap in case you are choosing your only mobile device,
* stay away from 95% of Android phone brands unless you know how to customize
their firmware, because the stock one comes with malware preinstalled,
* the same applies to Google Pixels: reflashing to GrapheneOS or DivestOS or
LineageOS is now a must if you care about your data,
* what gets into the cloud no longer belongs to you, no matter what you're
told,
* if a task can be completed offline, this should be a preferred option,
* in case a program is not FOSS, most likely it's designed to spy on you or
otherwise rob you (even if you already paid for it), and there's no legal 
way to prove otherwise.

There, I said it. These are the principles that anyone (again, with a
functioning brain) should place before any thought of convenience in order 
to not fall victim to scammers (both small-scale and large-scale) and to 
even have a chance of survival in an event of a tech-induced crisis. Another 
piece of advice, although it obviously might not apply to everyone out 
there, is to learn a programming language yourselves and start DIYing your 
own tool ecosystem. I think both Tcl and Python are good starting points for 
those who know nothing about programming yet. The point of this is to feel 
enough independence from the vendors who make decisions for you. 
Additionally, if you already are so into LLMs that you can't imagine your 
life without them, opt for privacy-first solutions and run them locally as 
much as possible. Although I have several things against Ollama and LM 
Studio, both of them are good starting points for newcomers as well.

At the end of the day, my point is: keep the balance between consumption and
creativity. Use the tools but don't let yourself become one. Hope for the 
best, prepare for the worst.

--- Luxferre ---