← Feed
@tinker@infosec.exchange
Post #3858149
2026-07-16 12:08 UTC
The argument that "AI works therefore we should use it" is akin to "Torture works therefore we should use it."
I don't mean that as hyperbole.
The underlying purpose of AI is at least twofold:
1) It puts a layer of abstraction over information.
- Primary Source: Raw data (eg video footage of an event)
- Secondary Source: Analysis (eg Low bias news and analysis of what this event means in the context of other events)
- Tertiary Source: Meta Analysis (eg Encyclopedia entries summing up Analysis)
- Quaternary Source: News Opinion and Propaganda (eg a "News Entertainment" channel that tells you what to think regarding the event)
- Pentanary (?) Source: TLDR summations divorced from underlying sources (eg AI summations, guides, etc that describe events or provide recommendations and guidance but follow behind the scenes algorithms that control the output)
The purpose of this layer of abstraction is control of information and removing the distribution of information from the actual events. This allows for influence campaigns, tuned marketing, tuned propaganda, and therefore control.
2) Centralization of compute power
- Computers in our hands that we control allow us a lot of power.
- Moore's Law no longer applies to localized compute power so we are in an era of repair and maintenance.
That means two things for corporations. Loss of income and loss of control.
- Forcing us to use centralized compute forces us to continue to spend money
- Localized AI engines still allow for centralized algorithms. (open sourced algos and localized engines dont factor in here as use and usecases are low)
AI is being pushed so heavily because both governments and corporations need it to further their wealth and power.
They push it via both making it cheap and accessible and having a veneer of usability.
This is a similar pattern to "Underpricing competition to drive them out and then jacking up rates when you have a monopoly", "Embrace, Extend, Extinguish" and similar approaches.
So.
I am not arguing whether AI is useful. I am arguing that its use is detrimental to us. In fact, the more it becomes useful the worse it becomes BECAUSE more people adopt it. It hurts us as we use it. It hurts us more as we use it more.
Tying it back, I don't care if torture is useful and gets your desired results. Torture is bad and it hurts us.
#Linus #AI #Linux #solarPunk
Replies (3)
-
@tinker@infosec.exchange Torture mostly doesn't, as victims will agree to anything to make it stop.
Open ##3858436
-
@tinker@infosec.exchange Considering the abysmal economics of it all, that they burn through far more money than they make, also indicates there are motivations other than pure profit.
Open ##3859349
-
@tinker@infosec.exchange
"The argument that "AI works therefore we should use it" is akin to "Torture works therefore we should use it."
Sorry .. but i call BS on that.
It's like say, we cannot ever use certain information about biology and the human body because some of the data was based on "work" of dr. josef mengele.
Which _IS_ bad and did cost lots of lives for _real_ .... but?
You can see where your comparison completely breaks apart?
Not using certain cosmetics because they did experiment on animals in the development is one thing.
It's like, you would rather let ppl die because of diabetes instead of giving them insulin just because we had to resort to animals as a source...
Yes ... LLMs and all the stuff around - including the massively intense drain on the resources of this planet - is a problem .... We need to deal with it and ASAP learn to get a healthy "relationship" with this new tech....
But saying it is like using torture?
No .. not really ........ it's just more like "they" try to create new digital overlords ... and we have to stand up to make them into useful unicorns instead of all seeing and all controlling monsters.
Open ##3862338