Don’t deny real risks, but don’t accept the AI industry’s mythology as the starting point for discussing risks. The useful question is what currant AI’s actually do, who controls them, and what infrastructure and power relationships we are building around them.
AI is useful technology wrapped in a massive social, political and financial mythology. The #OMN is less interested in the usual argument between “AI will save humanity” and “AI will destroy humanity”. as both become noise in the current #LLM mess. To get a clearer picture we need to step back to ask a much simpler question – what does the technology actually do, and what are we building around it? Because there is a trap here.
We take an interesting piece of software, call it “AI”, then AI becomes AGI, AGI becomes superintelligence, and suddenly we are discussing machines as if they are becoming digital people. The #dotcons are happy with this story, as this story justifys gigantic data centres, enormous energy use, massive investment, extraordinary infrastructure contracts and the concentration of yet more technological power in the hands of the #nastyfew.
But there are serious questions about “AI” safety – autonomous weapons – cybersecurity – biological research – critical infrastructure – economic disruption – surveillance – disinformation – concentration of power. The replacement or degradation of human work. But there is a difference between looking at concrete risks and accepting a mythology about an imaginary future machine.
lets look at some of these risks: “AI will destroy humanity” argument starts somewhere like this – AI becomes more intelligent than humans – it develops goals – those goals don’t align with human interests – it becomes capable of improving itself – it gains access to infrastructure – it protects itself from being switched off – it expands its ability to act. Eventually humans become an obstacle, therefore humanity is eliminated. If were we actually building #AI not the current LLM’s that would be a possible speculative scenario.
But notice how many assumptions are packed into that chain. AI → AGI → autonomous agency → self-improvement → independent goals → infrastructure access → resistance to control → human extinction. Each arrow needs evidence, to judge this #KISS becomes useful. Don’t start with the end of the story, start with the machine in front of you. What can it actually do?
An #LLM generating confident bullshit isn’t evidence of an emerging artificial mind, a badly isolated “autonomous” system causing damage is an engineering and security problem. A corporation using dysfunctional AI to replace workers is an economic and political decision. A military using LLM’s based on the history of litterateur, feed partial data to kill people is a legal, political and military decision. A government using AI for surveillance is a governance problem. A data centre consuming enormous amounts of electricity and water is an infrastructure and ecological problem.
These things don’t require us to believe that a machine has become conscious, they require us to look at the #nastyfew are pushing over us. If we don’t look at the problems from this view we make the same mistake we have made with technology for decades, turning a social problem into a technical problem. Have you noticed that it is the priests of the #deathcult who loves AI.
But the are real problems in the #AI arms race – military competition → autonomous weapons → humans delegate decisions because machines are faster → competing systems interact → escalation/control failure → mass killing. Here extinction result from autonomous weapons, nuclear systems, biological weapons or other military infrastructure. That argument doesn’t require AGI it is simply the current mess with added tech mythology.
The #OMN critique isn’t so much about whether LLM’s are “good” or “bad” it is about the economic and social system surrounding the current #techshit paths, that we already have piles of #techchurn to compost with out adding more. The #OMN doesn’t want to repeat this mess with AI, what we need is ways of keeping power visible enough that people can challenge it, for this we need to compost the mythology.
This is why the #OMN is deliberately boring compared with the AI storeys. Our focus is not technological salvation, or technological apocalypse. We need useful #KISS technology embedded in social structures where people can understand, challenge, change and, when necessary, walk away from demonstrated technological risks of speculative technological mythology. The mistake is always turning the tools into a mythology and then building society around the mythology. The #deathcult has had fifty years of economic “thinking” on this, look at the mess, now lift your eyes.
Let’s compost it and grow something alive.
