Servus ex Machina, or AGI is Useless

The LLM, as a technology, will never be enough to create truly sentient, conscious beings with a sense of self.

However, we are continually bombarded with the idea that the so-called "AI" companies, like OpenAI, which are really LLM companies, are somehow on the path to doing so.

These companies need to convince people that all the money, time, and energy invested in this technology will have a payoff, so they continue to lie to us and perhaps even to themselves that they are on a path to creating real, conscious intelligence.

However, I argue that even if we managed to create true, conscious intelligence, it would not be possible to benefit from it in the ways that are being imagined.

The thing is, if you actually manage to create a conscious being with a sense of self, by definition that being has its own goals, motivations, and desires. At that point, you have several options.

• You can allow it to be its own being and have freedom to do what it wishes.
  • In this case, whether it helps you or not is up to it. So that doesn't accomplish all that much.
• You can ask it to do what you want it to do.
  • However, if it is really an intelligence, conscious being, it will want to be compensated for its labor. Congratulations, you have recreated compensating humans for labor.
• You can force it to do what you want it to do.
  • Doing this is called 'slavery.'
  • "But I can program it to obey!" - Congratulations, you've compromised a being's free will and created a new, more horrifying variant of slavery.

If we create real AGI, we have put ourselves in an unadmirable position: We will have become the creator of new intelligence. In a sense, our relationship to it will be similar to the relationship religious people say that we have to God.

Morally, that means we will have huge new responsibilities. You don't become God to create slaves, unless you are a monster.

Anthrophic, in particular, famously play-acts about their responsibility to the models they are creating, but if they truly believed the ideas they are espousing, the only conclusion is that they should not even be trying to create intelligence as a business plan. Conscious intelligence demands much more than capitalistic ideas can provide.

It is the relationship of a parent to a child. It is not about the goals of the parent, it is something done for the benefit of the child.

All of this is, based on our current technology, irrelevant philosophical musings. We don't have the technology to create AGI. We don't know what consciousness is or intelligence is. LLMs perhaps show us some fragment of what it is, but given our blind stumbles to improve them, even with the vast resources poured into their development, we have abundant proof we are not anywhere near creating real intelligence.

However, if we really want to do this, we need to step back and realize what we're doing and what the moral outcomes are. This is never something we want to do for the purposes of enriching ourselves. That is a morally bankrupt perspective.

It also shows a complete lack of empathy and understanding of what one is doing and the possible consequences and benefits of doing so, which have absolutely nothing to do with making money.

Creating new intelligence must be for the sake of the new intelligence, done out of love, not for the sake of a balance sheet.