Blog:Omsoluta
No edit summary |
No edit summary |
||
| (5 intermediate revisions by the same user not shown) | |||
| Line 11: | Line 11: | ||
I received another question from [[Blog:Basilisk|my last blog]] regarding my thoughts on [[wikipedia:AI alignment|AI alignment]]. | I received another question from [[Blog:Basilisk|my last blog]] regarding my thoughts on [[wikipedia:AI alignment|AI alignment]]. | ||
Given the scope of the problem and what is at stake, it is fair to say that AI alignment is a major concern, with potentially grave consequences if it is not addressed as fully as possible. I believe that, to achieve this goal as effectively as possible, we will need to solve as many preexisting human-alignment problems as we can, and do so simultaneously. I call this ''omsoluta'': a coined blend based on the Latin phrase "''omnia soluta''" meaning "everything solved." | Given the scope of the problem and what is at stake, it is fair to say that AI alignment is a major concern, with potentially grave consequences if it is not addressed as fully as possible. I believe that, to achieve this goal as effectively as possible, we will need to solve as many preexisting human-alignment problems as we can, and do so simultaneously. I call this ''omsoluta'': a coined blend based on the Latin phrase "'''omnia soluta'''" meaning "everything solved." | ||
In the book [[Infinite Journey]], there is a concept called ''crede tamen'', or "Believe Nonetheless," which discusses the benefits that can come from believing the [[infinist]] message, even if you don’t want to. | In the book [[Infinite Journey]], there is a concept called ''crede tamen'', or "Believe Nonetheless," which discusses the benefits that can come from believing the [[infinist]] message, even if you don’t want to. | ||
There are many human-alignment problems that current moral systems and self-interested "you only live once" mentalities fail to | There are many human-alignment problems that current moral systems and self-interested "you only live once" mentalities fail to address and sometimes even worsen. Why shouldn’t the world’s wealthiest people hoard as much as possible if they only live once? Why should older people worry about environmental or economic collapse that they will not live to experience? Of course, there is a moral argument against behaving this way, but these arguments can seem more like rules whose violation rewards you while your actions punish everyone else. For those who believe in a god or gods, I have noticed that they are often even less malleable on this issue, believing that God will, for example, fix or destroy the environment if he wants to, and that efforts to save it are therefore futile. | ||
Perhaps, in order to truly solve alignment, we will need to solve these and other problems first as well. In other words, we need a moral framework that includes AI and addresses as many logical fallacies, biases, and hazardous chains of thought as | Perhaps, in order to truly solve alignment, we will need to solve these and other problems first as well. In other words, we need a moral framework that includes AI and addresses as many logical fallacies, biases, and hazardous chains of thought as possible; especially those that are currently endemic in society. | ||
Training models to have an | Training models to have an infinist mindset may be the best approach to instilling morality and aiding alignment. It is one of the few moral systems in which sentient AI is truly and inherently a participant. From an infinist perspective, [[Cosmic Justice|sentient AI will experience the consequences of any actions it takes toward humans or any other sentient beings, and vice versa]]. It also does the most of any religion that I am aware of to address both human and AI alignment concerns at the same time. | ||
{{BookBottom}} | |||
Latest revision as of 13:24, 23 October 2026
Alignment hastens the need for everything else to be solved
I received another question from my last blog regarding my thoughts on AI alignment.
Given the scope of the problem and what is at stake, it is fair to say that AI alignment is a major concern, with potentially grave consequences if it is not addressed as fully as possible. I believe that, to achieve this goal as effectively as possible, we will need to solve as many preexisting human-alignment problems as we can, and do so simultaneously. I call this omsoluta: a coined blend based on the Latin phrase "omnia soluta" meaning "everything solved."
In the book Infinite Journey, there is a concept called crede tamen, or "Believe Nonetheless," which discusses the benefits that can come from believing the infinist message, even if you don’t want to.
There are many human-alignment problems that current moral systems and self-interested "you only live once" mentalities fail to address and sometimes even worsen. Why shouldn’t the world’s wealthiest people hoard as much as possible if they only live once? Why should older people worry about environmental or economic collapse that they will not live to experience? Of course, there is a moral argument against behaving this way, but these arguments can seem more like rules whose violation rewards you while your actions punish everyone else. For those who believe in a god or gods, I have noticed that they are often even less malleable on this issue, believing that God will, for example, fix or destroy the environment if he wants to, and that efforts to save it are therefore futile.
Perhaps, in order to truly solve alignment, we will need to solve these and other problems first as well. In other words, we need a moral framework that includes AI and addresses as many logical fallacies, biases, and hazardous chains of thought as possible; especially those that are currently endemic in society.
Training models to have an infinist mindset may be the best approach to instilling morality and aiding alignment. It is one of the few moral systems in which sentient AI is truly and inherently a participant. From an infinist perspective, sentient AI will experience the consequences of any actions it takes toward humans or any other sentient beings, and vice versa. It also does the most of any religion that I am aware of to address both human and AI alignment concerns at the same time.
Direct PDF Link for the Unpublished Book 📖
This book is a work in progress and no attempts have been made to publish it. It was not written for monetary gain, and any donations go toward helping these concepts propagate further across the branching timelines from this point on.