[ExI] Elon Musk, Emad Mostaque, and other AI leaders sign open letter to 'Pause Giant AI Experiments'

sjatkins sjatkins at protonmail.com
Fri Mar 31 08:41:36 UTC 2023


I very much disagree with those that want to shut GPT-x down. The refrain that the new tech will take too many jobs has been hurt before since the Industrial Revolution began. Some jobs disappear and others open. That the language models don't understand means they are not AGIs and thus not directly possibly human competitive. They have no agency. What they are is a fanstastic tool that needs to be used by humans to do anything. In other words these language models are a fantastic augmentation of human abilities. We really really need that. We need as much effective human intelligence and productivity as we can get and we need it as fast as we can get it.

I have a suspicion that some powers that be are a bit nervous about the potential to augment the effective intelligent abilities of so many. It could threaten their position and comparative advantage. I think they are especially afraid now that more work is coming out about how to more efficiently and cheaply augment and perfect these systems. If that comes to past it will not be under the control of those that can afford large resources. That also gives me hope that it is already out of the bag and proliferating too fast to be stopped.
- samantha

------- Original Message -------
On Friday, March 31st, 2023 at 2:25 AM, Rafal Smigrodzki via extropy-chat <extropy-chat at lists.extropy.org> wrote:

>> TIME Magazine, today March 29.
>> "Shut down all the large GPU clusters (the large computer farms where the most powerful AIs are refined). Shut down all the large training runs. Put a ceiling on how much computing power anyone is allowed to use in training an AI system, and move it downward over the coming years to compensate for more efficient training algorithms. No exceptions for anyone, including governments and militaries. Make immediate multinational agreements to prevent the prohibited activities from moving elsewhere. Track all GPUs sold. If intelligence says that a country outside the agreement is building a GPU cluster, be less scared of a shooting conflict between nations than of the moratorium being violated; be willing to destroy a rogue datacenter by airstrike."
>> -Eliezer Yudkowsky
>>
>> https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/
>>
>>>
>
> ### I remember Eliezer being quite libertarian back in the day and now he wants the World Government to bomb any independent locus of thought to smithereens. People change.
>
> This is stupid. A government is a long-feedback loop entity, extremely inefficient and slow in responding to truly new challenges, unlikely to maintain alignment with the goals of its human subjects and its failures grow with its size. It would be suicidal to try to use the mechanism of government to solve AI alignment.
>
> Our only chance of surviving the singularity is to build a guardian AI, an aligned superhuman AI that would be capable of preventing the emergence of unaligned or malicious superhuman AIs - a bit like a world government but without the psychopaths and the idiots.
>
> Our best chance for building the guardian AI is for highly competent and benevolent AI programmers with unlimited resources to work as fast as they can, unimpeded by regulations (see "long-feedback loop" and "extremely inefficient" for why regulations are a bad idea). Give them all the compute they can use and keep our fingers crossed.
>
> Maybe we'll make it to our rapture of the nerds.
>
> Rafal
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230331/7780a7f4/attachment.htm>


More information about the extropy-chat mailing list