[ExI] ai in education

[spike at rainier66.com]

From: John Clark <johnkclark at gmail.com> 
.

> The same reasons the military distrusts Anthropic would cause me to distrust it:

 

>…But do you distrust Anthropic more than you distrust the US military?

 

We are not paying Anthropic to defend our country.  Those two are not comparable.

 

>…I don't, not when the commander in chief of that military is He Who Must Not Be Named….

 

Anthropic was not elected in a 7-0 clean sweep to lead the executive branch of government.  Those two things are not comparable.

 

 >… you don't answer questions that might cause you to doubt your worldview, so I will repeat it now for the third time….

 

John I am following your rule you have for Extropolis: don’t be boring.

 

Political matters are more at home on Extropolis than they are here.

 

>…Who do you believe has a history of telling fewer lies, the scientist Dario Amodei who is the head of Anthropic, or the most famous twice divorced TV game show host in America?

 

Was Dario Amodei elected to run the executive branch?  Does POTUS run an AI company?  Those two guys are not comparable any more than we can compare Amodei to Gary Kasparov or to Alysa Liu.  

 

…

 

>…I understand why the military might not want to purchase Anthropic products but they have done much more than just that, they have designated the company a supply chain risk, something that has never happened before to a US company….

 

My best guess on how that happened is that Amodei caused the top brass to suspect that AI agents could invade software in existing defense systems with a priority on global safety, rendering them useless or harmful to the army which fired it.  Consider the comment you made recently about an AI controlled defense system taking into account who gave the orders to fire it.  If Amodei said anything like that when pitching his company to the pentagon, I completely understand their reaction.  John if you tried to sell an AI system to the pentagon, good chance you too would be shown the door.   

 

>…So do you really believe, as POTUS does, that Anthropic deserves to be assassinated because it places too much emphasis on AI safety?

 

I have nothing at all against Amodei.  I have seen his videos and think he is a very good guy.  I don’t know or care about his political views.  The military (every one of them in every country) places its priority on the safety of its own country.  If any company presents as prioritizing global safety, they will be shown the door with enthusiasm.

 

If one is trying to sell something to the pentagon, one must read their often crazy specifications, and follow those specifications to the letter, even if they don’t make sense.  Those specifications are written the way they are because often the pentagon is buying a system to use in ways they do not wish to reveal.  An example might be they are looking for an agent which will invade the control systems of the other guy’s rockets and guns, causing them to do something other than what its army intended.  

 

 

 

 > John you assure us with complete confidence that such a system doesn’t already exist 

 

>…The primitive system you described certainly does exist, but who cares.

 

We cares.  Iran has missiles and might have nukes, or will soon.  Or would have, had not they gotten involved in a conflict, and started firing their missiles in all directions at their own neighbors.

 

 

> and that anyone who makes it to the top of the military is stupid. 

 

>…I would maintain that there is empirical evidence that the person at the very top of the US military is not only very stupid and showing clear signs of Alzheimers, he is also evil. And no, I am not afraid to use that word…

 

That happens.  But it didn’t stop his predecessor.

 

I am not so much a believer in the current POTUS.  I am neutral on him.  But I am a firm believer in the process that got him there.  That process selected him in a 7-0 clean sweep of all the states in play.  Shrug.  I didn’t think that would happen, so I lost money. 

 

 

> We are buying AI.  We need complete control of it before we can trust it with our defenses,

 

>…If the US military demands complete control and certainty at how an AI will behave before they use it then the US military will NEVER be able to use AI. And the Chinese military will beat the US military into a bloody pulp. 

 

John K Clark

 

There’s more to it than US vs China.  The first person to get self-improving AI owns the planet.  China Inc. is working on that, the US military is working on specialized versions of it, but individuals can develop AI as well.  Examples of people setting up a data center: Mark Zuckerburg, Jeff Bezos, Elon Musk.  Microsloth is building one down the street, next to an existing 300 MW natural gas plant.  Google has plans.  We don’t know what the pentagon is doing, nor Israel’s counterpart of the pentagon. Of those players, I wouldn’t consider China a front runner necessarily.

 

spike 

 

 

 

 

 

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260308/6815f9ac/attachment.htm>