Divine works


eng deu рус


FLI error:

(Open Letter to the Future of Life Institute (FLI) based on his open Letter " Pause giant AI experiments " suggesting a 6-month pause in AI experiments )

The mistake of the initiators of the FLI Letter is not that the use of AI poses the greatest threat to the existence of Humanity. All other threats - biotechnology, nuclear weapons, climate change and "are not lie near", are relatively minor.

The error is not that a 6-month pause can be very useful:
1. The pause will indeed temporarily reduce competition, which will allow attention to be paid to safety:
“Laboratory leaders themselves would like to take a break, but feel the pressure of competitors.
A public commitment by all laboratories to pause will reduce this pressure."
2. Not just a pause is proposed, but a period of very intensive development of AI security measures.
The initiators of the FLI Letter offered specific directions - the excellent "Policy recommendations" - for their development during the proposed pause.

The mistake of the initiators of the FLI Letter is the absence of the main policy - "political".
"Policy recommendations" assume the negotiability of participants.
The main problem is that developers in authoritarian states (Russia and China) will be deliberately and secretly encouraged to violate the "Policy recommendations".

Reasons for the success of precedents:

Previous international agreements to limit the development and spread of technologies in the fields of chemical, bacteriological, nuclear, genetic weapons have been successful due to the properties of these weapons:
1. Disprofitability:
For example, Hitler did not start chemical warfare, since Germany's losses in it would have been greater than the losses of the Allies.
After the 2nd World War, chemical weapons for "deterrence" turned out to be less effective than nuclear ones.
2. Non-selectivity:
For example, a genetic weapon that would kill Americans, but not Russians, could not be created due to high mixing haplogroups
3. Verifiability:
Storage bases for nuclear, for example, weapons are quite reconnoitered.

However, AI manipulation weapons do not have all of these properties.

"Policy recommendations" may well be effective enough to prevent the threat - in democratic countries.
But they will by no means prevent the main threat - the deliberate development of AI psychic weapons - in authoritarian states.

AI weapons are widely used - now.
- Kremlin propaganda is using Big Data and "Deep" technologies not only against Russians, but also against democratic countries.
Moreover, already today the main weapon of the Kremlin is propaganda, not tanks.
- Moscow's investment in propaganda has grown over the past 2 years several times more than investment in conventional weapons.

Scenario of the beginning of the End of Humanity:

To fool 50+% of the US population, you don’t even need Super-AI, superior to a person in intelligence.
1. AI psychic weapons will not be automatic, but automated.
That is, the AI-psychic weapon will not act independently, but it will be used as a tool that increases the productivity and efficiency of the "labor" of highly intelligent people propagandists and political-technologist by an order of magnitude.
2. "Managed democracy" is more beneficial to the US establishment than true democracy.
3. "Checks and balances" are no longer sustainable in the digital age.
4. 50% of the US population are not sufficiently reasonable people, and, like the Russians in 1993, they can support what is harmful to them.

After zombifying 50% of the US population, it will even be possible not to storm the US Senate, but to bring a pro-Russian government to power - democratically.
And already the pro-Russian US government will deploy "Policy recommendations" in the opposite direction.

 

Berezin Commentary on his Criticism of the FLI Letter on Pause in AI Experiments:

The Letter from the Future Life Institute states:
"Should we let machines flood our information channels with propaganda and untruth? 
Should we risk loss of control of our civilization?" 
( https://futureoflife.org/open-letter/pause-giant-ai-experiments/ )

Berezin objects ( https://naked-science.ru/article/nakedscience/kontrol-nad-demonami ):
"Newspapers with 'propaganda and lies' are already full.
ChatGPT won't be able to mess up the world around us for the simple reason that even before ChatGPT the planet was "under-equipped for gladness."
However, the "corruption" of the world does not mean that it cannot be spoiled even more, and in the end completely destroyed.

The threat from AI propaganda - in the near future will not be that people will be manipulated - full automata, but in the automation of demagogy, that:
1. "This is an excellent tool" that will allow you to issue more propaganda per unit of time. But most importantly -
2. This "tool" will raise the "quality" of propaganda.

Berezin contradicts himself:

Berezin writes:
"The production of journalists will be of 2 types:
- Low quality ChatGPT output , and
- Written by people, by hand."

But he also writes:
"The level of most journalists is so terrible that ChatGPT will even improve everything here."

Un-intelligenceable AI:

Berezin argues that ChatGPT is no smarter than a slime mold that has no nervous system at all.
However, it doesn't matter if the AI is self-aware or not. What matters is the result.
- AI is a very effective tool of demagogy and its effectiveness in this capacity is growing exponentially.
It is not necessary to have self-consciousness, it is enough to emulate it.

Berezin himself admits that systems like ChatGPT are already now improving the "quality" of Russian "news" (propaganda).
The same looms mentioned by Berezin not only increased production volumes, but also made it possible to produce more complex products.
And by 2030, the mentioned tomatoes will also be harvested by machines already - without people. Moreover, the development of recursive AI is accelerating.

Yes, until 2060, so far, recursive AIs will not be able to surpass intellectuals like Berezin, they will not be able to compete with humans in the production of high-quality mental products.
But this is not required:
1. AI tools will help human demagogues increase their efficiency - their demagogy, which, like the products of "professional translators", can no longer be called " Hand made ".
2. In propaganda, quantity is often more important than quality. -
2.1. For the inhabitants of the same Russia, most of whom believe in sorcerers, high-quality deception is not required.
2.2. The amount of cheap propaganda will make it possible, for example, to create for the inhabitants of Russia - "Virtual World" - a completely closed information environment in which black is white, Nazis are anti-Nazis, and anti-Nazis are Nazis, terrorists are anti-terrorists, and anti -terrorists are terrorists, and the enemies of the people are their benefactors.
3. Individualization of demagogy:
You can't put a FSB-agent or a frauder-psychologist to every Russian person.
However, by 2030, individual information environments will be used, similar to targeted advertising today, but also using "big" data about the behavior and statements of a person outside the Internet.

The threat lies precisely in the fact that this is only the beginning of the "process".
The pre-end of the process is 100% control of the people by the " symbiotes ";
0% real freedom of choice in people, "humane", but "Brave New World".
The end of the process is the provoked death of Mankind because of its problematic nature and uselessness for the almighty "gods", in which rulers evolve as a result of natural selection for unscrupulousness and selfishness. Peskov and Prigozhin are only remotely similar to the monsters of the future.

Berezin is right that:

1. The threat does not come from AI itself, but from the people who use it, the threat from AI is only as a tool.
In the 18th century, too, it was not the looms that were to blame, but the people who ordered people to cut off fingers, and the social organization of people.
2. There is only 1 threat from AI - propaganda, manipulation.
There really is no threat to jobs.
AIs will "replace us" not because they will ever surpass us in empathy, but because of our side qualities that AI do not have, because we are bad, capricious slaves, and bad property is destroyed.

There is no threat from other nations too:

Berezin writes that "it's short-sighted to prevent your own 'pseudo-demons' from arising, if as a result you have to fight off strangers."

But colonies in the 21st century are not profitable.
NATO does not threaten the Russians, but only the interests of their enemies - the usurpers of power.
On the contrary, for the Russians themselves, the European path is beneficial.

AI as a weapon has a fundamental difference. - Uncontrollableness by the people.
By the intellectual "weapon" it will be beneficial to influence precisely the intellect of all people , and their social organization of security.
And the result of the AI-manipulation race will not be the subjugation of Europeans by the Chinese, but first the subjugation of both of them by the "gods" of life, and then the death of both of them.

Further, the authors of the letter from the "Future of Life Institute" do not suggest self-renunciation of AI in the face of Putin's call of 09/01/2017 to become "ruler of the world", but offer international control over new technologies, which was impossible in the 19th century, but is possible in the 21st -m.
True, for this, it may not be possible to do without the destruction of propagandists, corrupt developers and "rulers of the world" - today's.

Unfortunately, Berezin also talks in the narratives Kremlin's terrorist not controlled by society.
The division into "us" and "them" today objectively lies not between nation, but between people and usurpers of power.

International Control of AI:

Berezin denies the efficiency of international agreements.
They say, as in the ancient world, today the "law of the jungle" is in effect: "If you do not kill first, then they will kill you"
However, FLI provides an example of precedents for successful international interaction.
Asilomar Conference on Recombinant DNA allowed scientists and government experts to ban certain experiments and develop design rules that allowed the technology to be safely explored.
Other examples: Reducing and limiting the spread of nuclear weapons, prohibition of chemical, biological.
That is, today international control is quite efficient.
True, there really is one exception - a state that continues to live according to the "law of the jungle": This is Russia. -
Russia has withdrawn from many international agreements and does not recognize the priority of international law and courts.

Opportunity of survival of Mankind:

In the digital age, any government is unsustainable,
but at the same time, only in the digital age is self-government without a government possible.
In the digital age, all the few issues that affect many people can and should be decided not with the participation of people, but without the participation of the government - only by the people themselves.
However, for this, a very complex, but flat, with free access, Deliberative structure can and should be created, which also uses AI, but:
- 100% transparent.
- not controlled by anyone, but working according to open Protocols developed by the Deliberative structure too.

The French should not demand in the squares either pensions or a ban on roads, anything - they must decide everything responsibly - themselves, but not with the help of primitive and ineffective referendums, but with the help of the Deliberative Structure.

Production Control Theory VS Cost Reduction Theories:

The reason for industrial automation was not so much the much lower cost, but the ability to control production to centralize profits.
In the 18th century, the cost of machine-made fabrics was almost no lower than manual production - the production of means of production was still expensive.
However, it was cheaper for the British oligarchs to control a 100-worker weaving factory than 10,000 objectively much more autonomous manual makers – only 10 overseers were needed, not a hard-to-manage pyramid of 1,000 terrorists.
When the cost of textiles really fell in the 20th century, due to increased competition, the British oligarchs moved from weaving to the "intellectual" industries of new developments, which they already control with the help of "intellectual property" rights.
Now intellectual labor is manual, but is rapidly mechanized.
The threat to Humanity is actually a by-product of the mechanization of intellectual labor - the greatest threat compared to other side effects - climate and others.

"Policy recommendations" from FLI:

( https://futureoflife.org/wp-content/uploads/2023/04/FLI_Policymaking_In_The_Pause.pdf )
1. Mandate robust third-party auditing and certification.
2. Regulate access to computational power.
3. Establish capable AI agencies at the national level.
4. Establish liability for AI-caused harms.
5. Introduce measures to prevent and track AI model leaks.
6. Expand technical AI safety research funding.
7. Develop standards for identifying and managing AI-generated content and recommendations

Alarmism?

Yes, maybe we are vain alarmists - we overestimate the threat.
However, at stake is the greatest thing that can be - the Meaning of the Universe.
Therefore, even a small risk is unacceptable.
At a minimum, the risk needs to be clarified. Do you agree?
And since the risk is associated, among other things, with the government agency of security that are supposed to deal with risks, an independent risk assessment is needed. Do you agree?
And since the threat is very complex, a complex adequate public organization is necessary to assess the risk. Do you agree?
"Deliberative network" we propose as such.

Why do those in power so easily fool people even before strong AI?
- Because people are busy with personal everyday problems. They are only smart about this.
That is why it is necessary to create a structure for a flat solution - all issues - independently, without rulers.
See You have to rely on yourself.docx # The Need for independence

We are against AI-alarmism:

It can even be political manipulation by scaring ordinary people away from AI.
AI itself is almost safe compared to its use by man against man, people.
On the contrary, we should welcome the use of AI, but - flat, distributed - DAI.
Only DAI will be able to counter the centralized manipulative AI.

" Konnor", http://dw2020.narod.ru

 

Death of Freedom:

Humanists are wrong - people are creatures to a greater extent - meaningless, selfish and vicious.

Ideologies, philosophies, religions give meaning to existence, purpose, direction.
Among the followers of some ideology, even outward behavior differs in contrast from the aimless life of nihilists.

However, as a rule, ideologies, like religions, are dogmatized. – The idea proposed by the founders becomes conservative.
This is the result of the work of the already social mechanism - the idea that unites more people survives. Mostly hierarchical organizations won.

Despite all this, there has until now been freedom to create and disseminate ideas, in particular, the freedom to search for the meaning of existence.

Soft dictatorship:

Even if humanity does not die - physically - in the next 300 years - as a result of a jump in the possibilities of manipulating people, then freedom will die. -
People will remain quasi -free - they will have a sense of personal freedom.
But they will no longer have the opportunity for the social realization of their freedom.

This process has already begun - even before the creation of a strong Artificial Intelligence.
Compare the spread of ideas 100 years ago and today. - Today, people are successfully fooled by propaganda in Russia and distracted from "unnecessary" views - in the West - even in the conditions of rapid and free growth of formally independent social networks. networks. – Opinion makers use all newly emerging technical possibilities – ahead of schedule .

Example - Our idea of "Flat democracy" in the era of social networks, noise generation and active jamming in them - "successfully" fades out not only without finding supporters, but generally remaining unknown.
Paradoxically, during the oral dissemination of opinions, her chances of survival would have been much greater.

Death of Freedom:

Yes, true freedom will remain - with individual " symbiotes ".
But they will not be motivated to seek answers to the "eternal" questions that have accompanied Mankind throughout its history.
- The free search for answers to social questions has always been inextricably linked with social processes, and "feeds" only on them. -
"Gods" will not need unpredictability.
With the death of society, social spirituality will also die.
"Humanity" if it survives physically, then only physically.

The death of "God":

If something like God exists, it probably isn't self-sufficient.
If the development of Mankind is not 100% accidental, if, for example, the previous "Civilization" contributed to the emergence of our Universe with the necessary parameters (see the version of the " Relay Race of the Spirit "), then the existence of a certain Humanity is an aspect of the meaning of the existence of "God" too.

2302