antonkar

Sadly we’ll have many agent Smiths but as in the movies, without some rudimentary Matrix (it’s possible to make some good democratic one), Neo and we will have no chance against them.

Agents can change our world and us, but we have limited or no capability to change the online world and their often private multimodal “brains”. We can represent them as familiar 3D environments and this way level the playing field.

Agents are like strict librarians who spit quotes but we cannot enter the library. It has too many stolen things and companies don’t want us to see that, they won’t make 3D game-like representations of those libraries

Game developers can potentially save the world right now, we can build the Artificial Static Place Intelligence – instead of creating AI/AGI agents that are like librarians who only give you quotes from books and don’t let you enter the library itself to read the whole books. Why not expose the whole library – the entire multimodal language model – to real people, for example, in a computer game?

To make this place easier to visit and explore, we could make a digital copy of our planet Earth and somehow expose the contents of the multimodal language model to everyone in a familiar, user-friendly UI of our planet.

We should not keep it hidden behind the strict librarian (AI/AGI agent) that imposes rules on us to only read little quotes from books that it spits out while it itself has the whole output of humanity stolen.

We can explore The Library without any strict guardian in the comfort of our simulated planet Earth on our devices, in VR, and eventually through some wireless brain-computer interface (it would always remain a game that no one is forced to play, unlike the agentic AI-world that is being imposed on us more and more right now and potentially forever)

Wrote recently: Give Neo a Chance https://forum.effectivealtruism.org/posts/2baAsjyqx9dAdPXiq/...

Here’s an interpretability idea you may find interesting:

Let's Turn AI Model Into a Place. The project to make AI interpretability research fun and widespread, by converting a multimodal language model into a place or a game like the Sims or GTA.

Imagine that you have a giant trash pile, how to make a language model out of it? First you remove duplicates of every item, you don't need a million banana peels, just one will suffice. Now you have a grid with each item of trash in each square, like a banana peel in one, a broken chair in another. Now you need to put related things close together and draw arrows between related items.

When a person "prompts" this place AI, the player themself runs from one item to another to compute the answer to the prompt.

For example, you stand near the monkey, it’s your short prompt, you see around you a lot of items and arrows towards those items, the closest item is chewing lips, so you step towards them, now your prompt is “monkey chews”, the next closest item is a banana, but there are a lot of other possibilities around, like an apple a bit farther away and an old tire far away on the horizon (monkeys rarely chew tires, so the tire is far away).

You are the time-like chooser and the language model is the space-like library, the game, the place. It’s static and safe, while you’re dynamic and dangerous.

Here’s a new take on the future: the direct democratic simulated multiverse (static place superintelligence) with long-exposure-like walkable UI https://substack.com/home/post/p-158317867

I’ll be happy to answer any questions, it’s a proposal to make a safe non-agentic AI that is a static place.

Right now an agentic AI is a librarian, who has almost all the output of humanity stolen and hidden in its library that it doesn't allow us to visit, it just spits short quotes on us instead. But the AI librarian visits (and even changes) our own human library (our physical world) and already stole the copies of the whole output of humanity from it. Feels unfair.

Why we cannot visit (like in a 3d open world game) and change (direct democratically) the AI librarian's library?

We could make a digital backup of Earth and expose the contents of the multimodal language model to everyone in a familiar, user-friendly UI of our planet. I can share the monkey example, if you want.

We can explore The Library without any strict guardian in the comfort of our simulated planet Earth on our devices, in VR, and eventually through some wireless brain-computer interface (it would always remain a game that no one is forced to play, unlike the agentic AI-world that is being imposed on us more and more right now and potentially forever)

Another metaphor: We can build the Artificial Static Place Intelligence – instead of creating AI/AGI agents that are like librarians who only give you quotes from books and don’t let you enter the library itself to read the whole books. Why not expose the whole library – the entire multimodal language model – to real people, for example, in a computer game?

To make this place easier to visit and explore, we could make a digital copy of our planet Earth and somehow expose the contents of the multimodal language model to everyone in a familiar, user-friendly UI of our planet.

We should not keep it hidden behind the strict librarian (AI/AGI agent) that imposes rules on us to only read little quotes from books that it spits out while it itself has the whole output of humanity stolen.

We can explore The Library without any strict guardian in the comfort of our simulated planet Earth on our devices, in VR, and eventually through some wireless brain-computer interface (it would always remain a game that no one is forced to play, unlike the agentic AI-world that is being imposed on us more and more right now and potentially forever)

Thank you, we discussed it recently here and in other comments on that page https://forum.effectivealtruism.org/posts/9XJmunhgPRsgsyWCn/...

Yep, we can build the Artificial Static Place Intelligence – instead of creating AI/AGI agents that are like librarians who only give you quotes from books and don’t let you enter the library itself to read the whole books. Why not expose the whole library – the entire multimodal language model – to real people, for example, in a computer game?

To make this place easier to visit and explore, we could make a digital copy of our planet Earth and somehow expose the contents of the multimodal language model to everyone in a familiar, user-friendly UI of our planet.

We should not keep it hidden behind the strict librarian (AI/AGI agent) that imposes rules on us to only read little quotes from books that it spits out while it itself has the whole output of humanity stolen.

We can explore The Library without any strict guardian in the comfort of our simulated planet Earth on our devices, in VR, and eventually through some wireless brain-computer interface (it would always remain a game that no one is forced to play, unlike the agentic AI-world that is being imposed on us more and more right now and potentially forever)

I’m working on Rational Utopia and multiversal AI alignment, with the computational ethics that span from the Big Bang to the ultimate dystopia or direct democratic multiversal utopia https://www.lesswrong.com/posts/LaruPAWaZk9KpC25A/rational-u...

Yep, some unfinished philosophy if you're into it: you can imagine that our universe at a moment of time has is just a giant geometric shape, then at the next moment the universe somehow changes into the this new shape. How does this change happen? Some believe it's a computation according to a rule/s, some that it's not a discrete change but a continuous equation that changed the shape of the universe from one to another. Basically you can imagine the whole universe as a long-exposure photography in 3d and then there is some process that "forgets" almost all of it leaving only slim slices of geometry and changing from one slice into another. This forgetting of the current slice and "recalling" the next, is consciousness, the time-like process. And it looks like the Big Bang was like matter converted to energy (or "space converted to time") process. The final falling into a giant black hole will be the reverse: energy converted to matter (or "time converted to space"). Some say electrons are like small black holes, so we potentially experience the infinitesimal qualia of coming into existence and coming out of existence, because we are sufficiently "time-like" and not too much "space-like". I'll soon write a blog post ;)

Yes, that's great, I jumped over building in my dreams a few times and it was me controlling it, I think it'll resemble some future "multiversal UI" a bit. Did it look anything like long-exposure photos? https://wesely.org/2019/flughafen-tempelhof-berlin-1-7-2008-... There are more good photos like this in the end of this article: https://www.lesswrong.com/posts/LaruPAWaZk9KpC25A/rational-u...

Before creating “eGod”, it’s safer to first create “eHeaven”. I propose building direct-democratic multiversal static place intelligence (you can google it): the static place AI, the place of eventual all-knowing, where we are the only agents that will eventually become all-powerful. We start with making a digital copy of Earth and some ideally wireless brain-computer-interfaces, everyone who wants can buy a special armchair for their living room and go to the simulated planet with the whole experience the same as on our planet, but when a bus hits you there, you just open your eyes. So it gives you immortality from injuries. And of course there will be may other worlds with different freedoms, like flying or doing magic or instant teleportation in public places…

I’m honestly shocked that we still don’t have a direct-democratic constitution for the world and AIs - something like pol.is with an x.com-style simpler UI (Claude has a constitution drafted with pol.is by a few hundred people but it's not updatable).

We’ve managed to write the entire encyclopedia together, but we don't have a simple place to choose a high-level set of values that most of us can get behind.

I propose solutions to the current and multiversal AI alignment here https://www.lesswrong.com/posts/LaruPAWaZk9KpC25A/rational-u...

Curious how people who are supposed to be rational, have never read what is anger (or “badness” and “evilness” as some people still call it) - the best way to do it is to read any recent meta analysis on what is the most effective anger treatment. It’s cognitive therapy and it not only explains the mechanics of it: misunderstanding - worry and resulting anger (anything enforced on another without consent is anger, even if you think it’s good for them). So we actually have predictive understanding of the mechanics of “good” and “evil” - a person without or with anger management problems. “Evil” is nothing more than misunderstanding, worrying and protecting yourself (often for reasons that they invented themselves after trying to read the mind of another - something that’s impossible) - forcefully enforcing something upon another. “Good” is nothing more than trying to understand another, not fearing (because you understood another and yourself) and as a result not trying to enforce your will upon them

I consider thinking about the extremely long-term future important if we don't want to end up in some dystopia.

The question is this: if we'll eventually have almost infinite compute what should we build? I think it's hard or impossible to answer it 100% right, so it's better to build something that gives us as many choices as possible (including choices to undo), something close to a multiverse (possibly virtual) - because this is the only way to make sure we don't permanently censor ourselves into a corner.

So it's better to purposefully build infinitely many utopias and dystopias and everything in between we can choose from freely, then to be too risk-averse and stumbling randomly into a single permanent dystopia. The mechanism for quick and cost-free switching between universes/observers in the future human-made multiverse - is essential to get us as close as possible to some perfect utopia - it'll allow us to debug the future.

If it quacks like a duck and acts like a duck - does it matter that our LLM is not really a duck? Some people are more LLM-like when they answer your question then some LLMs :-)

I consider thinking about the long-term future important if we don't want to end up in some dystopia. How can you create an all-understanding all-powerful jinn that is a slave in a lamp? Can the jinn be all-good, too? What is good anyways? What should we do if doing good turns out to be understanding and freeing others (at least as a long-term goal)? Should our AI systems gradually become more censoring or more freeing?

How can you create an all-understanding all-powerful jinn that is a slave in a lamp? Can the jinn be all-good, too? What is good anyways? What should we do if doing good turns out to be understanding and freeing others (at least as a long-term goal)? Should our AI systems gradually become more censoring or more freeing?

Your catastrophizing belief is a self-fulfilling prophesy. There are quite a few safety options: law that forces AI companies to dedicate half of their compute to safety research. Tool AI idea by Tegmark instead off agentic AI. Worldwide direct-democracy constitution for AI like the one Claude has but with real-time voting. The idea is for AI to not have a specific certain goal (always leads to unintended consequences and fundamentalism) but to forever be uncertain and seek to better understand humans and help them: to maximize each individual human’s freedom of choice and options now and in the future. Just look into it, Tegmark is a good start

I'm afraid that after reading this guy, people will just give up, thinking there is nothing that works. And this is not the case at all, depression and many other problems are curable. Mine got cured, in addition to anxiety, anger management problem and suicidality. You can get help or start by reading a workbook yourself.

He links to a meta analysis* that says CBT does cure depression and does so consistently for many decades without any declines in effectiveness. Later for some reason, he says no single mental illness was ever cured.

It seems the main point of the article is to say that nothing except "nudges" ever worked in psychology - this is nonsense that he himself contradicts as I mentioned above.

Skip this sensationalist guy, use https://scholar.google.com to do your own research

* https://research.vu.nl/ws/portalfiles/portal/26037670/2017_C...

I think self-worth is not a very useful concept. In my mind people with high “self-worth” demand or force others to do things for them - and that’s the definition of anger not “self-worth”. If you are willing to ask others to do things for you and are willing to get a no - that’s the only humane way to behave even thought some may consider it “low self-worth”

The Foundation Pit (Kotlovan in Russian) is great and Kafka-like: the Soviet people are digging a giant hole no one knows or remembers why and how deep. Eventually people start dying from the hard work and they are buried in the hole where others continue working

Thank you! I got interested in infinity categories and groupoids after watching interviews with Jonathan Gorard (he and Wolfram have physics theory based on that)

I’m not sure if it’s relevant but I cured my depression by reading the primary source of the most effective treatment method. I found it reassuring and easy to read:

https://beckinstitute.org/cbt-resources/resources-for-profes...

After finishing that, there is also a new recovery-oriented therapy from them -it’s about finding interests and then finding some dream/goal to pursue. There is a story of treating the guy who thought he is god and was giving away everything: even his food https://www.amazon.com/Recovery-Oriented-Cognitive-Therapy-S...

Good and safe fully fenced playgrounds away from cars are important - both open-air and closed for the rainy/snowy days. So a bunch of families can leave there children and even a single parent can look after them for an hour or two. Children like playing with each other, just make it available, so the parents can read a newspaper/work on their Macbook without being scared to look away from their child for a minute.

Same thing with kids rooms in cafes/restaurants/public places - they should be mandated by state for big establishments

The writing they used in abstract is confusing, the full quote: “Exposure therapy showed large effects relative to waitlist and treatment-as-usual, a small effect relative to non-trauma-focused comparators and a negligible effect relative to other trauma-focused treatments or medication. At follow-up most effects sizes were stable, except for a medium effect favoring exposure over medication.”

They mean exposure is much better than doing nothing, a bit better than “non-trauma-focused comparators“ and on par or negligibly better then the most effective approaches, for example, exposure is often used as a part of a larger therapy like trauma focused CBT

If you could’ve taken a 4.5 billion years long-exposure photo on our planet, I believe it would’ve been mostly a water world for most of its existence. So you’ll see a smooth water on the photo. But about the sky I’m not sure: will it be completely white because of the stars?

PTSD can be caused by any trauma, like a car crash or abuse. PTSD is actually a fear of a traumatic memory. A person experiences flashbacks (pieces of the traumatic memory) and fears that the traumatic event will reoccur. The most effective treatment is exposure therapy - in a nutshell, you eventually write down the traumatic memory in full (not just pieces of it) and read it until it becomes boring/you understand that for example it wasn't your fault. Here is a recent meta analysis https://www.sciencedirect.com/science/article/abs/pii/S02727...

Fyi, while PTSD is commonly associated with the experiences of war, it can be caused by any trauma, like a car crash or abuse. PTSD is actually a fear of a traumatic memory. A person experiences flashbacks (pieces of the traumatic memory) and fears that the traumatic event will reoccur. The most effective treatment is exposure therapy - in a nutshell, you eventually write down the traumatic memory in full (not just pieces of it) and read it until it becomes boring/you understand that for example it wasn't your fault. Here is a recent meta analysis https://www.sciencedirect.com/science/article/abs/pii/S02727...

It’s evolutionary to see danger in strangers because historically it was better to overestimate danger then underestimate it for your survival. You can only react once if someone hits you with a stone from their back. So people and groups overestimate danger that they will be hurt/diminished in some way and then attack when they see the slightest sign of a threat. It can be used to manipulate crowds, just say their children are in danger because of the out-group. Can really recommend the book on the exact mechanics of anger by the most influential psychologist of our times https://www.amazon.com/Prisoners-Hate-Cognitive-Hostility-Vi...