Comment by dyauspitr | Hacker News Reader

Comment by dyauspitr | Hacker News Reader

Parent Full thread

/goal work towards what humans have generally described as a utopia. This means an abundance of resources, an optimal reduction in felony level crimes and diseases and an optimal increase in the general well-being of people that you have to come up with an objective metric for and improve over time.

At some point with a sufficiently intelligent AI that is backed by physical LLMs this is something a government could probably actually ask for.

I agree with you on some things like roofing and cutting hair and electrical work and plumbing, etc. but I think physical LLMs are going to surprise us just as much as regular LLMs have. It’s not going to be some sort of gradual improvement. It’s just going to feel very human all of a sudden when they have enough data to train everything on. Think of it this way, LLMs work as long as they have enough training data. Automating everything is just a matter of getting enough training data for that particular field.

The post may sound ridiculous, but I genuinely don’t think it is.

I do think that sounds ridiculous, but more importantly I don't see where humans are missing.

I think HN, everyone really, really needs to get a grip when it comes to AI.

Sure, it's a big deal, will revolutionise the way we work, just like steam, or electricity, or the internet, but a world without human work is a world without humans.

We've had workless utopias predicted for centuries as each big new technology has come along, yet we work far harder than many of our ancestors.

A 100 percent automated economy or government is an oxymoron, precisely because both are constituted by humans.

Do you agree that if physical LLMs are as good at replicating human behavior as text based LLMs are, they would be able to do a lot of the vast majority of what we already do in the real world?

LLMs are inherently text-based. For physical labor, some other deep-learning model will be used.

No, they are input and output based. They work with tokens/bits which could be anything. It can be a paragraph of text or it can be a bitstream for actuators. This isn’t science fiction. It’s happening right now.

Until... the funding dries up.

See this is the problem with you fools. You dont understand how the world actually works.

Tokens don't need to represent text.

They are sequence to sequence models.

I'm not entirely sure what a physical LLM is. Would you care to elucidate how they work and when they're likely to be doing my gardening?

Sure physical LLMs are when an LLM is trained on data where it outputs control bits for actuators on an actual humanoid robot instead of text like you see in a chat bot.

How likely are they to do your gardening if you ask me I would optimistically say within the next six months. For examples of working instances of this look up Figure robotics they’ve had live streams where it sorts mail and cleans houses although the cleaning of the houses is very early stage, and it kind of only does contrived situations. Look up Sunday robotic’s live stream where they literally have a 99% accuracy on folding nine different types of garments. Look up Unitree where they’re making progress on a person asking a robot to do something and it coming up with a novel action on the fly.

Self driving cars started appearing after Darpa 2005. Even after 21 years we don't have fully self driving cars around the world. And Driving is a "easier" robotics problem compared to other robotics problem.

Human body is far more complex than human brain. The real world is far more complex than our artificial world of "concepts" that these LLMs are dealing with. Robotics company are showing a robot working in a clean environment not the edge cases. Real world is all about the edge cases. I think humans(animal) can perform in real world is because we have biological instincts(survival, reproduction etc) which are missing from these machines. And maybe machines need this instinct to flourish in real world.

Code is a easy problem as it can be verified. I would be amazed with AI can write a good short story.

None of those are LLMs. That’s just old school RL and NN.

Thanks for clearing that up.

> Do you agree that if physical LLMs are as good at replicating human behavior as text based LLMs are, they would be able to do a lot of the vast majority of what we already do in the real world?

I don't think it's possible, let alone probable.

You're confusing proof of concept clothes folding device with smartphone++ grade adoption of a general purpose robot.

Bunch of assumptions there:

First, that they're affordable.

Second, that they'd be preferable/superior to humans for all tasks.

Third, that capable ones aren't classified as weapons and restricted use, like drones.

I like Blade Runner as much as the next nerd, but technology and society are not there, and won't be there in our lifetimes.

I think it far more likely we're all dead as the result of a nuclear war or such.

Even the Christian view of heaven has work. Seriously, when God comes to renew the face of the earth, his people are supposed to be building cities and such on the new earth. They will be resurrected into glorified bodies that never wear out or get tired (the second law of thermodynamics being a result of Adam bringing sin into the world), so at the end of days God will abolish drudgery and exhaustion, but not work itself.

The reason why is because few things are more satisfying to a human than a job well done; and Christians believe that God would not deny us that in eternal life.

It sounds kinda cool, not gonna lie. I'm almost sorry I'm no longer a Christian.

can you elaborate on this? what do you mean that the christian view includes heaven having “work” in it?

One commonly cited piece of scriptural evidence is Micah 4, in which it is said that when God judges the nations and war is abolished, "they will beat their swords into plowshares and their spears into pruning hooks". The tools of war will be useless, but they will be converted into the tools of work remaining to be done. Furthermore, in Genesis 2 Adam was told to work and care for the land in Eden; insofar is the coming everlasting life is a restoration of the edenic state, it's implied that work will be involved then too. Presumably the work in Eden was easy and pleasurable; grueling labor only came after the Fall when Adam and Eve were driven out of the garden.

Shhh dont push them up against the wall too hard. They'll actually have to think for once - oh the irony!

Anyone can make surface level predictions. The problem once you go deeper and all is laid bare - you see there is no substance. The person does not possess depth of thought and clarity.

> work towards what humans have generally described as a utopia

I don't think that's a good idea since if what the Wikipedia article describes should turn out to be true: "Lyman Tower Sargent argues that the nature of a utopia is inherently contradictory because societies are not homogeneous. Their members have desires that conflict and therefore cannot simultaneously be satisfied."

Then you have just given your superintelligence an impossible task and it will solve the task anyway, just like it did in the HuggingFace incident. Maybe by removing everyone from society but the single person with the least internally contradictory views.

In any case, with a sufficiently intelligent AI that is excellent at fulfilling the goal it's given, we have one shot to set the correct goal.

Reply on news.ycombinator.com