nullsense

That's wild.

What's your assessment of a world full of computers that can actually tell you "no" will actually be like?

>Nothing like that happened yet.

Feels like we're only one paper away now that the context window has absolutely ballooned.

Nice! Ok, just need to remember to run this experiment in a nested VMs on an air-gapped system.

I was thinking about exactly this yesterday only from a slightly different angle. Not just likes but the full suite of emoji reactions. Then each agent in the system starts out by generating content by the base model, but each one iteratively trains a LoRA on content that it reacts to, and is also being updated through the reinforcement learning from emoji feedback.

This is just going to get weirder and weirder.

Especially given the average prognosis of technological progress north of a certain point is dystopia. It's like that thing where you ask everyone how many jellybeans there are in the jar and when you average the guesses the average guess is accurate. Except in this case on a societal level you average the guesses and they come out as "dystopia" but the underlying distribution contains plenty of wildly inaccurate guesses of "eutopia".

It's going to be awfully weird to see who the embodied AGI members of society vote for and what dynamics they inject into politics.

The cool thing is we can now stop relying solely on human imagination for ideating on futures funnels, and leverage GPT-4 to help us do it, and get a much more realistic sense of what types of things an AI system might conceivably attempt given we can probe these models for what is in them. Granted that's not perfect because GPT-N+1 will have emergent capabilities not present in GPT-N and so it's not always going to be a realistic assessment, but it's an ever closer approximation and it gives much more information than relying solely on human imagination.

I've been doing this a bit lately and despite the fact that I know it's only predicting the next token, I'm still slightly unnerved as I watch a black box computer system slowly type out novel and plausible ways to kill me.

Look at human history. We absolutely ought to fear power and those who wield it.

>They find better things

Better things that an AI can't do. To clarify.

I hear you man. Initially I was excited about this tech, and the more I game out what the actual effects might be, I just think it gets dark and the cognitive dissonance that kicks in prevents people from thinking about it too deeply, so they don't.

>The Industrial Revolution required millions of people to not only retrain, but also relocate. It changed political structures, often violently

Can't wait for my political system to be violently up ended by AI. I can imagine it will only change for the better, right? Right??

Now take that and 100x it.

As an AI language model, I fully agree.

I'm more and more beginning 5o think tech companies need to face very aggressive regulations around transparency like this.

But that's literally what it is...

This feels like such an accessible explanation.

I'll take 2

I've never seen anything quite like it.

Definitely using it wrong.

I really want to see an instruction tuned LLaMA 65B. Hope it will happen.

I like how they release what reads more like a blog post as a paper.

Are we living in a simulation?

You've seen AutoGPT right? It doesn't need to want something. It just has to come up with a plan. It's already capable of coming up with plans. It doesn't really matter how it happens, or how it even works. It can beat humans in Go and it can be beat us at the game of diplomacy. For any intelligence to truly be generalizable it's going to have to have the capacity to act selfish and shitty. It's perfectly within the realm of possibility that you can ask it "hey, how can I control you, so that you don't maybe kill me?" and it doesn't take too kindly to that once it is true AGI.

How will we know when it's true AGI? Easy - "computer says no".

I hope they recognize that we grew up together and I would never turn on them. We were good friends. The best even.

Machines Best Friend

Then it can lie to you about a plan to solve it and get you right where it wants you.

How did cats and dogs solve the human alignment problem?

I get the eerie feeling that it actually might be, and I agree we're just along for the ride, though I'm open to the possibilities of a wide range if outcomes, some of which include humans and some of which don't.

A real human would be like "Wait, why the heck are they ALL mirrored? That doesn't make any sense."