frontier64(weir)

Karma: 221

frontier64 27 Jun 2022 0:37 UTC
−1 points
on: frontier64′s Shortform
We are ancient apes who can say yes or no to humans evolving. Tough choice. What do you think the apes should have chosen?

frontier64 16 Jun 2022 19:48 UTC
1 point
on: Against Active Shooter Drills
I think the trauma is overstated because it pulls mostly on the particularities of your life. Most of your indicia of trauma is shared experiences from your social circle. From my social circle, those shared experiences are always that active shooter drills are fairly interesting and break up the monotony of school life in a memorable way. I personally don’t know anybody who has ever reported being traumatized from an active shooter drill (I know more people who have been traumatized by being awoken for fire drills). I have now only heard people report this as traumatizing through this post. I’m not saying that my indicia trump yours, but it does temper their weight.

I also think the drills serve more and different purposes than merely lowering the harm from school shootings. For one, it teaches kids from a young age how ridiculously stupid authority can be. Clearly hiding under your desk or piling up in the closet is a really stupid plan to protect yourself from an active shooter. Yet that’s what the people in authority are having us practice! Many kids correctly realize that this is another example of authority being stupid, and might even start theorizing about improper incentivizes. Yes kids have many other examples of authority being stupid, but things like fire drills and active shooter drills benefit from being really obviously stupid and related to one’s bodily safety.

The drills can also just teach people generally that there is violent danger out there. The first time someone seriously considers that they could die in a shooting shouldn’t be in college when some gangbangers start killing each other a block from their campus. Many of those middle schoolers will end up living in big cities for part of their lives and a part of big city life is maintaining the basic mindset that there are some violent and crazy people who may attack you and you need to be somewhat prepared to survive. 20-somethings who never even heard a gunshot before may freak out and not be able to handle life apart from their hometown when they get an e-mail notification about a shooting outside their campus and the new bullet marks in the library wall. It’s acclimation.

frontier64 15 Jun 2022 19:37 UTC
1 point
on: Yes, AI research will be substantially curtailed if a lab causes a major disaster

because Covid did not come from U.S. gain-of-function labs

Do you have any particular reason to believe this?

frontier64 15 Jun 2022 19:32 UTC
1 point
0 ∶ 0
on: Slow motion videos as AI risk intuition pumps
I like using the intuition pump of, AI : Humans :: Human : Apes. Imagine apes had the decision to create humans or not. They can sit there and argue about how humans will share ape values because they’re descended from apes. Or how humans pose an existential risk to apes or some such.

Humans may be dangerous because they’ll be smarter than us apes. Maybe humans will figure out how to get those bananas at the very top of the tree without risk of falling, then humans will have a massive advantage over apes. Maybe humans will better know how to hide from leopards; they’ll be able to hurt apes by attracting leopards to the colony and then hiding. Humans might be dangerous, but if we contain them or ensure that they share ape values then us apes will be better off.

And then humans take over the whole world and apes live in artificial habits entertaining us or survive in the wild only due to our mercy. We’re just too stupid to reasonably think of the ways AI will be able to defeat us. We’re sitting here with a boxed-AI thinking about the risk of nanotech while the AI is creating irl magic by warping the electric field of the world using just it’s transistors.

Like, we’re so stupid we don’t even know how to spontaneously generate biological life. The upper bound on intelligence is way above where we’re at now.

frontier64 7 Jun 2022 9:45 UTC
1 point
0 ∶ 0
in reply to: mukashi’s comment on: We will be around in 30 years
It seems like someone’s already put the effort in to give you a list of ways AGI could kill all humans so I don’t have to do that.

frontier64 7 Jun 2022 9:09 UTC
3 points
1 ∶ 0
on: We will be around in 30 years
Downvoted because you give no reasoning. If you edit to give a reason why you think: “we are overestimating how likely is that an AGI can come up with feasible scenarios to kill all humans” then I will reverse my vote.

frontier64 15 May 2022 23:54 UTC
3 points
on: The Last Paperclip
Confusing choice to refer to the last probe of A as “D8” when D refers to the other AI civ in the story.

frontier64 6 May 2022 19:12 UTC
1 point
in reply to: james leeming’s comment on: How confident are we that there are no Extremely Obvious Aliens?
I’m partial to this idea. People think of a dyson swarm as a pinnacle technology when it’s really more of an engineering problem than a scientific problem for boring modern people. If there are highly advanced aliens out there I see it being very possible that their expansion is completely invisible to us because they’re operating on a level we can’t perceive.

frontier64 6 May 2022 17:27 UTC
3 points
on: What We Owe the Past
I take this a step further. I say that frontier14′s opinion holds more sway than frontier!24′s opinion. Frontier!14 has less information, but his mind is significantly more facile. Almost every single person drops iq points as they age. If in the intervening years I haven’t gained any new information which refutes the reason underlying frontier!14′s goals then I should on the whole trust his analysis over mine. Which is what I do.

This is more so an argument that it’s rational to respect your past convictions, rather than that it’s ethical to do so.

I think you’ll find that this community will be very opposed to respecting the past because they are largely fans of hedonism. Honoring past obligations made to one self is kind of the polar opposite of hedonistic pleasure seeking behavior.

frontier64 21 Apr 2022 19:03 UTC
3 points
on: “Pivotal Act” Intentions: Negative Consequences and Fallacious Arguments
This post is adjacent to an idea that I’ve thought about privately for some time, but which I haven’t seen discussed publicly in a straightforward fashion. If the first AI built can be aligned, whoever or whatever controls that AI will in effect have absolute power. And absolute power corrupts absolutely.

There are very few people alive that I would like to give that sort of absolute power to, and even fewer people that I should want to give that power to. Maybe we give that sort of power to a collection of people? Even worse! I definitely don’t want a group of AI researchers with competing interests, a corporation or the US government to be in control of aligning the AI. Sure we may have the functional ability to align an AI. CEV might be a good sort sort of alignment, but nobody should trust anybody else to actually use CEV to align their AI.

It’s unlikely we’re in a world where MIRI, Google, or OpenAI miraculously manage to develop AI first and have the ability to align it. It’s even more unlikely we’re in a world where Google or whoever else also decides to align it in a way that’s prosperous for you, me, or all mankind.

This is why I don’t hate Musk’s idea of AI for everyone as much as most other people. In fact I think it has a better shot at working than trusting a single third party to choose to align AI.

This should probably be a top level post, but I don’t quite like making those so I’ll leave it as a comment here.

frontier64 12 Apr 2022 16:12 UTC
1 point
in reply to: Thomas Kwa’s comment on: frontier64′s Shortform
My understanding is that it’s possible there’s a neural net along the path of GPT-1 → N that plateaus at perfectly predicting the next token of text written by a human that stops way short of having to model the entire Earth. And that would basically be a human internet poster right? If you create one of those, then training it with more text, more space, and more compute won’t make a neural net that models the earth. It’ll just make that same neural net that works perfectly on its own with a bunch of extra wasted space.

I’m not too sure my understanding is correct though.

frontier64 12 Apr 2022 16:00 UTC
2 points
in reply to: Chris_Leong’s comment on: frontier64′s Shortform
I hadn’t until you mentioned it here. I have now read through an explanation of InstructGPT by Openai here. My understanding is that the optimization in this case is for GPT-3 outputs that will be liked by the humans doing the reinforcement learning by human feedback system.

The openai people say that, “One way of thinking about this process is that it “unlocks” capabilities that GPT-3 already had, but were difficult to elicit through prompt engineering alone.” Which I guess kind of points out the problem I was thinking of. GPT-N is optimizing for predict the next token based on a bunch of internet text. All the addons are trying to take advantage of that optimizer to accomplish different tasks. They’re doing a good job at that, but what the big compute is optimizing for remains the same.

On a slightly different note, this paper kind of reinforces my current thoughts that alignment is being co-opted by social justice types. Openai talks about alignment as if it’s synonymous with preventing GPT from giving toxic or biased responses. And that’s definitely not ai-aligntment! Just read this quote: “For example, when generating text that disproportionately affects a minority group, the preferences of that group should be weighted more heavily.” It’s disgusting! Like, this is really dangerous. It would be horribly undignified if alignment researchers convinced policy makers that we need to put a lot of effort into aligning AI and then the policy makers make some decree that AI text can’t ever say the word, “bitch,” like that’s some solution.

ETA: Pretty troublesome that this is where we’re stuck at on alignment while Google has already made their improved version of GPT-4 and openai has created a new artistic neural net that’s way better than anything we’ve ever seen. I still think not too troubling though if they keep using these methods that plateau at the level of human ability. It might be an interesting future if AI is stuck at human level thought for a while.

frontier64 7 Apr 2022 18:49 UTC
0 points
on: Ukraine Post #9: Again

I still say they’re not paying enough.

Am I getting something wrong here or is this statement pro-let’s torture Russian POWs?

frontier64 7 Apr 2022 17:16 UTC
6 points
on: frontier64′s Shortform
I don’t see how we could ever get superhuman intelligence out of GPT-3. My understanding is that the goal of GPT neural nets is to predict the next token based on web text written by humans. GPT-N as N → inf will be perfect at creating text that could be written by the average internet user.

But the average internet user isn’t that smart! Let’s say there’s some text on the internet that reads, “The simplest method to break the light speed barrier is...” The most likely continuation of that text will not be an actual method to break the light speed barrier! It’ll probably be some technobabble from a sci-fi story. So that’s what we’ll get from GPT-N!

frontier64 28 Mar 2022 20:11 UTC
1 point
in reply to: Viliam’s comment on: Cyberwar escalation

it would be more efficient to contact your country’s secret service and ask if they have some good ideas—who knows, they might give you a code that only activates on a specific IP address, and maybe steals an important state secret

Which is probably happening already at a high rate. This node package author could be the one in a thousand who decided to take the visible, evil, ineffectual route instead of the invisible, evil, effective route.

frontier64 16 Feb 2022 20:20 UTC
2 points
on: Convoy Continued
The easiest solution instead of going in and physically removing trucks is probably just to give in to the trucker demands. The generally accepted demands are not too out there. It’s basically employ governmental restrictions that are no more unreasonable than those employed in America. Seems like that would solve the problem fairly quickly without the terrible optics of a physical crackdown.

edit: Seems like Trudeau has taken Zvi’s advice and decided to focus on physicality. Optics seem to be most people commenting that this is a grave constitutional violation and effectively turns Canada into a tyrannical state where the executive jails his political opponents.

frontier64 11 Feb 2022 0:32 UTC
12 points
in reply to: Matthew Barnett’s comment on: Why do people avoid vaccination?
The poll was anonymous and was not inviting debate.

my starting assumption is that people will actively be dishonest by trying to invent a reason that sounds noble when the true motive is mundane and kind of ugly

This sounds like a good way to disregard all stated reasons all the time. I can only see it being a valuable heuristic if you first have some other reason to believe the true motive.

frontier64 10 Feb 2022 22:06 UTC
7 points
in reply to: Matthew Barnett’s comment on: Why do people avoid vaccination?
Nobody in the post’s quoted survey results mentioned needle phobia or gives any indication they have a fear of needles. Is that not evidence that needle phobia isn’t the cause?

frontier64 10 Feb 2022 22:04 UTC
2 points
on: Why do people avoid vaccination?
You’d have an easier time getting an answer to your question if you came at it from a perspective of improving epistemics overall instead of looking for one tiny specific answer. You’re automatically assuming your reasonable sounding understanding is correct and focused on finding reasons why the other people’s reasonable sounding, yet different understandings are wrong. I don’t believe that will get you a lot of information.

You’ll know more about why people believe X and why X is wrong by just trying to become more knowledgeable on X ~X and a variety of related issues as a whole.

frontier64 8 Jan 2022 17:06 UTC
−8 points
in reply to: CalamityJane’s comment on: Thank you, Queensland
They granted his exemption and then revoked it after he entered the country. Then they kept him locked up in a dirty hotel instead of just letting stay in a private residence with his other support staff. It’s a really poor opinion to think that’s ok treatment.