Corn Woman 🌽 @WomanCorn@schelling.pt
Corn Woman 🌽
boosted
In 1999 to 2000 I was going through a very rough patch in my life. In the span of a few years, I'd gotten married, lost a parent, birthed a child, gotten divorced, changed careers, and relocated 3,000 miles away from everything and everyone I knew. What with the state of things, I didn't get to see #GalaxyQuest when it was first released.
Finally, on a day I was feeling particularly low, and shortly before the movie left the theaters, I canceled work and took myself out to a midday matinée.
I was the *only* person in a vast theater, of a size that no one builds anymore, sitting and waiting for the lights to go down. At one point, I turned and peered up at the projectionist's booth. I saw a shadowy figure moving back and forth behind the window, bending and straightening. This was in the days before automated, digital films. The Phantom Menace had been released digitally in 1999, but the equipment to show such films was extremely expensive and most theaters hadn't converted yet; "projectionist" was still a real job.
While I was looking, the figure paused, strode to the window and peered back at me, then disappeared quickly.
I turned back around and continued to fidget and ponder the misfit pieces of my life.
At the top of the aisle behind me, the theater door swung open and banged loudly on the wall. The projectionist strode down the aisle toward me, a tall barrel-shaped man with a thick beard and glasses. My first thought was that the matinee was canceled due to low turnout, and I'd be getting a refund. Just as I'd resigned myself to that, the marching projectionist shouted out in a booming voice,
"WELCOME to your PRIVATE viewing oooooooof GALAXY QUEST!!!"
He stopped in front of my row, and I saw that he had an *armload* of STUFF. One by one, he began presenting each thing to me, and as the pile in his arms dwindled, the one in my lap grew.
"As our SPECIAL VIP Galaxy Quest GUEST today, YOU are entitled to..."
"- A commemorative t-shirt!"
"- A poster suitable for framing!"
"- A limited edition refrigerator magnet!"
"- A button to pin to your lapel!"
The list and the shwag went on. With every ridiculous item, I laughed harder and harder, until there were tears leaking out of the corners of my eyes.
Then he bowed and shouted, "WE HOPE YOU ENJOY THE SHOW!" and turned on his heel to march back up the aisle and out the exit door.
Alas, of all the shwag only the magnet has stood the test of time. But the humor and kindness of the unknown projectionist lives on.
@ami_angelwings yeah. I'd rather have more episodes if they're needed.
The ones that get me are <here's a 12 episode season adapting books 1-3 of a 12 book series. If you're good and buy enough figurines, maybe we'll animate more.>
@ami_angelwings I prefer adaptations that complete their story in a season. I don't care of it's 12, 24, or 36 episodes, but I really don't want half the story then wait + hope they actually bother to finish.
I have seen all three of these seriously discussed by people downstream of the LessWrong memeplex.
In version C, we're talking about Reward Hacking, and the AI has "maximized" something by overwriting the memory cell where it tracks the count. The "paperclips" are tiny memory cells that the AI can use to increase the number it cares about.
In version B, we're talking about Inner Alignment failures, where the AI is programmed to maximize human happiness, and the "paperclips" are 10-neuron constructs that count as human to the AI and can only feel happiness.
In version A, we're talking about the Orthogonality Thesis, and the paperclips are actual paperclips*, because the point is that a superintelligent AI might not care about what you care about.
* This also applies to bolts, or Facebook share prices.
The Paperclip Maximizer is a type error.
There are at least three versions in play, depending on what's being talked about.
Corn Woman 🌽
boosted
extra-punish people who euler you without anything backing it up
@empathy2000 is this just because we use the jargon "tacit knowledge" for that category, or do you think there's more discussion missing?
Corn Woman 🌽
boosted
the kryptonite fallacy: steelmanning the opponent's argument and then pulling out a counterargument that seems to hold, but actually only works on the steelman and not on the regular version
@flats I think the instrumental convergence argument is still pretty good. It does rely somewhat on the idea that the AI will be trained to optimize a single metric.
When reinforcement learning seemed like the winning technique, this was a big risk. Now that LLMs are the most promising technique, it's less clear. <Minimize next token prediction error> doesn't obviously call for conquering the universe.
@flats right. The question is how many of the fundamental arguments were worked out assuming that the goal was to build a CEV sovereign and never rechecked to see if they still apply now that that goal has been abandoned.
@flats If the AI isn't going to acquire godlike power, how many of the issues devolve into the principal-agent problem?
But no one wants to double check 1000 pages of blog posts to see if the conclusion relies on an unstated assumption.
@flats I think the problem is that a lot of their thinking on AI has a presumed final step <then we give it control over everything and it instantiates heaven on earth> and a lot of the threats hinge on the implicit assumption that you will give the AI control over everything.
So, an AI might conceal its real goals... Is that an issue if it is only going to get enough power to run the factory?
Maybe, maybe not. But we have to check every argument.
Waiting for service marks:
Networkless: Doesn't communicate with the seller or nay service provider.
Updateless: Simple enough that they won't ever be shipping a software update.
Microprocessor-free: The ultimate in assurance that it doesn't spy on you.
How many people have died since you decided nonviolent protest was the right option.
If you had just gone to war, the death total would have been smaller.
Alice: LLMs are not thinking. They're just rephrasing things they've already read. Their alleged essays revel that they have no understanding.
Bob: By that logic, most humans are not thinking either.
Alice: Okay.
Does anyone have good examples of Progressive Web Apps that are not based around disconnected use?
(Disconnected use is the acid test in a lot of ways, but it's not the build-around goal.)
