~2 years ago, I predicted we would have to hit the following key milestones for a positive resolution by 2028

>Key milestones to look for:

2025: the first human-edited video >30min that is actually watchable

2026: the first low-quality fully-ai-generated movies

2027: AI generated movies proliferate across the internet. You cannot scroll youtube without encountering them (some of which have millions of views). Occasionally it takes you a minute or two to realize a movie is ai-generated.

2028: This market resolves positive (idk, certainly I am a buy at 43%, probably not a buy at 90% though).

I think it's worth noticing that we have now blown past: "first human edited video >30min that is watchable" and "you cannot scroll youtube without encountering them" [at least I can't]. I think the low-quality market is imminent (could resolve right now if someone had sufficient budget/motivation).

I still think we are 6-months to 1-year behind the schedule required for a positive resolution. But given that the asymmetric risks are all on the upside, I would not be selling this market at 50/50.

@logan RE: asymmetric risk- I've said this before, but I don't understand the limit orders above ~75%

Some startup could come out of stealth mode with this solved, and it wouldn't even be the most surprising AI thing this year.

But there's big mana bets that any news that pops the market that high will turn out to be false.

I don't think this market is a sure bet, but a false pop seems unlikely given the asymmetry.

@robm I disagree I think if anyone released a model that could do this, it would definitely be the most surprising AI this year

@Gen interesting take on the website where we measure how surprised we would be by things 😜

I'm really bad at predicting when you all would push this one low or high. I kinda expect it to go to 80% at some point, but it really takes its time, random dips to 30%.

It's such a dilemma of buying now and wasting chance to buy better when it goes lower later, or waiting and potentially missing the chance to buy while it's in low consensus regime.

@Lavander limite orders are the way

can somebody rig this market into yes by just starting a company to do it

@Iflewto done

CONVICTIONDIFF

@Bayesian nice

@Bayesian okay 50% is too high

@bens I’ve had enough

@Bayesian Joshua fills me and jim at 36%. Rocketman shows up and suddenly he doesn't want any action. 😭

@cthor it’s rly wild i have placed like 400k mana worth of orders and crickets

I smell FEAR. Thinly veiled attempt at bait

@Bayesian If I were a true whale I would keep going, but my entire liquidity isn't worth moving this since I don't care about leagues any more.

@Bayesian my limit order is never getting filled is it?

@EliezerYudkowsky “non-epistemic" (IYKYK)

The old base rates are dead. LK-99 already resolved 'yes' (Navier-Stokes Equation)

Eliezer Yudkowsky (@allTheYud) on X

@jim Hi @EliezerYudkowsky just a heads-up—you might want to sell your 'no' shares

@jim The credits at the end suggest it was written and directed by a human, which makesit significantly less impressive than if it was produced one-shot by a prompt.

@WilliamKiely yes but 2028

@WilliamKiely yeah, and there isn't much incentive to get the workflow down from one to zero human directors. Cutting down studio sizes from a hundred to one is ninety-nine salaries you don't have to pay, but going from one to zero is a huge leap and doesn't make much more money (or any at all).

The idea that a consumer would actually want this is some typical mind fallacy. Most people don't want to make anything. They want TV. 1% rule - Wikipedia

So this won't YES from any large-scale effort to make such a system, but as emergent behaviour, delivered as a tech demo rather than a dedicated product.

@jim so my (correct) take exists for future posterity to show how clairvoyant I am, I share my unrequested opinion.

The response to Overman has made me convinced I think this will resolve YES, and not necessarily because I agree with the decision, but because the general pub and tech bros (who don't know ball (cinema)) are some of the least discerning morherfuckers ever.

@No_uh This is now my number one concern re: a wack resolution. I have to realize that these people are somehow missing so many quality markers that my theory of mind for them may have been too generous. Still, great for my self esteem re: taste.

@No_uh this is a good take but I suspect Alexander is a bit different.

@No_uh I'm leaning towards thinking this should resolve N/A. This thing was a more enjoyable 25 minutes than anything superhero I've ever seen, but mostly just because it took one concept all the way ad absurdum. The story on its own didn't make a lot of sense and the ending is in a weird contradiction to everything that came before, and IMO there were only two slightly funny moments (debate, monuments).

My point is, there isn't a single quality scale that AI movies and regular movies exist on (same for any other medium). This was good in that it didn't have to follow the formula you would expect, and bad in every aspect of execution or thematic coherence. I expect that the future will bring lots of new and different stuff but not more of the things we currently consider masterpieces.

all markets ought resolve N/A, the journey is the worthier part

The story is ok but the rest is dull the same way all these AI films are. I'm quite sure finally one of these will be done by someone who understand cinematic writing and be really good. However a good one-shot from a prompt in early 2028, this is still no.

@jim if N/A didn't claw back all profits and losses during the whole trading life lowkey this would be fine (i liked the perpetual markets im sad they were born and aborted in such short time)

@No_uh resolve to current probability at a random time is kind of like resolving NA without claw back