Which of Casey Newton's 2024 predictions will come true? [Multiple Choice]
191
831
αΉ€10K
2025
84%
GPT-5 begins training.
75%
The quality of search results degrades as Google proves unable to reliably detect AI-generated content.
71%
Fragmentation in social media will make it more difficult to understand which narratives are resonating most with voters.
71%
The US presidential election is full of synthetic media, and it mostly doesn’t matter.
70%
TikTok skates unscathed through another presidential administration.
65%
Meta kills off its incomprehensible celebrity-based AI characters.
63%
Scaled-up AI-generated sludge outcompetes many digital media companies for advertising and affiliate-link dollars, sparking further waves of job losses and consolidation.
60%
Bluesky begins to wither as its development team prioritizes building its underlying protocol over growth, community management, and making improvements to the user experience.
59%
The number of people who say they are in romantic relationships with AI companions will increase sharply.
42%
Google mostly catches up to OpenAI in LLM quality and begins to neutralize GPT's lead.
33%
Apple's Vision Pro is successful enough to revive interest in mixed reality and the metaverse.
7%
Threads overtakes X in daily users and becomes the leading text-based social network.

SPONSORED by @BoltonBailey! Huge thanks to Bolton for volunteering both the idea & the seed mana for this market (how generous! 😁)

***

All of the above predictions are quoted verbatim from his newsletter, but I tried to excerpt only the most pertinent information from the paragraph-predictions. See here for a full explanation on his blog, Platformer. (The predictions post is free to read without signing up, but it will ask for your email. You can say no.)

One very important point all bettors need to understand is that the only additional criterion I add on top of Casey's is that the predictions must have verifiably come true by EOY 2024. This does NOT require the prediction content to have occurred in full or in part in 2024; e.g. we may have reliable reporting in January 2024 about GPT-5 being trained in 2023, and that would count as a "come true" for the purposes of both markets. That said, nothing that came true only after January 1st, 2025 counts for fairness/clarity/intelligibility. If you disagree with this rule, please do not bet on this market.

NB: the time of reporting on an event =/= the time the event occurs (or fails to occur).

Given this qualifier, I reserve the right to extend the resolution date (tentatively scheduled for January 3rd, 2025) for any reasonable extent beyond the closing date (firmly at December 31st, 2024). I can't foresee any reason it will need to extend beyond January 31st, 2025 at this juncture, but at the absolute latest I will resolve it by Valentine's Day, 2025 if there are serious ambiguities that need time to consider or research. This outcome is especially likely if Casey fails to post a free 2024 retrospective newsletter.

Given the subjectivity inherent in evaluating some of these questions, I will be betting in neither and any resolution outcomes are up to my sole discretion. (That said, if you have good-faith advice, I'd be happy to hear it, especially about resolving to "probability" because I don't have much experience doing that. )

Something important to note about this market's differences from the numerical one is that any of these items could resolve at any time over the course of the next year.

If I am extremely confident in resolving I will do so immediately and without warning, but if there is significant uncertainty I will poll the comment section below. I may resolve any individual item 24hrs+ after I post the poll--I want to allow for those who aren't on Manifold 24/7 to have a chance to voice their opinion--but also resolve quickly enough so that (hopefully) 99% of bettors are satisfied with my management of this market. As such, I'll put an upper limit of one week after poll posting to make a firm decision as to "Yes" "No" "N/A" "Prob" or "Decision Postponed (Awaiting Further Reporting)". The only situation in which any will resolve "N/A" is if it becomes clear after the fact that something was inherently flawed in the wording/intent/factuality of the prediction and therefore a rational prediction couldn't possibly have been made (or the prediction no longer makes sense for some reason).

If you bet on this market, YOU are responsible for understanding and agreeing to ALL of the resolution criteria linked below and listed above. If you have any general questions about the Platformer 2024 Predictions, please post them in the numerical market so that we can have a centralized location to discuss (any queries specific to this market can, of course, be left in the comments below).

Please click attached to read how I plan to resolve both 2024 Platformer Predictions markets:

Thank you for reading, thank you for betting, and may the odds be ever in your favor!! 🎲🎲

Get αΉ€200 play money
Sort by:
Scaled-up AI-generated sludge outcompetes many digital media companies for advertising and affiliate-link dollars, sparking further waves of job losses and consolidation.

That's kind of hard to resolve to yes or no. There are some layoffs and some AI-generated sludge already, but the criteria are extremely fuzzy.

Isn't AI generated content better than most of the human generated trash SEO optimized sites anyway? I don't see how AI output is a loss.

GPT-5 begins training.

I thought Altman has already said this has started?

@BTE These are verbatim Casey’s comments predictions, not my own.

For resolution purposes, any of the above 12 only has to occur by EOY 2024 to count, so if it happened in 2023 and only got widely reported on in 2024, that would count.

Scaled-up AI-generated sludge outcompetes many digital media companies for advertising and affiliate-link dollars, sparking further waves of job losses and consolidation.

I'm sure the first part is going to happen, but not the second part. The companies can just adapt and focus on other parts of the market.

Fragmentation in social media will make it more difficult to understand which narratives are resonating most with voters.

According to who? It seems to me that getting user data is getting easier rather than harder, but it's the kind of question a motivated reasoner can reach any conclusion on

@TheBayesian

Simplified Decision Heuristics for the market:

1) Casey’s own evaluation in his 2024 retrospective iff reasonable and evidenced

2) If 1 can’t be used, it goes to my own evaluation based the consensus emergent from multiple sources reported on by multiple credible journalists/institutions

3) If 2 is impossible (no consensus, no credible sources, very low confidence) then I will be polling the comment section and asking for resolution arguments from veteran Folders, weighting the latter with a 90% input and 10% input from the polling section.

My goals are to be as factually accurate, trader-fair, and outcome-neutral to the greatest extent possible, but ultimately traders need to trust I’m not going to pull some ridiculous market manager BS of the kind recently seen in at least one other unrelated market (which shall go unnamed).

If the subjectivity bothers you despite the above clarification and absence of any self-dealing on my part, then perhaps it would be better not to place any bets here.

@Ophiuchus thank you! that works

bought αΉ€10 of Meta kills off its i... YES

πŸŽ‰