Skip to main content
MANIFOLD
Daily LLM assistant personal usage exceeds 2 hours for >10% of users by end-2025?
18
Ṁ1kṀ1.1k
resolved Apr 16
Resolved
YES

As https://manifold.markets/JoeandSeth/daily-llm-assistant-personal-usage

But for end 2025

  • Update 2026-04-11 (PST) (AI summary of creator comment): - Character.ai is currently being considered as a qualifying platform, despite the creator's hindsight that it should likely have been excluded

    • If data shows Character.ai's average user time exceeds 1 hour (implying top 10% exceed 2 hours), the creator expects to resolve YES once more granular data is available

    • Final resolution is open to discussion

Market context
Get
Ṁ1,000
to start trading!

🏅 Top traders

#TraderTotal profit
1Ṁ104
2Ṁ56
3Ṁ40
4Ṁ33
5Ṁ2
Sort by:

@traders my current understanding is that character.ai has average user time of greater than 1 hour, which would imply the top 10% at above this 2 hour threshold. On hindsight, this platform should likely have been excluded from consideration, but as things stand, I would expect to be able to resolve this market YES as soon as this data is made available and more granular.

Open to discussion.

@JoeandSeth I'm not very familiar with it; would it count as an "assistant"? Talking to characters sounds different to me

@ImaPerson this is why I'd mentioned that particular platform being counted disrupts the intent of the question - but character.ai has been explicitly mentioned in the list of qualifying platforms.

That said, I do not think talking to characters means they are not useful as chat assistants. Therapists, outside perspectives, even validation are all legitimate use-cases - and there's likely little chance of getting those uses differentiated from the rest of the pure role-playing data.

If someone talks to Dr. House about their medical concerns or to Luke Skywalker about the pains of being given unwanted responsibility or Chuck Norris about being bullied?

All that said, my intent in creating this question was to measure the rate at which use of LLMs for personal reasons was increasing. This question was poorly worded, and while I put some of that on the fact I wrote it with AI, the rest falls on myself for taking that shortcut.

@JoeandSeth https://qz.com/a-startup-founded-by-former-google-employees-claims-tha-1850919360
Feels like a thirdhand claim, but the source claims 7x the engagement that would resolve this market YES.

With no other trader responses in the last couple days to indicate to the contrary, I will be resolving this question YES and adjusting the 2026 question to specifically exclude roleplay-first LLM assistants.

Will wait to resolve for a little while to try and get survey data more recent. Last-Nov information clearly indicates NO but I'd hate to close down on such a fast-moving domain with only year-old data.

https://rethinkpriorities.org/research-area/estimating-the-usage-and-utility-of-llms-in-the-us-general-public/