
Background
OpenAI recently introduced a $200/month ChatGPT Pro subscription tier, which includes advanced features like research and operator.
Resolution Criteria
This market will resolve YES if I am able to generate at least $200 in additional income that can be directly attributed to using the ChatGPT Pro subscription features within the first month of subscription. This includes:
Revenue from projects or work completed using Pro features
Cost reductions in existing expenses
The market will resolve NO if:
I fail to generate $200 in measurable value within the first month
Considerations
Success will largely depend on my ability to effectively leverage the advanced features for revenue-generating activities
The $200 threshold must be reached through new or additional income/savings, not from existing revenue streams
While the subscription includes powerful features, monetization requires active effort to create value from the tool
Results may vary significantly based on use case and implementation strategy
Update 2025-02-22 (PST) (AI summary of creator comment): Time Valuation:
The creator is currently unemployed and assigns a value of 0 to time spent, meaning that time investment will not be factored as a cost.
Expense Deduction:
Any direct monetary expense incurred as a result of using ChatGPT Pro (for example, purchasing AWS storage) will be deducted from the net generated income.
Update 2025-02-25 (PST) (AI summary of creator comment): Exclusions:
Benefits like food stamps will not be included in the $200 calculation.
Verification Requirement:
Only revenue that is confirmed via a voice-based phone interview will count toward the $200 threshold.
Update 2025-02-28 (PST) (AI summary of creator comment): Advanced Subscription Features Only:
Only revenue generated from deep research, operator, and sora—features exclusive to the $200 subscription—will be considered.
Any income generated using free tier or lower tier capabilities is excluded.
will you compare other tools? Do you try deep research by google/grok too, or image generation by imagen? I'm worried that you may use it and think it's good without trying other free/cheaper services, and market will resolve to yes just because AI tools in general helped you
@eamag ok I see what you are saying, but taking the cross product between all possible tools I could use at all price points and all possible prompt combinations I could have used is just not tractable. I will make sure I am super clear what made the money in the end, and other people can see if these cheaper tools would have worked. I haven’t seen any other browser users and that has been my main money making tool so far.
@bessarabov as clarified below I am not using any free tier features. So I am only using deep research and operator and sora.
@tftftftftftftftftftftftf I don’t want to go through and re-verify if it could have been done with free
@SimoneRomeo I’m currently unemployed so I’m treating my time as 0 but my money as 1. So if it convinces me to buy a bunch of aws storage I will deduct that cost from the net.
@tftftftftftftftftftftftf do you consider earnings even if operator does actions that you could have done yourself in less time? For example, if operator spends 40 hours to click stuff on the web, in the process it makes 100 EUR, but you could have brainstormed with an open source model and clicked that stuff yourself in less time and earn the same.
@SimoneRomeo good question. I don’t think I could stand up operator with an OS model. Maybe someone could but I can’t.
@tftftftftftftftftftftftf what I mean I that you can click links yourself. Chatgpt can recommend what you should do and you can click what chatGPT suggests. This can be much faster than waiting for operator to click. My experience with operator is that it takes more time than it saves to you, so even if it might be able to earn money, I think you'd be better off just doing work yourself.
@tftftftftftftftftftftftf I just want to not that this isn’t a market closing event because the application hasn’t been approved yet
@tftftftftftftftftftftftf I am not including this in the 200- a voice based phone interview is required
@iicaptaineye Exactly, I would say that just being given the idea that you qualify is sufficient. But I’m not arguing today.