Can GPTStore generate more revenue for developers than the AppStore?

As a developer, I am both excited and concerned about the future of GPTStore.

My apprehension arises from a calculation made by another creator. If those numbers are to be believed, it appears that the earnings potential for creators on GPTStore might be somewhat restricted.

This brings up important considerations regarding the platform’s potential for profitability and long-term sustainability for developers.

Addressing this issue is essential to ensure that GPTStore is not only a hub for innovation but also a financially viable and rewarding platform for creators.

The safest assumption is likely no. The GPT store is designed to allow people to share cool specialized uses for OpenAI’s models, and to encourage people to sign up for ChatGPT Plus.

It’s important to remember that ChatGPT is a very weird product, they make less money the more you use it. The platform was never really designed to be a “new App Store” and I think OpenAI may have made a mistake calling it the GPT Store.

The best of the best might make some money, but for the most part its safe to assume being a GPT developer won’t pay as well as a full time job.

This would significantly diminish the value brought by GPTs and the motivation of developers. Currently, many GPTs seem to be used primarily for attracting traffic. However, how many of these GPTs are genuinely creating value for users? I believe there should be a further reduction in the barriers to usage, and at the same time, GPTs could have individual pricing. The platform could simply take a commission from these sales.

This is almost certainly the case.

ChatGPT as a platform is definitely a weird beast. I don’t doubt there will be some GPTs that offer an increased value for certain niche use-cases compared to base GPT-4, but they’re going to be exceptionally few and far between.

Most—but by no means all—GPTs that I’ve seen are pretty lazy efforts, with much more attention paid to trying to “secure” the instructions than to actually creating something interesting.

Beyond that, the vast majority of GPTs appear easy enough to replicate even without accessing their instructions, because it’s just not that hard to deduce the general gist of their instructions.

Even the GPTs making use of external knowledge don’t add particularly too much value as, again, it’s not terribly difficult to identify the types of documents that the model would need access to in order to perform its job.

For most of these, the qualitative difference between any dozen attempting to do the same thing just isn’t going to be enough of a differentiator to allow any one to really stand out.

There will, of course, be some exceptions for those GPTs that have obscure, specialized, or proprietary knowledge bases or some kind of uniquely clever instruction set, but that’s absolutely not the case for almost any of the GPTs out there.

The only GPTs I’ve seen that add any real, true, value over stock GPT-4 make pretty heavy use of actions that extend the capabilities of the models in new, novel, and interesting ways that aren’t easy to replicate—particularly those that bring together multiple actions for different services or do clever things calling other language models or specialized tools to super-charge the planning and reasoning capabilities of GPT-4.

There may be a few flukes that go viral for the lulz, but I would be absolutely shocked if more than 500 or so GPTs ever made more than a few bucks a month.

But, I didn’t know, I would be exceedingly happy to be wrong. I know there are those at OpenAI who are very invested in the idea of ChatGPT as a platform and want developers to be able to make real money from it, and they know way more about it than I do.


1 Like

I understand your concerns about the practical value and potential earnings for developers in the GPT market. Indeed, many GPT applications in the market may not innovate enough to significantly differentiate from the basic GPT-4 model, potentially limiting their value and earning potential. However, I believe the future in this field still holds great possibilities.

Firstly, although many GPT applications currently seem similar in content and functionality, with the evolution of technology and changing market demands, we might see more innovative and specialized applications emerge. Such applications can offer unique value beyond the basic model, thereby bringing greater revenue to developers.

Secondly, the instances of GPTs you mentioned that “heavily use actions to extend the model’s capabilities” show that when developers creatively combine GPT technology with specific needs, they can indeed create new, interesting, and valuable applications. This is not just a technological challenge but also a market innovation opportunity.

Moreover, while most GPT applications might struggle to generate significant direct revenue currently, they lay the foundation for brand building, user engagement, and long-term development. These indirect benefits may become more significant over time.

Lastly, regarding the potential of ChatGPT as a platform, I agree with OpenAI’s perspective. Although it might not have reached the ideal level of commercialization yet, with ongoing technological advancements and a gradually maturing market, its potential could be further realized.

In summary, while the current market performance of GPT applications might not completely meet expectations, the future development in this field remains promising. For developers, now is an opportune time to explore new possibilities, experiment with innovative applications, and prepare for future market shifts.

Earning money solely from a GPT is not easy but you can add GPT to demonstrate your tool functionality and draw traffic to your tools. Few popular GPTs are doing this already such as Canva

Yes, it has gradually become a tool for market lead generation.

This depends on several factors. First, the extent to which Custom GPTs can establish their own barriers (as mentioned above, there are currently few barriers); second, the openness policy of OpenAI (the target user group that can use it determines the market size); third, the revenue of AppStore comes from developers’ control over their own business, but GPTs depend on the limitations of OpenAI’s GPT Plus. When can this change?