In the past year, a surge of AI tools has hit the market, with many identifying as AI startups. The advent of OpenAI’s ChatGPT, including GPT-3.5 and GPT-4 models, has revolutionised how we interact with technology. However, amidst this excitement, a trend needs addressing: the phenomenon of “API wrappers” masquerading as AI startups.
While it’s true that many of these products utilize the power of OpenAI’s GPT APIs, it’s essential to take a step back and consider the implications of relying solely on an external API for your business. Does wrapping GPT APIs and selling a service based on them warrant the label of an AI startup? Let’s take a closer look at the potential downsides of this approach.
Control Issues: The Risks of API Dependency
First and foremost, basing a company on an API you don’t control is inherently risky. Here are a couple of scenarios that could threaten the stability of such a business:
- License terms change: If OpenAI alters its licensing terms, and your “startup” violates these new terms, your business could face severe consequences.
- Increasing costs: If OpenAI increases its API usage costs, your business may no longer be able to operate, resulting in potential collapse. The GPT-4 model right now is a lot more powerful, but it’s expensive compared to GPT-3.5.
Remember when Twitter altered their API access, leaving several popular apps built around it in a lurch? That’s a prime example of the risks of building a business around an external API.
Redefining AI Startups: More Than Just API Wrappers
With a constant stream of AI products appearing on platforms like producthunt.com, it’s evident that people are excited about AI, and that’s fantastic! However, we must address the elephant in the room: should we call these products AI startups if they’re just wrapping the GPT APIs?
The AI ecosystem should foster genuine innovation and encourage businesses to develop unique algorithms, models, and solutions. By merely wrapping the GPT API, these “startups” may be stretching the definition of what it means to be an AI company.
In Conclusion: Embrace Innovation, But Be Wary of Hype
While there is nothing inherently wrong with using GPT APIs to build a product or service, it’s crucial to understand the limitations and risks associated with this approach. Entrepreneurs and investors should be cautious not to get swept up in AI hype and maintain a discerning eye when evaluating these AI startups.
I would also caution people signing up to these slapped-together AI startups to evaluate what they’re signing up for. Given how easy it is to create an AI tool now, you could open yourself up to possible data leaks. This rings true for any service or software, but a quickly thrown-together API wrapper built to capitalise on the AI hype and make some money could be a risk.
By acknowledging the potential pitfalls and encouraging true innovation, the AI community can continue to grow and thrive sustainably.