Should We Build Our Own AI Tool / Startup?

Mohammad Mohammadi
2 min readOct 7, 2024

--

As AI continues to evolve, one of the big questions we face is whether the future will belong to one all-powerful AI or if we’ll see a landscape filled with specialized AIs tailored to specific tasks.

In the first scenario, where a single AI dominates, it’s likely that giants like OpenAI will lead the charge, leaving little room for smaller players. But in the second scenario, there could be space for thousands, even millions, of AI startups, each excelling in a particular area.

Let’s push this idea further. Imagine we had artificial general intelligence (AGI) today — an AI that could think and reason like a human. Wouldn’t that solve every problem? To explore this, we need to compare AGI with the alternative we have today: us, humans.

I won’t dive into the technical details, but I believe AI systems will likely resemble human skill sets. We’ll have big, versatile AIs like OpenAI, capable of handling a wide range of tasks, but we’ll also see specialized AIs that outperform in niche domains.

There’s a huge opportunity in specialization for two key reasons: 1) better results with higher accuracy, and 2) reduced costs.

Think of OpenAI like a heavy-duty truck. It’s powerful, but you wouldn’t use it to deliver a small bag of potatoes. That’s where smaller, more nimble solutions — like bicycles — come in. Building these “bicycles” may not be a big opportunity for the AI giants, but for small startups, it’s significant.

By focusing on specific problems, startups can create AI tools that deliver better outcomes at a lower cost than large language models, giving customers exactly what they need with greater efficiency.

--

--

Mohammad Mohammadi
Mohammad Mohammadi

No responses yet