Misunderstanding with Founders in AI Startups

After the release of Open AI APIs, it's very easy to add AI as one of the features in our SaaS applications. But most of the time people often confuse this AI with AGI (Artifitual General Intelligence).

What we see every day while using Open AI, Claude AI, Gemini, or anything else is all part of ANI (Artificial Narrow Intelligence). These AI models are meant to work on already existing data and can generate output on top of it. They can be used for Predictions, Search, Summarization, and Generations. But you cannot expect them to work fully on something new for which they don't have any data.

Now we often compare these models and think that we have created JARVIS from IronMan or StarWars Robot like BB-8. This is not true, the AI that you see in movies is often AGI having the ability to generate and do new things without giving them the data.

One of the most common examples you can have is,

Let's suppose in your SaaS there is some sort of campaign creation that is needed to go over 6 Steps to complete. Now you can use AI to help people understand what each Step is but you cannot use AI to automatically fill all the fields and make changes to the previous fields just by a chatting interface between the user and the AI.

You might make it work, but for that, you either have to provide a lot of training data and strong prompts or you need to hardcode some steps manually.

The only way to achieve features like this is to use AGI, which we have not seen yet.

So don't confuse the AI that you see every day with AGI, today's models are best workable on already existing data.

You could waste your time and effort to build such features that require real human-like intelligence.

Let me know if you have any questions in the comments I would love to discuss more.