Why I am not going to build anything with AI (for now)

Published December 18th, 2022 2 min read

The purpose of this article is:

  • To state my opinion that AI is currently overhyped and not understood enough to make substantial bets on
  • Which products will emerge and benefit from the current AI hype
  • To remind you to focus on your made commitments and execute existing plans
  • How the landscape is going to change in the coming years. How to be early

Artificial intelligence has gained substantial traction globally in the past few months. Much of it is attributed to two groups of generative AI tools that have made waves more recently - art and text. There are mainly three big players in the art sphere: Stable Diffusion, MidJourney, and Dall-E. Text is almost entirely occupied by OpenAI's ChatGPT.

The amount of discourse and experimentation shared on social media platforms that followed the release of these tools quickly blew up. They intrigue through their novelty and "magic", but at the same time, the technology is challenging to understand and control. This leads to vastly unrealistic expectations, fear, hopes, and predictions. But due to the significant misunderstanding, betting on "AI" may lead to both dramatic fortune and hardship.

But what opportunities are there to build products with AI? It is probably a bad idea to train (or, god forbid, entirely build) your own models. You will compete with elite academics immersed in the field and start at a massive disadvantage. Working on anything that goes beyond the capabilities of existing no-code tools is, in all likelihood, a vain attempt. The goal should be to draw value from AI, as the recently surfaced AI tools are merely toys in the current state.

Your best bet is integrating them by synthesizing interfaces for a specific purpose, granting easy access to non-techie people. Jasper, a copywriting assistant, currently valued at $1.5 billion evaluated startup, got ahead of the curve doing precisely that (in less than 2 years. As a bonus, #AI (Twitter hashtag is meant metaphorically for all potential content portals) is a gateway to free attention - from the nerdy fraction of internet people and potential investors. Ultimately, an AI-driven approach will enhance and disrupt existing software solutions. Many products will be battle-tested, and some of them will die.

Despite the opportunities that may lay ahead, I decided not to FOMO on the hype train. Nothing good ever cames from FOMOimg. It is a massive diversion of focus, so I will hold on to my current plans. However, I will keep carefully observing new developments in the AI space and adjust my strategies to either hedge against or thrive on AI.

Addition from 15th January 2023:

I am thinking now, that it would be simply a healthy idea to get my feet wet with AI in one way or another. Finding time should not be an issue for a prototype or MVP ever, however I still don't want to dump time on an AI-first project. Instead, I will lean towards enhancing existing projects with naive AI features and see where it leads me.

I am currently writing a stimulating piece about AI and its implications on the future of software jobs. If you want to get notified when it goes online please consider subscribing to my newsletter below 👇.

Maxim Zubarev Avatar

A personal publication by Maxim Zubarev.
I use software as a leverage for business.

Join my Newsletter

Want to receive more articles like this? I write about programming and how to grow your business with software. Subscribe to my newsletter to receive all my upcoming posts delivered to your inbox.