Is the Era of Free AI Over?
OpenAI has just introduced its new subscription plan for $200/month. By the end of this article, I’ll analyze whether it’s really worth it, but the trend is clear: the era of accessing large AI models for free or almost free is over. We’re heading toward a subscription model with high costs and/or the inclusion of ads.
Disclaimer: They were never truly free. We’ve been offering our data and feedback at no apparent cost. Remember this phrase: “If something is free, you’re paying for it with your data,” and this applies to absolutely everything.
There’s a phrase I really like: “Google wants to be OpenAI, and OpenAI wants to be Google.” What does this statement mean? On the one hand, Google aspires to be like OpenAI because, after having become the leader of the internet and search for years, it fell behind in the race for artificial intelligence. When OpenAI burst onto the scene with ChatGPT and gained a competitive edge, Google lagged behind, forced to react so it wouldn’t miss the boat in a technological revolution it believed it had under control — especially considering it was a pioneer in AI with projects like DeepMind. From there emerged the foundations that expert Gustavo Entrala explains so well, including the transformers.
On the other hand, OpenAI looks to Google in search of a similar profitability model to the one the latter has maintained thanks to its extensive advertising ecosystem. Google, with its enormous user base and advertising machinery — the core of its business — has enjoyed stable and long-lasting revenue. Now that OpenAI has achieved notoriety and a mass market, its main challenge is to replicate that ability to monetize its products, something Google has successfully done for decades.
Believe me, it’s no easy task. Training such powerful models, maintaining them at scale, and continually innovating is complex and, above all, extremely expensive.
In fact, we recently learned that industry giants like Google are exploring the possibility of using nuclear reactors to train their models. Yes, you read that right: nuclear reactors. This gives us an idea of the immense amount of energy resources needed to power these large AI systems. On the other hand, the salaries of the engineers who develop and maintain these technologies are not exactly cheap either.
Now we reach the moment of truth: Is it worth paying the $200/month?
Let me give you two answers from two different perspectives:
- Yes, absolutely: You can find multiple comparisons online. If you’re a large company, a professional developer, or someone who really makes the most out of every cent of ChatGPT, the investment can be very profitable. In fact, for these companies, the price will seem quite affordable for the value they receive. And even the $20/month option still feels almost like a gift for many of us who use ChatGPT daily… for now.
- You have no other choice: Companies are there to make money, or at least not to lose it. All of them — be it Anthropic, Google, or OpenAI — have invested huge amounts of capital to get to this point. If they don’t want to disappear, they need to recoup their investment.
Are there any alternatives for now?
In life, there are always alternatives, but it’s clear that at least for the time being, they won’t be as powerful as OpenAI’s models. Still, for certain tasks, they may be more than sufficient.
If you found this insightful, I’d love for you to follow me on LinkedIn: Ángel Molina. You can also follow my company MOLA DATA for more updates. But most importantly, don’t miss out on my course on automating data flows with KNIME: Join here. Let’s shape the future together!
🧑💻 LinkedIn 🐦 (X) 🌄 Instagram ⏰ TikTok 📘 Facebook ⏯️ YouTube
