Have you noticed that interest in ChatGPT and its usage has started to decline recently?

Some claim that AI is just a temporary hype, and once the excitement dies down, the entire field will disappear like a soap bubble. These skeptics believe that AI has thrived solely due to inflated expectations, and when that “fuel line” stops pumping, the whole structure will collapse.

However, over my long career, I have witnessed too many innovations, research, and implementations to fall for illusions. I call those who deny the value—or even the very existence—of AI, AI deniers.

But absurd views can also be found on the other side of the spectrum, among those I refer to as superhumanity evangelists. These people are convinced that artificial general intelligence (AGI) is already at the doorstep, with some even predicting the arrival of the singularity—a term coined by Ray Kurzweil. Just take a look on YouTube, and you will find dozens of channels filled with panic-driven predictions about the future of AI.

Over the years, I have heard a lot of fantastic forecasts about how AI will change the world: from the domination of malevolent conscious machines to more grounded, yet still exaggerated, critiques claiming that AI will never become “truly intelligent.”

We are living in an era where the line between fact and fiction is blurrier than ever. It’s hard to distinguish objective data from speculation, especially considering the spread of fake images, AI-generated text, and misinformation. So how can we understand the real value of AI? And where is generative AI really heading?

What will AI bring?

If ideas about AGI and the singularity remain speculative, what will AI actually deliver in the near future?

Here, it’s crucial to refer to Eric Siegel’s words: “The antidote to hype is focusing on concrete value.” This simple yet brilliant insight highlights the key criterion—real, tangible benefits for users.

Real value does not come from flashy demonstrations or technological tricks that stoke interest in AI. It emerges in products that solve specific problems, such as Midjourney for image creation, ELSA AI for language learning, or Grammarly for text improvement.

Take Grammarly, for example. Its value lies not in impressing us, but in genuinely helping users by not only correcting errors but also enhancing sentence structure, making the text sound more natural. For people like me, for whom English and Polish are not native languages, this service has become a lifesaver. I use it almost daily, and for a good reason, I pay for its subscription (though I have no commercial ties with Grammarly—it’s simply a product I genuinely like).

Yes, demonstrations like ChatGPT or the Sora video generator from OpenAI are impressive. They showcase the potential of technology, but an important nuance is that ChatGPT was a demonstration tool, not a product designed to solve a specific problem.

Since its sensational launch, ChatGPT hasn’t been particularly good at any specific task. It’s just a general chatbot that can be used occasionally. But as the novelty wears off, users return to their real lives and stop using it regularly.

And this is natural. ChatGPT, like many other demos, doesn’t provide real value. That value appears in the next stage of AI development.

The first wave of generative AI consisted of demos like ChatGPT and Sora. The second wave is now about real products with clearly defined value. We’re already seeing examples of such products: Grammarly, Midjourney, Udio, and others. These offer real utility and solve specific problems.

However, creating such products is an incredibly difficult task. Turning a technological idea into a product that is in demand is a real challenge. Many startups struggle with implementation and finding real market demand. AI projects often solve non-existent problems instead of focusing on real value.

Here are examples of products that already demonstrate their value and effectiveness in practice:

1. Quillbot is a tool competing with Grammarly, offering not just error correction but also text rephrasing. Quillbot uses powerful machine learning algorithms to help improve text structure and style. Its strength lies in generating alternative phrasings, which is especially useful for non-native speakers. Like Grammarly, this tool has a clear purpose—improving written language.

2. Photoroom is an image editing tool that competes with Adobe, providing advanced AI tools for photo editing. Photoroom makes the process of creating high-quality images accessible to people without professional design skills. Its value lies in significantly simplifying routine tasks, such as background removal or creating marketing materials.

3. ElevenLabs is a startup offering AI-powered voice generation based on text. This tool stands out for its high-quality speech synthesis, which approaches the natural sound of a human voice. ElevenLabs can be used in video dubbing, podcast creation, or audiobooks, making it a valuable tool for content creators.

4. Udio is an AI product that generates music based on text prompts or uploaded audio. Its uniqueness lies in not only creating music but also generating song lyrics, giving users the ability to create their own musical compositions. Udio makes music creation accessible to a broad audience, even for those without any musical experience.

5. Vocal Remover is a tool that allows separating vocals and instrumental parts in music tracks. This product is perfect for creating karaoke tracks or remixes. Its value lies in its ease of use and high-quality results. I tested Vocal Remover on Pat Benatar’s legendary song “We Belong,” and the software easily separated the vocals and instrumentals, creating two high-quality tracks. This is an excellent example of a product with a clearly defined function.

We are on the verge of the second wave of generative AI, which will resemble the Cambrian explosion: many products will hit the market, and many will disappear. But the few that survive will form the foundation of a new technological era, bringing real value.

Success in the field of generative AI will require not only technological ingenuity but also the ability to manage inflated expectations and false hopes that have developed around this area.