Why I'm still sceptical about generative AI
We are just seeing the low hanging fruit
⏰ Reading time: 2-3 minutes
Experiment: This is an opinion piece on a recent tech development. Let me know if you like this at the survey below.
Some people think Generative AI will hit the mainstream market in a year. I’m sceptical. Here’s why.
The cliché warning of job destruction
Everyone indeed has seen the ChatGPTs and Dall-es. People are posting lists of generative AI tools by the bunch on Twitter and Linkedin with warnings that you need to step up your game. AI is here in its final form to take our jobs, right?
Well, are they? Depending on your circles, AI adoption seems to vary. Sure, everyone can play with Dall-e or ChatGPT and be amazed for a few days. And then we get used to the most recent development and our amazement drops, only to be reinvigorated after a new release of the underlying foundational model.
Generative AI has inflated expectations
To explain what’s happening, I use the Gartner Hype Cycle (wiki for explanation). Foundational models are the fundament of all this generative AI.
Last year, Gartner plotted it on the ‘Peak of inflated expectations’, with a time to hit the mainstream in 5 to 10 years. In 2023, Gartner shifted it to 1 - 3 years. I’m sceptical of the predicted product-market fit.
Generative AI has many evangelists
Even though we see all these tools, you should ask yourself, are you using them in your day to day? Who is using it in their day-to-day?
If we think back on tech adoption, innovators & tech enthusiasts get in first. They get in for their love of technology and their love of the potential of technology. This is what the media shows us: tech fanatics explaining the technology.
You see the media's favourite tech nerds demonstrating it on talk shows. Everyone is an expert all of a sudden. Twitter and LinkedIn posts flow in. But: this is not the same as tech adoption. This is just tech evangelism.
Gartner’s take on this stage of a technology is as follows: the only people profiting from it are people giving workshops and conferences on the technology. Speculation and positive tech prophecy are business models on their own.
Generative AI is extremely accessible
This technology is uniquely, extremely accessible. Generative AI is often free for everyone. Everyone can experience it, so everyone can experience wow. Not everyone can experience the HyperLoop in the same way.
Furthermore, it has APIs to integrate with other tools. Normally, it can be quite expensive to double down on a new technology. Building a HyperLoop is impossible as a weekend project.
We are seeing the low-hanging fruit now
But integrating these generative AI things such as Dall-e and (Chat)GPT has been extremely simple. And therefore, ProductHunt has been flushed with generative AI-tools.
It’s text-based and gives text or images back. Everyone that has a text-based app can imagine where generative text comes into play, such as Notions AI assistant. Not extremely creative: low-hanging fruit.
No, I’m not saying it will never take off
The abundance of tools is not an abundance of adoption. I’m sceptical about the current adoption rates of all these tools. I think it still is incredibly niche. Sure, there are some people generating Tweets, social media content, or improving their brainstorming for ideas.
I tried to let it generate article ideas for my newsletter. The articles were vain and uninspiring. I might be biased. Do you know what I use it for, for fun? Generating clickbait titles. And often I stick with my own title.
I think that we humans still need to find the best applications for these new tools and underlying technologies. The low-hanging fruit is very gimmicky and interesting.
But for something to get mass adopted, it needs to be 10x better than any alternative. Give me mass adoption examples, please, I’m open to changing my opinion. I haven’t seen them yet.
How was this opinion piece?
Great - Good - Meh
Should I do more of these? Let me know in the text field after you vote.