Is AI magic?
Is equating AI with casting spells and wielding wands an accurate analogy for AI's popularity and Web3's lack of mass adoption?
In 2021 and 2022, we were in the midst of the cryptocurrency bubble and the peak of the public's interest in trends like Web3 and NFTs. The promise was mainstream adoption, fuelled by a word-of-mouth strategy (proven to be successful to some extent by projects such as NBA Top Shot) that would support it, along with the prediction that true asset ownership and the control users had over their assets would create a free market, unconstrained by regulations, and would act like a magnet for the public. Gaming was often mentioned as the main driver (examples can be found here), but apart from Axie Infinity, which reached the mainstream, and games like Alien Worlds, Splinterlands, or Planet IX, with a notable number of users, nothing else came as close.
Then came ChatGPT, Dall-E, and the vast variety of tools that have their roots in concepts such as machine learning, statistics, and neural networks. Most of these solutions are equivalent to a Swiss Army Knife and are popping up every day. However, to use some of them, one has to have the tech fluency of a programmer. This is similar to the crypto space, where to use the tooling, one has to make an effort (a crucial for using wallets, DEXes etc.), and be tech-savvy to some extent. But despite this complexity, AI-based features seem not only to be a must in software nowadays (e.g., Microsoft, Canva) but also must be user-friendly. They must not be forced upon the user but provide value and solve a real problem and be ease to use. Web3 failed to do that, but Generative AI tools provide just that. And it stuck, not only within the software industry but with the general public. It stuck because it was similar to magic—presentations generating themselves, AI-based assistants performing tasks at a push of a button or given a voice-base command, ideas coming to life through prompts (which can count as a magical factor—a spell has to be pronounced in a certain way to work, doesn't it?).
What caught my attention are the types of pictograms and icons that are being used with AI tools. The logical choice would be those of robots and pictograms based on them or mechanical tooling. This is evident when we observe companies that are focused on high-end and complex concepts—and even lighter ones like PatentPal where the robot appears to be a very friendly character. However, what is interesting is when we delve into the specific GenAI-driven functions of a given tool or solution. Many of them seem to be related to magic, once again.
Some projects have even used the term as a synonym for the capabilities of their solution.
As we can see, performing AI-related tasks that require repetitive, tedious, and often mechanical actions (e.g., creating presentations, asking questions) is becoming as easy as waving a wand and chanting a spell.
In his book, "The Coming Wave," Mustafa Suleyman claims that "what seems like magic engineering one day is just another part of the furniture the next." The magic engineering Suleyman refers to encompasses utterly complex concepts that are not for the casual user to grasp. However, these are hidden beneath a fairly accessible interface and the magic wand icon. While the popularity among the general public does not rest on developers' decisions about which icons to use for a GenAI function, it reflects the sentiment. Crypto is hard; it requires a learning curve, but with GenAI tools, all it takes is the push of a button and things “just” happen.
Magically.