All fully written by Isaiah, 9/12/2025
The AI revolution of the past three years has been a phenomenal sight to see, as everyday it seems like we're witnessing technology reach heights we didn't think were possible in our lifetime. However, it's also been a smokescreen that presents a story that isn't entirely true. Looking beyond all the smoke and mirrors is a way less harsh reality: AI's rise to glory has been in the works for way longer than the past 3-5 years. And while it's shook the world majorly throughout this boom period, it's still way lower on the Richter Scale than some experts may claim. We're likely some years away before AI is the earth shattering socioeconomic earthquake that many of us fear it soon will be.
Let's first analyze AI's timeline. AI (artificial intelligence) has been accessible to the masses as early as the first video games. The same tech needed to power video games (GPUs and machine learning concepts) is the backbone of what powers AI. Think about it, when you play a video game, the "CPU" is smart, responsive, and intuitive.
On the consumer scale it really took major leaps with the arrival of smart home/IoTs (Amazon Alexa, Google Home, Apple TV, Ring, etc.) in the late 2010s. The ability for household essentials to be "smart" enough to compute tasks autonomously and learn your habits and preferences is once again the bedrock of how AI works.
As time progressed, we began to witness even more instances of it in the technology we know and love: autocomplete on Gmail, subtitles/transcript tools on YouTube, and social media recommendation algorithms. The space took another major leap in the late 2010s with self-driving cars. As AVs like Tesla's demonstrated just how independent and dynamically learning AI could be, they also introduced one of the best use cases of AI, computer vision. The eyes that allow AI to see what's in front of it and plug visual comprehension into its dynamic learning patterns.
Then in 2023, ChatGPT arrived and shook the world as the first of its kind LLM (Large Language Model), essentially a perfect blend of yesterday's technology trained on mountains of data. AI's entire intelligence is predicated on its knowledge base: the more it knows, the better it performs, and the inverse holds true. Now LLMs are capable of what's called "generative AI," a functionality in which their giant brain of training data allows them to take user input and generate conversational responses, or even images, diagrams, video, etc.
The issue is that since AIs are limited to what they know, and they store so much data just to stay afloat, they often "hallucinate," meaning they forget what they know or spit out untrue information because their training data is full of falsehoods.
So while AI is awe-inspiring, and in its current iterations has massively streamlined productivity and unlocked whole new possibilities of what technology can do, it's still just a power drill for creative construction, not a one-man construction crew.
The Over-Sensationalized AI Bubble
The over-sensationalized hype surrounding AI has unfortunately cultivated what myself and many experts deem to be a bubble with dot-com-like implications. Part of why there's such an illusion around AI's capabilities is the staggering amount of money big tech and VCs alike are pouring into anything with "AI" in the name.
In 2024 alone, Meta committed over $50 billion to AI development, Google pledged $25 billion, and more than 50% of employers are now building their own internal AI workflows. The result has been a frenzy where entire colonies of startups, many with little to no real AI capability beyond the letters "AI" in their branding, are receiving overvalued funding rounds or acquisitions. Investors are still not seeing returns anywhere near the scale of their outlays.
Even the leaders, Meta and OpenAI, have introduced frontier, consumer-facing products that millions use and love. Yet both continue to suffer from major shortcomings: hallucinations, misinformation, bias, and easy manipulation. Consequently, it almost seems inevitable that unless the technology begins closing the gap between perceived potential and actual performance, a market collapse may be imminent.
As alluded to earlier, the main crux of AI's shortcomings is its absence of frontier training data. While recycled news articles and social media posts help mold AI's "personality," its understanding of context and nuance remains shallow. That's where SocialDNA emerges as the obvious brain fuel AI needs to continue developing and maturing.