From Luxury to Utility: Why Today's AI Will Soon Be Everywhere
What does this mean for business
Last July, we explored how technologies evolve from precious resources to everyday utilities. Now, DeepSeek's breakthrough in AI efficiency offers a perfect case study of this transformation happening in real-time.
Let's break down where we are and what it means for your business.
Michio Kaku’s Four Stages of Evolution:
Stage I (Precious Resource)
Until recently, advanced AI was the domain of elite labs with deep pockets. Training models like GPT-4 cost upwards of $40 million, making AI development as exclusive as early computers that filled entire rooms and cost millions.
Stage II (Mass Production)
ChatGPT changed everything in late 2022. Suddenly, anyone could access AI capabilities for free or a small fee. Companies like Anthropic (Claude), Google (Gemini), and Meta (Llama) quickly followed, making AI tools widely available. This is similar to how personal computers first brought computing power to businesses and homes.
Stage III (Ubiquity - We Are Here)
We're witnessing the early signs of this transition. DeepSeek's recent breakthrough shows how AI models can be built and run much more efficiently, drastically reducing costs. Just as PCs got smaller and cheaper until they were everywhere, AI is following the same path.
Stage IV (Abundance)
While we're not here yet, the destination is clear. AI capabilities will likely become so commonplace and inexpensive that they'll be embedded everywhere (ubiquitous), often invisibly - much like how we stopped thinking of our thermostats and cars as "computers."
What This Means for Business
Just as electricity transformed from luxury to utility, AI is following the same path. Ben Thompson from Stratechery summarizes three new business realities:
Basic AI capabilities will be cheap and everywhere - standing out means using AI better, not just having it
The money moves from selling AI access to building valuable solutions on top of it - Think Amazon with the internet - they won by using it creatively, not by selling access
Distribution becomes King - Companies with large user bases (Microsoft, Meta, Apple) gain advantage, while AI-only companies must find specific problems to solve.
Your Next Move: Look at your core business processes. Don't ask "How do we add AI?" Instead, ask "If AI was as cheap and available as electricity, what would we build?"