Investors might be feeling a growing unease about the profitability of generative AI, but the technology itself is still advancing at an impressive pace. It’s on track to not just meet but potentially surpass human cognitive abilities.
In the previous year, there was a flurry of reports suggesting that elite AI research labs like Google, OpenAI, and Anthropic might struggle to create more advanced AI models due to a scarcity of high-quality training content. However, this notion was challenged by Eric Schmidt, Google’s former CEO, who argued that there’s no real evidence indicating that the scaling laws have hit a wall. OpenAI’s CEO, Sam Altman, also chimed in, insisting, “There is no wall.” It seems these industry leaders are confident in the ongoing evolution of AI.
We might be on the cusp of achieving artificial general intelligence (AGI), which is a form of AI meant to handle any intellectual task a human can perform. Unlike its narrow AI counterparts that specialize in specific functions, AGI would boast the versatility and ingenuity of human intelligence. Reflecting on this, Anthropic’s CEO Dario Amodei suggested in a Wall Street Journal piece that AI could be outsmarting humans by next year.
Speaking at the Journal House in Davos, Amodei shared his thoughts: “I can’t pinpoint exactly when AGI will arrive—it might be sooner or later than 2027—but it’s feasible we could see AI outperforming humans in almost everything. Eventually, it could surpass every human in every task, even in robotics.”
Amodei’s insights come at a pivotal moment when the demand for AI computing power is skyrocketing. For instance, OpenAI’s Sam Altman recently acknowledged their struggle in deploying the new GPT-4.5 to ChatGPT Plus users due to GPU shortages. The company simply couldn’t satisfy the computing needs.
Anthropic is facing similar challenges, with Amodei hinting that they might end up using over a million chips to power their AI developments, like Claude, by 2026. While engaging with the Wall Street Journal, Amodei observed that as AI advances, it may accelerate the evolution of robotics. He remarked, “If we create AI systems that are sophisticated enough, they’ll contribute to building better robots. When that unfolds, we’ll need to discuss—especially at events like this—how we structure our economies. How will humans find meaning when AI takes over many roles?”
This raises concerns about AI-driven robotics possibly replacing human workers rapidly. If we reach this point, it’s crucial to reassess our values, particularly how we value work and humanity.
In parallel, Microsoft co-founder Bill Gates has made predictions that AI could gradually edge humans out of many jobs. Gates acknowledges that the swift rise of AI is unsettling people, as it implies that very soon, humans might not be necessary for numerous tasks in this new era of innovation. Yet, he insists it will ultimately be up to us to decide which aspects remain human domain: “There will be some things we keep to ourselves, but eventually, making, moving, and growing things will predominantly be AI’s work.”
Interestingly, last year, Sam Altman projected that AGI would become a reality within five years, suggesting it could “whoosh by” with minimal societal upheaval, despite the significant technological leap.
