Debunking the Hype: Artificial General Intelligence (AGI) by 2027?

The conversation around Artificial Intelligence (AI) is intensifying, with headlines proclaiming imminent breakthroughs. One prominent voice is Leopold Ashen Brener, a former OpenAI employee, who claims that Artificial Superintelligence (ASI) is just around the corner. In a recent 165-page essay, he elucidates why he believes AGI will surpass human intelligence by 2027. While his arguments are compelling, there are reasons to approach such predictions with skepticism.

” alt=”Artificial Intelligence future predictions” />

The Case for Rapid AI Advancement

Ashen Brener argues that burgeoning computing power and continuous algorithmic improvements are driving exponential AI performance gains. According to him, factors such as advanced computing clusters and self-improving algorithms will soon make AI outperform humans in virtually every task. He suggests that these advancements will continue unabated for at least a few more years, making AGI a tangible reality by 2027.

“The most relevant factors that currently contribute to the growth of AI performance is the increase of computing clusters and improvements of the algorithms.” – Leopold Ashen Brener

While I agree with his assessment that exponential improvement can lead to significant breakthroughs, the pragmatist in me questions the feasibility of his timeline. My background in Artificial Intelligence and Machine Learning informs my understanding, and I believe there are significant hurdles that need addressing.

Energy and Data: The Unsung Limitations

One of the major oversight in Ashen Brener’s predictions involves the massive energy consumption required for training and running advanced AI models. By his own calculations, advanced models will demand up to 100 gigawatts of power by 2030, equating to the output of about a thousand new power plants. This is not just a logistical nightmare but also a financial one – the costs will run into trillions of dollars.

” alt=”High power consumption of AI” />

Additionally, he dismisses the challenge of data requirements. As models grow, so does their need for data. Ashen Brener proposes using robots to collect new data, yet he underestimates the complexity of creating a robot-driven economy. Developing, deploying, and scaling a global robot workforce is not just a technical issue but one that requires a seismic shift in the current economic structure, likely taking decades to accomplish.

“By 2030, they’ll run at 100 gigawatts at a cost of a trillion dollars. Build 1,200 new power stations? You got to be kidding me.” – Me

My assumption is that AGI will indeed unlock monumental scientific advancements. AI’s potential to analyze vast amounts of existing scientific literature and prevent human errors is an undeniable advantage. However, this does not mean a rapid, uncontrollable intelligence explosion. Historical overestimations by prominent figures, such as Marvin Minsky in the 1970s and Herbert Simon in the 1960s, serve as reminders to temper our expectations.

Security and Ethical Implications

Ashen Brener also dedicates part of his essay to discussing the geopolitical tensions that AGI could exacerbate, mainly focusing on a U.S.-China dichotomy. He warns that as governments wake up to AGI’s full potential, they will compete fiercely to gain control over it, likely imposing stringent security measures. This is plausible but reductive, neglecting the broader global context and the impending climate crisis.

“The world economy is about to be crushed by a climate crisis, and people currently seriously underestimate just how big an impact AGI will make.” – Me

The risks associated with AGI are indeed enormous, from ethical considerations in deployment to potential misuse in warfare or surveillance. As someone who has worked extensively in cloud solutions and AI, my stance is that these security issues highlight the necessity for robust governance frameworks and international collaborations.