With the rise of AI technologies, widespread awareness about synthetic intelligence has become a prevalent phenomenon. AI will make ! AI can do ! AI will
The latest statement from OpenAI CEO Sam Altman, posted on his personal website this week, strikes a strikingly exaggerated tone.
As he speaks, we’re poised at the threshold of a groundbreaking era, fueled by the promise of superintelligence just a few thousand days in the making. This bold era will usher in “staggering achievements”, including “successfully mitigating climate change, establishing a thriving lunar colony, and groundbreaking advancements in our understanding of physical laws”.
Altman and his firm are seeking to navigate a complex situation, as they attempt to appease US authorities while simultaneously negotiating to acquire a stake in the company.
Notwithstanding the underlying motivations, it is worth scrutinizing several of the presumptions underpinning Altman’s forecasts. Upon closer examination, these proponents of AI’s most ardent enthusiasts betray a wealth of insights into their own worldview, as well as the glaring omissions that have characterized their thought process.
Steam engines for thought?
Altman’s remarkable forecasts are rooted in a succinct historical context spanning two paragraphs.
As human progress has accelerated, people have become increasingly adept at achieving feats previously thought unimaginable, with today’s individuals capable of accomplishing tasks their ancestors would have deemed impossible.
As humanity hurtles forward on an unwavering trajectory of advancement, driven by the insatiable curiosity and ingenuity of our species. The cumulative achievements of scientific inquiry and technological advancements have propelled us towards the development of computer chips, ultimately paving the way for the emergence of artificial intelligence that will propel us forward into the future. The stunning vista is heavily indebted to the pioneering futurism of the cinematic medium.
The allure of a familiar story is tantalizingly straightforward. As human intelligence continues to propel us towards unprecedented achievements, it is difficult to avoid the conclusion that significantly more advanced artificial intelligence will fuel progress even further and more effectively.
That is an previous dream. In the 1820s, as Charles Babbage observed the rapid growth of steam engines during England’s Industrial Revolution, he began contemplating the development of machines that could automate intellectual labor, inspired by the potential of these technological advancements. Charles Babbage’s unrealized vision of a mechanical brain, though unfulfilled, has had a lasting impact on the notion that humanity’s ultimate triumph might lie in mechanizing thought itself, perpetuating a dream that continues to captivate our imagination.
As per Altman’s cinematic vision, we find ourselves standing triumphantly atop the figurative mountain.
Pursuing knowledge with arduous dedication – yet, to what end does one strive so fiercely?
The explanation for why we’re so near the wonderful future is surprisingly straightforward, according to Altman: “Deep learning has worked.”
Deep learning is a specialized form of machine learning that incorporates artificial neural networks, loosely inspired by organic neural systems. Despite initial skepticism, this technology has proven remarkably successful across various sectors: deep learning has fueled breakthroughs in fashion, enabled precise predictions, and even influenced the design choices of certain companies.
The benefits of in-depth research will be far from insignificant. Their effects on society and finance are both profoundly positive and devastatingly negative.
While in-depth analysis may be effective for a limited range of problems, Altman is aware of this:
Scientists have discovered an algorithm capable of learning any distribution of data, effectively unlocking a profound understanding of the fundamental “rules” governing all information.
When you delve deeply into a subject, that’s how learning truly takes hold. While that’s necessary, its widespread applicability across multiple domains is hindered by a single significant limitation.
Not every drawback can be reduced to a simple sample matching scenario. Not all issues necessitate the level of meticulous examination required to effectively address them. Neither is this .
A behemoth of a hammer, ever vigilant for its next victim – a unsuspecting nail.
While intriguing, the notion that guidelines derived solely from scientific knowledge can single-handedly resolve all humanity’s problems is a premise worth scrutinizing further.
There’s a notion that when someone wields a hammer, they’re likely to perceive every problem as a screw in need of tightening – not just nails. Altman grips a behemoth of a hammer, its hefty price tag a testament to the magnitude of his latest acquisition.
Deep study can also be thought of as “work,” albeit only because Altman and others are reimagining the world as a collection of data distributions, a notion that challenges traditional notions of knowledge and understanding. A potentially insidious trend is emerging, where AI is hindering the types of complex problem-solving we’re accustomed to, rather than fostering their development.
What’s often overlooked in Altman’s tribute to AI is the growing infrastructure required for deep learning to truly thrive. While we commend the significant advantages and lasting legacies of modern medicines, transportation systems, and communication methods, let us not overlook their profound costs.
The environmental costs of exploiting natural resources have been steep, exacting a heavy toll from both humans and the natural world: for some individuals, the benefits of progress have yielded diminishing returns, while animals, crops, and ecosystems have suffered ruthless exploitation and destruction at the hands of capitalist extraction and technological advancements.
While Altman’s supporters might downplay concerns over pricing, the question of costs cuts to the heart of discussions surrounding AI’s future.
While acknowledging the limitations of AI, Altman is keenly aware that there are many specifics that still need to be worked out. One of the rapidly escalating costs associated with training AI models is undoubtedly their vitality expenses.
Microsoft recently launched an initiative to build AI knowledge centers and data turbines to power them? Microsoft, which has invested over $10 billion in OpenAI, has also partnered with the owners of the Three Mile Island nuclear power plant, infamous for its 1979 meltdown, to develop a groundbreaking new technology.
As frenzied outlays unfold, whispers of desperation begin to permeate the atmosphere.
Magic or simply magical considering?
Given the enormity of such challenges, it’s crucial to recognize that even if we accept Altman’s optimistic assessment of humanity’s advancements thus far, we still must concede that past progress may not provide a reliable foundation for the future. Sources are finite. Limits are reached. Exponential development can finish.
What’s most striking about Altman’s submission isn’t his hasty predictions. What stands out is the unmistakable essence of his irrepressible enthusiasm for scientific discovery and the relentless march of progress.
While acknowledging the potential drawbacks of technology, it’s crucial to consider how entities like Altman and OpenAI can mitigate these “downsides”? Despite much to accomplish, what’s there to dread about minor obstacles? As artificial intelligence edges closer to unprecedented achievements, what is the point of hesitating?
As hype surrounding AI surges, it’s not an “age of intelligence” unfolding, but rather an “age of inflation,” characterized by inflated resource consumption, overvalued companies, and most alarmingly, exaggerated AI promises.
Indeed, it’s undeniable that some practices we employ today would have seemed like pure sorcery to our ancestors 150 years ago. Not all progress implies a solely positive trajectory between then and now.
While AI undoubtedly holds vast promise across various domains, it’s unrealistic to expect it to single-handedly solve all of humanity’s complex problems.
