One well-known expert on artificial intelligence (AI) is Ray Kurzweil, an American computer scientist and techno-optimist.
In his best-selling 2005 book, “The Singularity Is Near,” Kurzweil forecasted that computers would reach human-level intelligence by 2029 and that humans would merge with machines by 2045, a concept he termed “the Singularity.” Now, nearly two decades later, Kurzweil returns with a sequel, “The Singularity Is Nearer,” where he reflects on the progress made and reaffirms his predictions. Principal researcher and AI visionary at Google is Kurzweil’s day job. He addressed the Observer as a futurist, writer, and inventor.
Why was this book written?
The future was discussed in The Singularity Is Near, but that was twenty years ago before AI was widely understood. Though not everyone understood what would happen, he knew exactly what would happen. These days, AI rules the discourse. Large language models (LLMs) are a pleasure to use, but it’s time to reexamine both our past achievements and our upcoming innovations.
Kurzweil’s predictions for 2029 and 2045 remain unchanged.
He has continued to be reliable. Thus, 2029 will mark both the advent of artificial general intelligence (AGI), which is somewhat distinct, and human-level intellect. Human-level intelligence, which will be mostly attained by 2029, is commonly defined as AI that has developed to the level of the most proficient humans in a certain field. (Although it will, there might be a transitional period after 2029 during which AI does not yet outperform the best humans in a few critical domains, such as screenplay writing that wins Oscars or producing profoundly novel philosophical ideas.)
Also Read: South Korea’s SK Hynix Announces $74.6 Billion Investment in AI Chips Development
Artificial general intelligence (AGI) refers to AI that is more advanced than humans in all areas. Though it sounds more complicated, AGI is being introduced concurrently. Furthermore, his five-year projection is genuinely conservative: According to Elon Musk, it will occur in two years.
What makes Kurzweil’s dates credible?
In actuality, he is the only one who foresaw the current surge in interest in AI. People believed that would take a century or longer in 1999. “Look what we have after thirty years,” he said. The exponential increase in processing power at a given price in constant dollars is the main motivator. Every fifteen months, price-performance is doubled. Due to the increase in computation, LLMs have only been in use for the past two years.
What is needed now to advance AI to the point that Kurzweil anticipates it will be in 2029?
One is increased processing power, which will happen soon. This will allow for advances in areas where inadequacies still exist, such as contextual memory, common sense thinking, and social interaction. Then, to address more queries, we’ll need more data and improved algorithms.
Also Read: Morgan Freeman Calls Out AI Voice Misuse: A Call for Regulation
By 2029, LLM hallucinations—which produce illogical or erroneous outputs—will be far less of an issue; they arise far less frequently now than they did two years ago. They don’t know that they don’t have a solution, which is why the problem arises. They search for what seems best, even if it’s inappropriate or incorrect. As AI develops, it will be able to recognize its expertise more clearly and tell humans when it is ignorant.
What is the Singularity precisely?
We can never get smarter than the size of our current brains. However, the cloud is becoming increasingly intelligent and expanding virtually infinitely. The physical metaphor known as the Singularity will happen when our brains combine with the cloud. Our natural and cybernetic intelligence will be combined to form one cohesive entity. Brain-computer interfaces, which eventually consist of nanobots—robots the size of molecules—that enter our brains through capillaries without causing any harm, will be the technology that makes it possible. By 2045, human intelligence will have multiplied a million times, raising our level of awareness and consciousness.
Also Read: Microsoft Unveils ‘Skeleton Key’ Attack Exploiting Generative AI Systems
Though it is difficult to conceive, this doesn’t sound very attractive.
Imagine having your phone in your head instead of your hand. Similar to how you currently use your phone, your brain will be able to query the cloud for an answer when you ask a question. The main differences are that this process will be instantaneous, there won’t be any problems with input or output, and you won’t even be aware that it’s finished—the answer will just appear. When someone says, “I don’t want that,” they also mean that they don’t want a phone!
What about the existential risk posed by sophisticated AI systems—the possibility that they will develop unexpected abilities and gravely harm humankind? The “godfather” of AI, Geoffrey Hinton, departed Google last year partly due to these worries, and prominent industry figures like Elon Musk have also voiced cautions. Workers at OpenAI and Google DeepMind demanded earlier this month that those who disclose safety issues be given more protection.
There is a chapter on dangers in Kurzweil’s book. To determine the right course of action, he has been involved in the establishment of the Asilomar AI Principles, a non-binding set of rules for responsible AI development that was published in 2017. We must monitor what AI is doing and remain mindful of the potential here. However, it makes no sense to just oppose it when the benefits are so great. Positively, all the big corporations are working harder to ensure that their systems are safe and consistent with human values than they are to develop innovations.
Also Read: Google Cloud AI Gemini 1.5: Flash and Pro Versions Now Available
Won’t computational power have physical bounds that put a stop to it?
Our current state of computers is essentially flawless, and it will only continue to improve with time. We can keep making chips better in a lot of ways. We’ve only recently started to utilize the third dimension, which will last us for a very long time—creating 3D chips. I don’t think we need quantum computing because its benefits have never been proven.
Kurzweil predicts that by 2029, artificial intelligence (AI) will be able to communicate through text just like a person. But AI will have to get dumber to pass it. In what way?
Humans are not very accurate and have limited knowledge! Today, you can ask an LLM a particular question about any theory in any discipline, and it will provide you with a very thoughtful response. But who on earth is capable of that? You would be able to tell whether it was a machine answering if a human did. Because the exam is attempting to mimic a human, that is why it has been made simpler. There have been reports that GPT-4 is capable of passing a Turing test. He believes we still have a few years to go before we settle.
Also Read: Google Launches AI-Powered Google Vids for Easy Video Creation
It’s unlikely that everyone will have access to the technology Kurzweil envisions for the future. Does he care about technological inequality?
Rich people can purchase new technologies early on, but it’s also at a time when their functionality is limited. When [mobile] phones were first released, they were incredibly costly and had subpar performance. They didn’t communicate with the cloud and had very limited access to information. These days, they are very practical and reasonably priced. Approximately 75% of individuals worldwide own one. Thus, the outcome will be the same: this problem eventually goes away.
The book examines AI’s potential to eliminate jobs in great depth. Do we need to worry?
Both yes and no. People will be impacted by the automation of some job types. However, new skills also generate new employment. Even ten years ago, a position like “social media influencer” made no sense. There are more jobs than ever before in the US, and when adjusted for current currency, the average personal income per hour worked is ten times more in the US now than it was a century ago. Beginning in the 2030s, universal basic income will lessen the negative effects of job disruptions. That won’t be sufficient then, but eventually, it will be.
Also Read: US Regulators to Intensify Antitrust Scrutiny on Microsoft, OpenAI, and Nvidia
Beyond the loss of jobs, AI is also predicted to drastically alter society in other unsettling ways, such as disseminating false information, creating harm through biased algorithms, and escalating surveillance. Kurzweil avoids focusing too much on those.
There are some difficulties that we need to resolve. The impending election raises concerns about “deepfake” videos. He believes that we can genuinely identify [what is phoney], but we won’t have time if it occurs soon before the election. Regarding prejudice, AI is picking up on human bias since humans are biased. Although we’re moving forward, we still have a ways to go. The legal system must be used to resolve concerns over AI’s equitable usage of data.
What role does Kurzweil play at Google, and was there a pre-publication assessment of the book?
He gives them advice on how to expand technology and make their products better, including LLMs. The book was authored by the author personally. Google is pleased that he published these, and no review was submitted.
Also Read: OpenAI Acquires Multi to Develop ChatGPT Desktop App
The forecasts made by Kurzweil regarding digital and physical immortality will not sit well with many individuals. He predicts the arrival of “afterlife” technology in the 2040s that will enable us to upload our minds so they can be restored—even put into convincing androids—if we experience biological death, as well as medical nanobots that will be able to enter our bodies and perform repairs so we can live forever in the 2030s.
Not only is computational power increasing exponentially, but so is our comprehension of biology and our capacity for far smaller-scale engineering. We should anticipate approaching longevity escape velocity in the early 2030s, meaning that for every year we lose due to aging, we will gain one back through scientific advancements. And we will truly regain more years as we go past that. Although there is no absolute assurance that you will live forever—accidents may happen—your chance of passing away won’t rise with each passing year. The potential for the digital resurrection of the dead will raise some thought-provoking ethical and legal issues.
Also Read: Oppo Reno 12 and Reno 12 Pro to Debut in India with Advanced AI Camera Tools
What is Kurzweil’s strategy for eternal life?
His primary goal is to live long enough to flee at a high enough speed. He takes roughly eighty medicines a day to help maintain his health. As a backup, cryogenic freezing is used. In addition, he plans to construct an afterlife AI avatar, which he believes will be available to everyone by the late 2020s. He collected all of his father’s written works and did something similar with him; it was almost like having a conversation. With access to greater resources, [his replicant] will be able to accurately capture his personality.
How can we effectively prepare for the future by acting now?
AI is entering our bodies; it won’t be us against it. It will enable us to produce new items that were previously impractical. It’s going to be an amazing future.
Recommended Readings:
Top 15 Robotics Books for Beginners
9 Best Books On Cryptocurrency
Bill Gates Recommends Must-Read Book on AI and Education
Top 10 Best Fintech Courses, Programs, and Certificates Online
Note: Tech Chilli researches products to bring you our best recommendations. When you buy through our links, we may earn a commission.
