As artificial intelligence (AI) continues to advance at full speed, some industry experts predict the arrival of artificial general intelligence (AGI) within the next decade. This form of AI matches and even surpasses human-level intelligence.
In a conversation with Ben Goertzel, CEO at SingularityNET, and David Vivancos, Scientific Advisor at Qubic, BeInCrypto learned about what needs to be refined for AGI to arrive and how it will impact the cryptocurrency market as we know it today.
Transitioning from AI to AGI
AI refers to the development of computer systems capable of performing tasks that typically require human intelligence.
These include learning, comprehending information, solving complex problems, making decisions, exhibiting creative behavior, and operating autonomously.
Last year witnessed breakthroughs specific to generative AI, a technology that can create original context for text, images, and video.
“Current AI tech has functionality that is impressively broad and general in some ways, but it achieves this generality via ingesting a huge and general training dataset, rather than by making creative leaps of generalization,” Goertzel told BeInCrypto.
As AI continues to refine itself, researchers, industry experts, and the general public wonder when technological superintelligence will arrive.
Since the inception of ChatGPT, the scientific community has started to consider the concept of AGI with increasing seriousness.
Unlike traditional AI, which guides itself by reading data and interpreting algorithms, AGI emphasizes the integration of logic and reasoning into its learning processes, seeking to emulate human cognitive capabilities.
“To frame it correctly, reaching AGI will mean that we build a human level artificial entity that works 24/7 (x) times faster than a very smart human, if we have enough computational capabilities for that (x) to be bigger than the 8 billion people living on planet earth, then we will have reached ASI or Artificial Super Intelligence,” Vivancos said.
While AGI does not currently exist and remains a subject of ongoing research and experimentation, its development could lead to significant advancements. It would enable AI systems to process and analyze data at speeds and sophistication that surpass current technologies’ capabilities.
Disagreements Over AGI’s Expected Arrival
Though there have been a string of breakthroughs in artificial intelligence over the past year, AI researchers and industry experts do not agree on how soon AGI can be achieved.
In December, OpenAI CEO Sam Altman predicted that the industry would begin seeing the first AGI examples in 2025.
Last April, Tesla CEO Elon Musk predicted the development of smarter-than-human intelligence by 2026. Anthropic CEO Dario Amodei agreed with Musk’s calculations in a separate statement.
Some industry leaders like Google DeepMind co-founder Demis Hassabis expect AGI-like systems to arrive within the next decade.
In his interview with BeInCrypto, Goertzel shared this prediction.
“Only recently do we have the compute power needed to approach human-level AGI without extreme amounts of algorithmic trickery. Now that we do, I think we will probably get to human-level AGI within a few years,” he said.
Vivancos employed a different approach in his evaluation. He broke down progress made in different sectors of AGI development.
“We are on average about 53% completed in the journey towards AGI, but it is true that in some areas it is practically already at 100% and in a few other at 0%,” he said.
But other experts have expressed even more tempered expectations.
Yann LeCun, Meta’s chief AI scientist, has repeatedly declared that AGI is nowhere near completion.
“All of this is going to take at least a decade, and probably a lot more, because there are a lot of problems that we’re not seeing right now. So we don’t know if there is an easy solution within this framework,” LeCun explained in an interview with Lex Fridman.
However, before entertaining the idea of AGI, hurdles related to its development need to be examined.
Challenges Facing AGI Development
Most researchers agree that several challenges must be addressed to develop true AGI.
“The main challenge is that we don’t know how the human brain works, so replicating it is tricky,” said Vivancos.
Today’s AI systems can recognize patterns in complex data and learn through trial and error. However, they struggle with truly complex tasks like deep thinking or understanding and using language like humans.
The finiteness of computational resources is another issue.
“Also, we are bounded by our current computation capabilities, energy storage, and generation,” said Vivancos.
A computer requires incredibly powerful hardware to emulate a human brain. This undertaking involves immense memory capacity to store the vast amount of information the brain processes and incredibly fast processors to handle this information at the speed required for human-like thought.
Furthermore, these powerful components must be seamlessly interconnected. They need to allow for the rapid and efficient transfer of information within the system to mirror the complex neural networks of the human brain.
Creating a truly brain-like computer demands computational power and interconnectedness that surpasses current technological capabilities.
Difficulties in Replicating Nuanced Human Behavior
AGI encompasses the capacity to accomplish various goals and tasks across different contexts and environments.
But unlike humans, technology does not register the environment that it lives in.
“There are many unresolved research questions we need to work through to get to human-level AGI. No one yet understands in detail, for example, how abstract conceptual knowledge needs to relate to perceptual knowledge or specific actions in a human-like brain, nor how attention and awareness emerge from the collective activity of large self-organizing networks of processing elements,” Goertzel said.
Vivancos coined his very own term to describe this phenomenon. He refers to Embodied Artificial General Intelligence, or E-AGI, as the convergence of biological and machine intelligence.
“For me, a critical factor is the lack of “physical” form, where our senses and actuators build real models of the world, and support the imagined ones the brain constructs,” he said.
This approach highlights the significance of the physical body and its interactions with the environment in developing machine-automated intelligent behavior.
It argues that true intelligence cannot be fully understood by solely focusing on abstract thought processes. Instead, it emphasizes the importance of examining how living beings interact with their physical surroundings and how they perceive and respond to sensory information.
Nonetheless, even without having embodied intelligence, AGI could greatly impact different areas of society.
The Potential Impact of AGI on Crypto Trading
AGI can completely re-envision the way that cryptocurrency markets work today.
“Based on the current growth in compute, if AGI is reached within the decade, it can beat by far best human crypto trader working with the best current trading bots, and it will disrupt the market, depending on how many of them are there,” said Vivancos.
To that point, Goertzel added:
“An AGI with roughly human-level general intelligence would likely vastly exceed human capability in financial matters including trading, strategic investment and risk management. So the financial markets would be almost totally dominated by AGIs, and to the extent the money-driven market economy was still a thing, it would be taken over by AGIs and those who own them.”
Once AGI is achieved and its capabilities are refined over time, so will its capacity to autonomously trade cryptocurrencies more efficiently.
“Imagine the current trading bots on steroids but with the capability to understand the nuances that a human expert trader can extract from the news and all the open and close data that flow every instant, and use it to track the subtle nano-trends that influence the price and the macro ones, all at the same time,” said Vivancos.
AI meme coins were one of the leading crypto narratives of 2024. These projects use AI to foster deeper user interaction through personalized content, meme campaigns, and real-time analytics.
With the rise of AGI, its influence on meme coins promises to have an even greater impact.
“One suspects an AGI would be better than humans at guesstimating which meme coins are likely to take off in a given culture at a given moment. As well, AGI market makers would be better than human market makers at coordinating mem ecoin tokenomics,” Goertzel said.
That being said, the market capitalization of AI agent tokens has surged to $13.3 billion.
Top AI Agent Coins. Source: CoinGecko
How Blockchain Technology Can Be Used to Rollout AGI
Currently, AGI development is concentrated in the hands of a few companies. Major players include Anthropic, Google DeepMind, IBM, Microsoft, and OpenAI.
“Currently beside a few open initiatives, 3 or 4 big closed companies/labs are at the forefront of AGI development, and transparency can only be loosely inferred since they don’t share the ‘weights’ or the neural architectures being built nor the training data used to build them. Also, they amass the best GPUs,” Vivancos told BeInCrypto.
In such cases, blockchain technology’s decentralized nature could be leveraged to increase transparency and democratize access to relevant information.
“Using a blockchain infrastructure, one can create an AGI system that runs across a huge network of globally distributed machines, with no small group of owners or controllers,” said Goertzel.
Blockchain technology can help AGI development by offering a decentralized and secure target=”_blank” rel=”noreferrer noopener”>Blockchain’s immutability features also ensure that the data cannot be altered or tampered with after it is recorded. Verifiability allows for the easy corroboration of the authenticity and origin of the data.
Meanwhile, traceability enables tracking data movement and usage throughout its lifecycle, promoting transparency and accountability.
“Blockchain seems the only available technology with a prayer of enabling AGI to be developed for the good of humanity as a whole rather than the narrow ends of particular companies or corporations. This does not guarantee democratic or ethical AI, but at least it creates a framework within which democratic and ethical AI can plausibly be unfolded. How to use this sort of framework to work toward AGI that is compassionate and broadly beneficial is an important and interesting challenge,” Goertzel added.
Though AGI has the potential to completely solve or vastly improve many of the issues humanity faces today, risks need to be mitigated effectively to avoid adverse situations. The deployment of this technology could help achieve this goal.
AGI’s Role in the Future of Civilization
The impact that AGI can have on society is enormous.
“An AGI with even modestly superhuman general intelligence would likely be able to solve most of the problems that now vex our societies, such as aging and death, material scarcity, mental illness and so forth. So if we can create an AGI that is beneficially oriented toward humanity, we will see a shift to a mode of existence that will feel relatively utopic compared to the current situation,” Goertzel told BeInCrypto.
However, guiding the fair and equitable distribution of AGI will be indispensable for this to happen.
“AGI could help solve most of the human problems, like health, education, environment, and all the others you can imagine, but it can exacerbate the inequalities too vastly, ethics is a human responsibility for now,” Vivancos added.
Goertzel highlighted the dangers associated with the limited ownership of AGI technology and its consequent centralization.
“There is also the risk that the first AGIs are created, not to love and help people, but rather to serve the narrow needs of some particular government or corporation… and then when it ascends to superhuman intelligence, it continues to maintain a narrow-minded orientation rather than a broad-minded and inclusively compassionate orientation,” he said.
The role of blockchain technology could be particularly helpful in distributing AGI equitably.
“I think what we need is for the first breakthrough human-level AGI system to be rolled out on a decentralized blockchain-based network with meaningful participatory, democratic governance and well-designed, non-corrupt tokenomics. If this happens then the transition from the pre-AGI to the post-AGI era may go a bit more smoothly,” Goertzel added.
Human involvement in every step of the AGI developmental process will be crucial.
Responsible Guidance in the Age of AI
Scientists and researchers have a wide consensus that once AGI arrives, civilization will change completely.
“AGI will be what I call the “last” human tool, so it is a strategic resource like no other, foreseen that the countries that harness it will thrive and be able to survive and compete, so cooperation is going to be tricky,” Vivancos explained.
Given the negative outcomes that can come with the improper deployment of this technology, ensuring transparency from the beginning is vital. Effective human oversight can help make this happen.
“AGI can help us leverage this and give all humanity an equal chance in all our pressing issues. But as long as we are all part of the process, AIs are and will be biased as all humans are. But if we “teach” now AI’s with best human traits, future AGIs it could be less biased and more aligned with agreed human values, mitigating the risks of the last technology that will redefine humanity like no other before,” Vivancos added.
Successful development and deployment of AGI will require a multifaceted approach. Ethical development, equitable distribution, and ongoing human oversight can contribute to a future where AGI serves as a force for good, benefiting all of humanity.
Read the full article here
