The following article is based on our recent FS Insider podcast: Dr Alan D. Thompson on AI, Brain Machine Interfaces, and Intelligence Augmentation. If you’re not already a subscriber to our FS Insider podcast, click here to subscribe.
In recent years, the field of artificial intelligence (AI) has experienced unprecedented growth and breakthroughs, with developments occurring at a pace that's leaving many experts astounded. To gain a deeper understanding of these advancements, particularly in the area of brain machine interfaces and their potential implications for society, we spoke with one of the world's leading researchers and consultants in the AI and robotics space, Dr. Alan D. Thompson at LifeArchitect.ai, who is also the author of The Memo, one of the most widely read AI newsletters on Substack.
Dr. Thompson's Journey in AI Research
Dr. Thompson's fascination and background with AI took a circuitous route starting in 2001 with degrees in computer science and psychology. His interest in intelligence led him to chair Mensa International's Gifted Families program, providing him with unique insights into exceptional cognitive abilities.
However, with the rise of modern large language models (LLMs), Dr. Thompson decided to leave Mensa after witnessing GPT-3 surpass the capabilities of his prodigy clients. Since 2020, he has been at the forefront of this research, conducting independent studies and providing analyses of datasets that are widely read by Fortune 500 companies, governments, and academic institutions worldwide.
As his research shows (see image below), AI continues to make astonishing progress in beating humans on a wide range of IQ tests, exams, and other benchmarks.
Because of his research and real time tracking of AI advancements globally, Dr. Thompson now advises "more than half of the Fortune 500s out there: governments, large research teams, and academic institutions,” he told FS Insider. “But at the end of the day, my focus is on understanding large language models inside of the field of artificial intelligence and then allowing that to be visible to the general public."
Google’s DeepMind Scores Another Breakthrough
One of the most significant recent developments in AI comes from Google DeepMind. Their systems, AlphaGeometry 2 and AlphaProof, backed by the new Google DeepMind Gemini model, recently took the 2024 International Mathematical Olympiad (IMO) exam. This achievement is particularly noteworthy given the difficulty and prestige of the IMO.
"The 2024 International Mathematical Olympiad is huge. It's something that is really revered in math circles. You'll see field medal winners who've also got gold medals at the IMO," Dr. Thompson explains. The AI scored 28 out of 42, just two points shy of a gold medal. This score was assessed by one of the official judges who had already taken the exam.
The significance of this achievement cannot be overstated. As Dr. Thompson points out, "The questions are at a level beyond what I could do, what any of my prodigies could do. There are only a handful of people in the world that can even attempt these questions."
This breakthrough was so unexpected that it prompted the national coach of the United States International Mathematical Olympiad team to compare it to the launch of Sputnik in 1957, stating, "Human civilization needs to move to high alert!"
AI's Performance Across Various Domains
Dr. Thompson tracks AI performance across a wide range of benchmarks, and the results are consistently impressive. Recently, Claude 3.5 Sonnet achieved a 67.2% score on the Google Proof Question and Answer benchmark, surpassing the typical 65% score of PhD domain experts. This demonstrates that AI continues to outperform human experts in specialized fields.
In other areas, the performance of AI is equally impressive:
- Essay writing and SAT exams: GPT-4 is 34% higher than the human baseline
- Social intelligence: 50% higher than the human baseline
- Empathy in medical contexts: GPT-4 showed 70% higher empathy than human doctors, with a 7.8 times higher prevalence of empathetic and very empathetic ratings
Dr. Thompson explains the latter finding: "GPT-4 having 70% higher empathy than human doctors; it's got a lot more time to listen to you. It's also got infinite patience with how it actually comes up with responses."
The Path to Artificial General Intelligence (AGI)
Dr. Thompson maintains what he calls "Alan's conservative countdown to AGI," which tracks progress towards achieving artificial general intelligence - a system that can perform at the level of an average human. As of July 2024, this countdown stands at 76% completion, with full AGI potentially achievable by 2025 or 2026.
"As of today, the end of July 2024, the AGI countdown, or progress, is at 76%, predicted to reach 100%, perhaps sometime in 2025," Dr. Thompson states. He clarifies that AGI refers to a machine that performs at the level of an average or median human, with an IQ of 100.
To reach AGI, Dr. Thompson explains that we need large language models to be integrated into humanoid robots for real-world interaction, improvements in areas like truthfulness and spatial awareness, and continued advancements in overall intelligence. He notes that current AI already demonstrates verbal linguistic intelligence in the 99.99th percentile, with an estimated IQ of 155.
Beyond AGI lies Artificial Superintelligence (ASI), which Dr. Thompson defines as "a machine that performs at the level of an expert human in practically any field." He suggests that ASI could follow within a year or two after achieving AGI, potentially leading to revolutionary advancements across all areas of human knowledge and capability.
"That will be, as some say, a culmination of all intelligence on earth and beyond. So, think of new inventions, new discoveries, new ways of harnessing energy, and really, that changes both the entire earth, the universe, and the way that we evolve as humans," Dr. Thompson explained.
Brain-Machine Interfaces: The Next Frontier
The conversation also touched on the rapid advancements in brain-machine interfaces (BMIs). Companies like Neuralink and Synchron are at the forefront of this technology, which allows people to control digital devices with their thoughts.
Dr. Thompson recalls a conversation he had with Leta AI (an AI avatar he created) in 2021 about the possibility of hooking GPT-3 to his brain. At the time, it seemed like science fiction. However, just three years later, Synchron in Australia has developed an integration between their brain-machine interface and ChatGPT.
"You can see people playing with it, and it's almost exactly what we were doing with Leta. They get a few different options to respond. Every time they hear a question in a conversation or even see something in the world, they get different tiles to choose from to respond to that scenario," Dr. Thompson explains.
Initially focused on helping those with severe disabilities, BMIs are expected to expand to address less severe cognitive impairments and eventually move into the realm of intelligence augmentation for the general population. Dr. Thompson predicts that the next steps in BMI development will involve overcoming learning difficulties and enhancing human capabilities such as memory, creativity, and general cognitive function.
While the timeline for these advancements is uncertain, Dr. Thompson notes that some companies have been in the research and development phase for a decade or more. "We can take Mary Lou Jepsen's OpenWater all the way back to 2015-2016. So, these have been played around within labs for a long time," he says.
Challenges and Skepticism
Despite the rapid progress in AI, there remains skepticism about the feasibility of AGI or ASI in the near future. Dr. Thompson counters these doubts by pointing to the wealth of data and research available, showcasing AI's achievements across various cognitive tasks and standardized tests.
He emphasizes that the advancements in AI over the past three or four years have been so explosive as to overshadow everything since Alan Turing's conception of intelligent machinery in 1948. "So if you haven't kept up with, say, the last 36 months, it's almost as if your knowledge set is irrelevant," Dr. Thompson states.
This rapid progress has left many, including some in academia and government, working with outdated information and underestimating the current capabilities of AI systems. Dr. Thompson notes, "Even in the US, you've got people through Congress working in technology that are not up to speed with basic technology. But then if you look at some of the major universities, they've had trouble keeping up."
Practical Applications and Tools
For those interested in leveraging AI in their work or creative pursuits, Dr. Thompson and FS Insider host Cris Sheridan discussed several websites and AI tools:
- Poe.com: an AI chatbot aggregator platform that allows users to interact with various AI chatbots and LLMs, including popular models like OpenAI's ChatGPT, Anthropic's Claude, and Google's Gemini. It serves as a centralized hub for accessing different chatbots, enabling users to ask questions, generate content, and even create their own custom chatbots without needing programming skills.
- GPT-4.0 through the ChatGPT interface
- Perplexity AI: an advanced AI-powered search engine and research tool that was launched in 2022. It utilizes natural language processing (NLP) to provide accurate and real-time answers to user queries. Unlike traditional search engines, which typically return a list of links, Perplexity generates concise summaries and includes citations from various sources, making it a more efficient tool for information discovery.
These tools can assist with tasks ranging from generating images and written content, providing advice on crafting speeches, programming assistance, language translation, travel planning, book and movie recommendations, and general research on a wide variety of topics. Dr. Thompson himself used these tools to write a 22-page paper on GPT-5's dataset in just a few hours.
"I started it at 6:00 p.m. I was pretty happy with it at 10:00 p.m. The only reason I was able to do that was using Claude 3.5 Sonnet via Poe.com and GPT-4.0 within the ChatGPT interface because it allowed me to interact with CSV (speadsheet and database) files," he explains.
Apple's Entry into AI
Apple's recent announcement of Apple Intelligence marks another significant step in the widespread adoption of AI. Dr. Thompson sees this as a move towards making AI ubiquitous, potentially bringing advanced AI capabilities to billions of iOS devices worldwide.
"ChatGPT opened access to these large language models from a few million people in 2020, maybe less, to 200 million plus per month, according to OpenAI… Apple does that even better," Dr. Thompson explains. “If you look at the count of iOS devices out there, it’s measured in the billions. So, we now have ubiquitous access [to AI] across regions.”
He notes that this democratization of AI access could have far-reaching implications, particularly in lower socioeconomic regions where access to cutting-edge technology has traditionally been limited.
Implications for the Future
As AI continues to advance at a rapid pace, it's clear that we are on the cusp of a technological revolution that could fundamentally change human society. The potential development of AGI and ASI within the next few years could lead to unprecedented advancements in science, technology, and our understanding of the world.
Dr. Thompson quotes James Campbell, who has been associated with prestigious institutions like Berkeley, Cornell, and Johns Hopkins, who recently made an astonishing statement: "Most of the staff at the secretive top labs are seriously planning their lives around the existence of digital gods in 2027." This aligns with Dr. Thompson's timeline for potential ASI development and underscores the profound changes that researchers in the field are anticipating.
However, these developments also raise important questions about the future role of human intelligence and creativity. As AI systems become more capable, how will human society adapt? What new opportunities and challenges will arise from the widespread availability of superintelligent AI systems?
Dr. Thompson sees this as an exciting development rather than a threat. "I think that's what's most exciting for me, is that we do get access to augmenting and amplifying our natural intelligence," he says.
Conclusion
The coming years promise to be a period of unprecedented technological change. As AI systems become more advanced and more deeply integrated into our daily lives, we may find ourselves on the brink of a new era in human history - one where the boundaries between human and machine intelligence become increasingly blurred. It's an exciting and potentially transformative time, but one that will require careful consideration, robust ethical frameworks, and proactive planning to ensure that the benefits of AI are realized while minimizing potential risks.
As Dr. Thompson's work demonstrates, the field of AI is moving at an incredible pace, with new breakthroughs and developments occurring almost daily. From the impressive performance of AI in academic competitions to the integration of AI with brain-machine interfaces, we are witnessing the emergence of technologies that were once the realm of science fiction only a few years ago.
The implications of these advancements are far-reaching. In the near future, we may see AI systems that can not only match but surpass human expertise in virtually every field. This could lead to revolutionary advances in science, medicine, technology, and beyond. At the same time, it raises profound questions about the nature of intelligence, consciousness, and what it means to be human.
As we look to the future, it's clear that AI will play a central role in shaping the world of tomorrow. By staying informed, engaged, and proactive, we can work to ensure that this future is one that enhances human potential and contributes to the betterment of society as a whole. The journey towards AGI and ASI is not just a technological one, but a deeply human one, challenging us to rethink our understanding of intelligence, creativity, and our place in the universe.
To listen to this full audio interview, see Dr Alan D. Thompson on AI, Brain Machine Interfaces, and Intelligence Augmentation. If you’re not already a subscriber to our FS Insider podcast, click here to subscribe.
For a link to our full podcast archive, see Financial Sense Newshour (All) and don't forget to subscribe on Apple Podcast, Spotify, or YouTube Podcasts!
To learn more about Financial Sense® Wealth Management, click here to contact us.
Advisory services offered through Financial Sense® Advisors, Inc., a registered investment adviser. Securities offered through Financial Sense® Securities, Inc., Member FINRA/SIPC. DBA Financial Sense® Wealth Management. Investing involves risk, including the loss of principle. Past performance is not indicative of future results.
