Overview of AI Development Puts Future in Perspective as Growth Explodes
Peter H. Diamandis, an entrepreneur, tech blogger, podcaster and named by Fortune as one of the “World’s 50 Greatest Leaders,” recently explained the explosion of AI now in his blog.
Diamandis is the founder and executive chairman of the XPRIZE Foundation, which leads the world in designing and operating large-scale incentive competitions. He is also the executive founder of Singularity University, a graduate-level Silicon Valley institution that counsels the world’s leaders on exponentially growing technologies.
Here’s part of the blog:
“AI was first discussed in 1956 at a Dartmouth Conference 70 years old. And the first deep-learning-like algorithms were described in 1965.
So Why is AI Exploding Now?
There are four reasons that are important to understand:
#1. Computational Power: While the power of computation has been doubling every 12 to 24 months for the past 50 years (Moore’s Law), it is just recently (the last 5 years) that there is enough computational power to run today’s deep learning algorithms.
#2. Massive Amount of Labeled Data: Global data has been doubling every two years, and is expected to reach 175 Zettabytes (175 billion-million-megabytes) in 2025. Not only is it a large amount of data, but it is also labeled data, enabling the training of today’s large language models (LLMs).
#3. Demonetized Training Costs: The cost to train these AI systems has plummeted 99.5% in the last five years. That’s extraordinary.
#4. Massive Investments: In 2021, corporations invested $160 billion in AI. By 2030, the global AI market is projected to be worth $1.6 trillion.
We saw ChatGPT go from 0 to 100 million users in just two months.
Proof of AI Domination:
Need more evidence of the power of AI? Below are just a few recent developments, what I call “Proof of AI Domination”:
- ChatGPT passes the Bar Exam, US Medical License Exam, and Wharton’s MBA Final Exam. For comparison, the US Medical License Exam normally requires four years of medical school and then a few more years of residency to pass. ChatGPT did it in two months.
- GPT-4, OpenAI’s newest model released last month, then beat us humans by 75% or more on a range of other exams: from the LSAT and the GRE to the SAT and the AP Math exam.
- An AI system defeats top doctors in tumor diagnosis competition in China. The AI correctly diagnosed 85% in just 18 minutes, compared to top neurologists who achieved 64% accuracy in 50 minutes. Soon, our best diagnosticians are going to be AIs.
- DeepMind’s AlphaCode AI outperforms human coders. When pitted against over 5,000 human participants, the AI outperformed 45% of expert programmers.
- An AI can clone your voice from just 3 seconds of audio. Microsoft’s AI Vall-E can synthesize your voice using only a 3-second audio sample and preserve intonation, emotion, and voice style.
- The recently released semi-autonomous AI Auto-GPT removes the need for writing prompts altogether. Auto-GPT, which is based on GPT-4, works through goal-setting, with the AI generating and completing tasks once a given goal has been specified—from creating and completing to-do lists to ordering food and booking flights.
- AI designs a potential cancer drug in 30 days. Researchers used AlphaFold and the drug discovery platform Pharma.AI to design and synthesize a potential drug to treat hepatocellular carcinoma (HCC), the most common type of primary liver cancer.
And so much more (every day)… it’s breathtaking!
The CEO of Stability AI Emad Mostaque, who I’ve had on my podcast and on the Abundance360 stage, recently made an extraordinary point
“Google and Microsoft are going all-in with generative AI as core to their future. There is no ‘we are still early’ here, trillion-dollar companies are shifting their whole strategy and focus.”
And we’re seeing this play out…
read more at diamandis.com