AI is everywhere, and as time passes, the technology is going to so smart and powerful, that it will be inseparable from human lives.
AI will be so ubiquitous that no future computers shall operate without it.
The the roads toward that is continuously evolving. OpenAI pioneered the commercialization of Large Language Models with ChatGPT, and that from the very beginning, the goal is AGI.
At first, pursuing that idea seemed far-fetched. At the time, AI was still a hot topic but far from a guaranteed success.
But Sam Altman, CEO of OpenAI, believes that the next-generation of AIs are just around the corner.
In an interview with Y Combinator, the company he was a former president, he said that:
"At a time when in the field you weren't allowed to say that because that just seemed impossibly crazy."
" [...] we were far less resourced than [Google] DeepMind and others so we said okay they're going to try a lot of things and we've just got to pick one and really concentrate and that's how we can win here."
"Most of the world still does not understand the value of like a fairly extreme level of conviction on one bet.
That's why I'm so excited for startups right now it is because the world is still sleeping on all this to such an astonishing degree."
In the interview with the president and CEO of Y Combinator Garry Tan, which happens 8 years after the original How to Build the Future, a lot of things have changed.
When at first, AGI was more a speculation than a potential use case, this time, Altman confirmed that AGI is really near.
But here's the thing: while an increasing number of people began to realize about AI's existence and began to make use of the technology to their advantage, most people are undermining the technology's massive potential.
AGI is a form of AI that is as capable as, if not more capable than, all humans across almost all areas of intelligence. It has been the ‘holy grail’ for every major AI researchers, and many predicted it would be a decade or more before it can be reached.
As 2024 is coming to an end, Altman expects to see AGI in 2025.
OpenAI defines AGI as "AI systems that are generally smarter than humans," and explains that there are five levels of AI:
- Level 1 is where chatbots reside. It's where AI has natural conversation language abilities.
- Level 2 is reasoners, or when AI reaches human-levels of problem solving across a broad range of topics.
- Level 3 is agents, which can be described as AI systems that can take actions independently or from human instruction.
- Level 4 represents innovators, or AI that can aid in the invention of new ideas and contribute to human knowledge.
- Level 5 is AGI, or the time when AI is capable of doing all of the work of an organization independently.
Whereas Level is is where ChatGPT was, Level 2 is what the world is seeing through models like OpenAI o1. Then, and happening at about the same time, is Level 3 where AI is able to perform tasks on their own, like through the leaked Google Jarvis and Anthropic Claude 3.5 Sonnet.
The final two levels are a big step up.
Being surrounded by the right people, Altman focused the company towards scaling deep learning and exploring how AI could evolve.
He tried tackling challenges, address criticisms, and deal with skepticisms.
By scaling up, Altman and his team managed to yield better results, and saw an "emergent behavior" in machine learning—where the learning process itself seemed to evolve in unexpected and beneficial ways.
This was a key insight that fueled their research.
Read: 'AGI Is The Equivalent Of A Median Human', Just Like A Remote Co-Worker
Not everyone agrees about AGI through, and the definition of the term is still very much undecided. AGI is still a controversial topic in the AI community, and that nobody can really agree to its clear definition.
In some definitions AGI also has to be able to learn, adapt, and perform tasks similar to human intelligence, going beyond just "knowledge". This would require it to create output not based on human input, moving beyond its training data.
In reality, AGI will be a gradual thing.
The transformation isn't going to happen and changes everything in an instant.
Much like the rise of generative AI, in which the technology should be improving step-by-step until it becomes seamlessly integrated into all aspects of life.
Altman also recognized that unforeseen challenges might arise along the way.
Nevertheless, he remains optimistic about AI’s potential to expand human capabilities profoundly.
Further reading: 'I Don't Care If We Burn $50 Billion A Year,' Because AGI Is Going To Be 'Totally Worth It'