A Conversation with Sam Altman on The Possibilities of AI

May 2, 2024 • 4 min read

In a recent conversation with Ravi Belani at Stanford University's Entrepreneurial Thought Leaders (ETL) series, Sam Altman, CEO and co-founder of OpenAI, shared his insights on the rapid advancements in AI, the opportunities and challenges it presents, and his advice for aspiring AI entrepreneurs. He also took the time to field a wide range of questions from students. It's a great conversation that I highly encourage everyone to watch in its entirety.

In this article, I highlight my top takeaways and summarize his most important points:

Altman emphasized the extraordinary pace of AI development, stating, "We can say right now with a high degree of scientific certainty, GPT 5 is gonna be smarter than, a lot smarter than GPT 4, GPT 6 can be a lot smarter than GPT 5 and we are not near the top of this curve." He stressed that AI is not just going to improve in specific areas or modalities, but rather, it will become smarter in a general sense. The gravity of this statement, Altman believes, is still underrated.

💡

Altman's assertion that AI will continue to get "smarter" at a rapid pace is significant. His perspective rebuffs concerns about LLM innovation potentially hitting a plateau, especially with the recent drop-off in performance that we have seen with GPT-4. As each new iteration of AI models surpasses the last, the potential applications and impacts across industries will be transformative. This underscores the imperative for businesses and individuals to stay informed and adapt quickly to the evolving landscape.

In a remarkably candid moment, Altman admitted, "ChatGPT is like mildly embarrassing at best. GPT4 is the dumbest model any of you will ever, ever have to use again by a lot." However, he emphasized the importance of iterative deployment and tight feedback loops in the development of AI systems. "It's important to ship early and often and we believe in iterative deployment," he said.

💡

I loved this. Altman's candor about the current state of AI is both refreshing and strategic. By acknowledging the limitations of GPT-4, he provides a much needed reality check to the overhyped discourse that often surrounds today's language models. At the same time, his statement serves as a bold challenge to OpenAI's competitors (Google, Anthropic, Meta—I'm talking about you) who have yet to truly surpass GPT-4's capabilities even a year after its release. OpenAI is likely still much much further ahead than the rest of the industry. Think about it, Altman is deliberately setting high expectations for OpenAI's next release. That level of confidence should pique curiosities and raise a lot of eyebrows among its competitors.

Altman warned against building AGI in isolation, without involving society. "If we go build AGI in a basement and then, you know, the world is like kind of blissfully walking blindfolded along. I don't think that makes us like very good neighbors," he explained. Instead, he believes that putting AI products in people's hands and allowing society to co-evolve with the technology is crucial. "People don't wanna be surprised, people want a gradual rollout and the ability to influence these systems." He further stressed "Let society tell us what it collectively and people individually want from the technology. How to productize this in a way that's gonna be useful. Where the model works really well, where it doesn't work really well. Give our leaders and institutions time to react, give people time to figure out how to integrate this into their lives..."

💡

A gradual rollout and encouraging society to influence the development of AI technology feels like the right approach. Not only does this foster trust and acceptance of this disruptive technology (and this will be disruptive), but it gives us an opportunity to think through and have public discourse on complex issues around governance, intellectual property, safety, access and more. This is precisely why AI literacy is so crucial. An informed society will ultimately ask better questions, demand more accountable systems, and shape AI development in a way that aligns with public values and ethical standards. This sort of engagement is essential to ensure that AI is broadly beneficial and does not worsen existing inequalities.

When asked about the significant expenses involved in AI development, Altman had a surprising response. "Whether we burn $500 million a year or $5 billion or $50 billion a year, I don't care. I genuinely don't. As long as we can stay on a trajectory where we eventually create way more value for society than that, and as long as we can figure out a way to pay the bills, we're making AGI. It's gonna be expensive, but it's totally worth it," he stated.

💡

I hope so. I really do. Like Altman, I too have "a general bias to be too pro-technology" and I genuinely hope that the investments made in AI will ultimately benefit society as a whole. The numbers are indeed staggering, and I do feel a sense of unease when considering the stark contrast between these investments and the persistent poverty and inequality that still plague our world today. This all hinges on AI "creating way more value for society" in the long term. This is our challenge. We ~~can~~ must do this.

While AI has the potential to revolutionize society, Altman offered a nuanced perspective on its impact. He predicted that AGI would significantly enhance human capabilities, but the world may not seem drastically different day-to-day. "In some sense, it will be super different, and in some sense, not different at all," he said, emphasizing that while AGI will enable us to accomplish more complex tasks, it won't change the fundamental nature of human experiences and interactions.

Altman also addressed the potential dangers of AI—both subtle and cataclysmic. While much attention is given to existential risks, he expressed more concern about the pernicious effects that may go unnoticed, drawing a parallel to the widespread attention deficit issues arguably caused by social media platforms like TikTok.

💡

I couldn't agree more. Unfortunately, too often, the narrative in the mainstream media around AI is focused on exploiting the fears of a misinformed public—fears that are founded more in science fiction than in reality. It is critical that we continue to advance research in AI safety and responsible AI development.

For those eager to dive into the world of AI entrepreneurship, Altman's advice is clear: trust your intuition, be prepared to adapt, and focus on being right rather than just contrarian. He encouraged aspiring entrepreneurs to chart their own course and pursue ideas that may not be obvious to others.

"I think this is probably the best time to start a company since the internet at least. And maybe kind of like in the history of technology. I think what you can do with AI is gonna just get more remarkable every year."

Chris McKay is the founder and chief editor of Maginative. His thought leadership in AI literacy and strategic AI adoption has been recognized by top academic institutions, media, and global brands.