OpenAI co-founder and CEO Sam Altman has marked the company's 10th anniversary with a deeply personal reflection, celebrating the 'crazy, unlikely, and unprecedented' journey from a band of optimistic nerds to a global force in artificial intelligence. In a blog post, Altman expressed profound gratitude and optimism, underscoring the transformative potential of AI while acknowledging the immense challenges ahead.
Announced to the world a decade ago and officially launched in early January 2016, OpenAI has evolved from a small, misunderstood collective into an organisation at the forefront of the AI revolution. Altman described the early years as a time of 'unreasonably optimistic' energy, driven by a conviction that the mission – to ensure artificial general intelligence (AGI) benefits all of humanity – was worth the long odds.
OpenAI's humble beginnings: 15 nerds and a sharp focus
Reflecting on grainy photos from the startup's infancy, Altman painted a vivid picture of youthful exuberance amid uncertainty. "When I look back at the photos from the early days, I am first struck by how young everyone looks. But then I’m struck by how unreasonably optimistic everyone looks, and how happy," he wrote.
The initial team of 15 fostered a culture of relentless discovery, tackling obstacles from securing funding for massive computing power to pioneering practical AI safety measures. Altman highlighted the 'great spirit of always just figuring out the next obstacle,' which propelled breakthroughs in deep learning and real-world applications.
OpenAI scaling in 2017
By 2017, OpenAI notched several landmark achievements that laid the groundwork for future success. Altman spotlighted the Dota 1v1 results, which elevated reinforcement learning to unprecedented scales; the unsupervised sentiment neuron, demonstrating a language model's grasp of semantics over mere syntax; and early work on reinforcement learning from human preferences, charting a path to align AI with human values.
"At this point, the innovation was far from done, but we knew we needed to scale up each of these results with massive computational power," Altman noted, crediting the team's ability to learn from both victories and setbacks.
ChatGPT explosion: From lab to global phenomenon
The past three years have been a whirlwind of intensity, Altman admitted, with the launch of ChatGPT in late 2022 catapulting OpenAI into the mainstream. Followed swiftly by GPT-4, these releases shattered perceptions of AGI as mere science fiction. "All of a sudden, AGI was no longer a crazy thing to consider," he observed.
This rapid ascent demanded "extremely difficult execution," transforming OpenAI from a research outfit into a massive enterprise overnight. Altman praised the team's decision-making under pressure, though he shouldered blame for missteps: "The ones we’ve gotten wrong are mostly my fault."
Altman's optimism for superintelligence
Today, OpenAI's models outperform most humans in intellectual benchmarks, and Altman is more bullish than ever. "I have never felt more optimistic about our research and product roadmaps, and overall line of sight towards our mission," he declared.
Looking a decade ahead, he predicted near-certainty in achieving superintelligence by 2035. While daily life may evolve subtly – with humans fixated more on each other than machines – the era's capabilities will defy current imagination. "The people of 2035 will be capable of doing things that I just don’t think we can easily imagine right now."
