From 14 People to 500 Million Users: Sam Altman Reflects on OpenAI Journey

AI Summary: Sam Altman at AI Ascent reflected on the company's journey from a small research lab founded in 2016 focused on fundamental AI challenges to a powerhouse valued at $340 billion with 500 million weekly active users. He discussed the company's current model of maintaining momentum with small, focused teams, the generational differences in how people use AI, and outlined a future vision where OpenAI becomes a "core AI subscription" aiming for models with massive context to integrate seamlessly into users' lives.

May 13 2025 08:17
At AI Ascent, OpenAI CEO Sam Altman offered a rare glimpse into the company's meteoric rise from a tiny research lab to an AI powerhouse with a reported $340 billion valuation. Reflecting on a journey that began with just 14 people trying to teach AI to play video games and has evolved into a platform with over 500 million weekly active users, Altman shared insights on OpenAI's evolution, its future direction, and how AI is fundamentally changing how we interact with technology.



OpenAI's modest beginnings in 2016 bear little resemblance to the company we know today. "We were sitting over there and there were, you know, 14 of us or something and you were hacking on this new system," Altman recalled, describing the company's first office. "It's almost impossible to sort of overstate how much we were like a research lab with a very strong belief and direction and conviction but no real kind of action plan."

In those early days, the idea of building consumer products seemed far-fetched. The team was focused on fundamental AI research challenges like getting systems to play video games and manipulate robot hands. Large language models (LLMs), which would later become OpenAI's breakthrough technology, weren't even on the immediate horizon.

"Not only was the idea of a company or a product sort of unimaginable, LLMs as an idea were still very far off," said Altman.



The road to creating ChatGPT wasn't straightforward. Contrary to popular belief, ChatGPT wasn't even OpenAI's first consumer product – that distinction belongs to DALL-E, their text-to-image model. Before that, OpenAI had focused on creating an API that other companies could build upon.

The development of GPT-3 marked a significant turning point. While the model showed promise, the team struggled to identify its practical applications. "By the time of GPT-3, we both thought we had something that was kind of cool, but we couldn't figure out what to do with it," Altman explained.

The financial pressures of scaling AI models also began to mount. "We were heading into the world of billion-dollar models. It's hard to do those as a pure science experiment unless you're like a particle accelerator or something," said Altman.

This financial reality pushed OpenAI to think about sustainability. In June 2020, they released GPT-3 as an API, hoping that others might find practical applications for the technology. While the broader world showed little interest initially, Silicon Valley took notice. Some startup founders even declared it as AGI (Artificial General Intelligence), though Altman didn't go that far.

The real breakthrough insight came from user behavior: "Even though people couldn't build a lot of great businesses with the GPT-3 API, people love to talk to it in the playground," Altman observed. Despite not being optimized for conversation, users were drawn to interacting with the model directly. This observation eventually led to the development of ChatGPT.

By the time ChatGPT launched on November 30, 2022 – about six years after OpenAI's founding – the company had created a product that would fundamentally change how people interact with AI.



Today, OpenAI operates at a dramatically different pace and scale. When asked about the company's recent product velocity, Altman emphasized his belief in keeping teams small and focused while expanding the scope of projects.

"A mistake that a lot of companies make is they get big and they don't do any more things. So they just get bigger because you're supposed to get bigger and they still ship the same amount of product. And that's when the molasses really takes hold," Altman explained.

Instead, Altman advocates for giving small teams significant responsibility across many projects:

I am a big believer that you want everyone to be busy. You want teams to be small. You want to do a lot of things relative to the number of people you have, otherwise you just have like 40 people in every meeting and huge fights over who gets what tiny part of the product.

This approach has allowed OpenAI to maintain momentum even as it grows. The company continues to release new capabilities and models, with Altman expressing particular pride in the quality of their current models: "The models are so good now... I'm amazed that one model can do so many things so well."


One of the most fascinating insights Altman shared was how different generations use ChatGPT in dramatically different ways. This generational divide reveals much about how AI might evolve to become more central to our daily lives:

• Older users tend to use ChatGPT primarily as a Google replacement – a sophisticated search engine
• People in their 20s and 30s use it as a "life advisor"
• College students and younger users treat it like an operating system

"They really do use it like an operating system," Altman noted about younger users. "They have complex ways to set it up to connect it to a bunch of files, and they have fairly complex prompts memorized in their head."

Even more striking is how integrated the tool has become in their decision-making: "They don't really make life decisions without asking ChatGPT what they should do. It has the full context on every person in their life and what they've talked about."

This behavioral difference points to a future where AI becomes increasingly embedded in our daily routines and decision-making processes.



Looking ahead, Altman outlined a vision where OpenAI evolves from providing an AI assistant to becoming a "core AI subscription" that integrates across all aspects of a user's digital life.

"We want to be people's core AI subscription and way to use that thing," Altman explained. While he acknowledges they haven't yet perfected the platform aspects – "We have not yet figured out exactly what the API or SDK or whatever you want to call it is to really be our platform" – he's confident they will work it out.

The ultimate goal appears to be creating an AI that can move seamlessly between services and devices while maintaining a continuous understanding of a user's life and preferences. Altman envisions: "You should be able to sign in with OpenAI to other services. Other services should have an incredible SDK to take over the ChatGPT UI at some point."

This vision extends to creating what Altman described as:

a new protocol on the level of HTTP for the future of the internet where things get federated and broken down into much smaller components and agents are constantly exposing and using different tools and authentication, payment, data transfer, it's all built in at this level that everybody trusts.

Perhaps most ambitious is what Altman described as the "platonic ideal state" for AI: "A very tiny reasoning model with a trillion tokens of context that you put your whole life into."

In this vision, the AI model would never need to retrain or customize its weights. Instead, it would continuously append new information as you live your life – every conversation, every book read, every email – while being able to reason efficiently across all that context.

"Your life just keeps appending to the context, and your company just does the same thing for all your company's data," Altman explained. While acknowledging this isn't achievable with current technology, he views it as the north star that guides OpenAI's development efforts.


Altman also offered glimpses into OpenAI's internal workings, revealing that the company's employees are already heavily using their own technology. When asked how OpenAI uses ChatGPT internally, Altman admitted, "It writes a lot of our code."

While hesitant to quantify exactly how much, he emphasized that it's writing "meaningful code" – "the parts that actually matter."

On research priorities, Altman indicated that OpenAI balances between giving talented researchers freedom to explore and providing coordination when necessary. "There are some projects that require so much coordination that there has to be a little bit of top-down quarterbacking, but I think most people try to do way too much of that," he explained.

This approach to innovation seems informed by historical models. Altman mentioned that when starting OpenAI, they spent significant time studying successful research labs of the past, adopting principles that had worked before. "We shamelessly copied from other good research labs in history," he said, expressing surprise that more organizations don't follow these proven approaches.


Looking to the near future, Altman sketched what he sees as the likely progression of AI capabilities:

• 2025: "A year of sort of agents doing work," with coding likely to be a dominant application
• 2026: AI assisting in major scientific discoveries
• 2027: AI moving from the intellectual realm to the physical world, with robots becoming "a serious economic creator of value"

While emphasizing these were "off the top of my head kind of guess," Altman's timeline suggests a rapid acceleration of AI's practical impact over the next few years.


In a more personal reflection, Altman touched on the challenges of leadership, particularly in the aftermath of his brief ouster from OpenAI in late 2023 (which he referenced indirectly).

"The hardest thing about the big challenges that come as a founder is not the moment when they happen," he observed. The real test, he suggested, comes later: "The thing that I think is harder to sort of manage your own psychology through is the fallout after."

Altman noted that while there's much discussion about handling the immediate crisis, there's less guidance on "how you pick up the pieces" in the weeks and months that follow – "on day 60 as you're just trying to rebuild after it."

This perspective on resilience seems particularly relevant given OpenAI's tumultuous journey and Altman's own experience with organizational upheaval.


As OpenAI reportedly seeks to raise $40 billion at a staggering $340 billion valuation, Altman maintains that there is "no master plan" beyond making great models and shipping good products. He expressed skepticism about overly deterministic planning: "I have heard some people talk about these brilliant strategies of how they're going to work backwards... I have never seen those people really massively succeed."

Instead, Altman emphasized adaptability: "We pride ourselves on being nimble and adjusting tactics as the world adjusts." This approach seems to have served OpenAI well, as the company has repeatedly pivoted based on user behavior and technological breakthroughs.

With 500 million weekly active users and a valuation that would place it among the world's most valuable companies, OpenAI's journey from a small research lab to a global AI platform represents one of the most remarkable growth stories in tech history. The vision Altman describes – of AI that intimately knows and assists us throughout our lives – suggests the company's most significant impacts may still lie ahead.