Lessons from Edwin Chen of Surge AI

Edwin Chen, the founder and CEO of Surge AI, has emerged as a significant voice in the artificial intelligence landscape. Having bootstrapped his data labeling company to a reported revenue of over $1 billion, his perspectives on data quality, AI development, and entrepreneurship offer a compelling counter-narrative to the venture-capital-fueled blitz-scaling common in Silicon Valley.

On Data Quality: The Bedrock of AI

The central thesis of Chen's philosophy is the paramount importance of high-quality data, a lesson he learned from his time at major tech companies.

On the core problem in AI: "I just used to work on ML at a bunch of these big companies and just the problem I kept running into over and over again was that it really was impossible getting the data that we needed to train our models." [1]
The founding principle of Surge AI: "We just really believed in the power of human data to advance AI, and we just had this really big focus from the start of making sure that we had the highest quality data possible." [1]
The GIGO (Garbage In, Garbage Out) principle: "AI models are only as good as the data that you feed them. If you feed your models poor data, then they'll mimic the bad data and give inaccurate predictions." [2]
The real bottleneck: "Without clean, contextual, high-quality training data, even the best models underperform.” [3]
Quality over quantity: "A thousand pieces of high quality human data, highly cured, really, really high quality human data is actually more valuable than those 10 million points [of synthetic data]." [1]
The failure of simplistic labeling: "When your labeling software is just an Excel spreadsheet, it's difficult to monitor and motivate performance." [4]
The problem with low-quality datasets: "The problem with low-quality datasets like these is that they cause your machine learning models to perform ineffectively, but they also mean your performance evaluation metrics are meaningless."
On the inadequacy of traditional crowdsourcing: "I've always felt the annotation landscape is overly simplistic... You can't ask five people to independently come up with a joke and combine it into a majority answer." [5]
The nuance of human intelligence: "High quality data actually really embraces human intelligence, creativity. And when you train the models on this like richer data, they don't just learn to follow instructions, they really learn all these deeper patterns." [1]
The danger of poor data in critical systems: "This is a severe problem when so many important products and services depend on AI—whether it's the content moderation algorithms at YouTube and Twitter, customer support systems at Uber and Amazon, or search engines at Google and Facebook." [2]
On relabeling existing datasets: "We've helped companies boost their ML model performance by 50%, simply by relabeling their existing datasets." [6]
The issue with context stripping: On Google's emotion dataset, Chen pointed out the flaw in presenting Reddit comments to labelers with no metadata, stating, "language doesn't live in a vacuum!" [7]
The need for skilled labelers: "As AI systems become more complex, we need sophisticated human labeling systems to teach them and measure their performance." [5]
Quality control is an adversarial problem: "It's a really challenging problem to detect high quality... and second if you actually take the folks from MIT who can code, they're actually just going to try to cheat you... so it's also this really, really challenging problem to detect low quality, it's actually really adversarial." [8]

On AI Development and the Future

Chen offers a pragmatic and forward-looking view on the trajectory of artificial intelligence, from the role of humans to the nature of future models.

The enduring role of humans: Chen is skeptical that AI will reach a point where human feedback is no longer needed and sees annotation becoming more difficult as models improve. [5]
On synthetic data's limitations: "Models trained on synthetic data often struggle in real-world scenarios, lacking nuance and diversity." [3]
The need for external validation: "It's almost like sometimes you need an external signal to the models like the models just think so differently from humans that you always need to make sure that they're kind of aligned with the actual objectives that you want." [1]
The diversity of frontier models: "Every company has... a different set of principles that they care about... I think that just will lead to different strengths for all the model providers." [1]
The future is multi-model: "I think today, you know, we already see like a lot of people including me, we will switch between all the different models just depending on what we're doing. And so in the future, I think that will just happen even more." [1]
On building rich RL environments: "One of the things that people really underestimate is how... complicated it is that you can't just synthetically generate it." [1]
AI Safety is a present-day issue: “The real risk isn't that AI becomes evil. It's that we train it toward the wrong objectives—and don't realize it until it's too late.” [3]
The goal of achieving AGI: "I think it really is to help achieve AGI... when you're a kid you literally dream of building AI that can do all these amazing things. And now we have the chance to do it." [8]
The industry is optimizing for the wrong objectives: Chen believes the climate has given AI companies warped priorities, citing leaderboards that incentivize being at the top at the expense of improving a model's quality. [9]
The value of specialized models: "Despite the dominance of general-purpose models, Chen sees enduring value in domain-specific approaches... 'Some products simply can't be built within the constraints of Big Tech companies.'" [3]

On Entrepreneurship and Building Surge AI

Chen's journey with Surge AI provides a masterclass in bootstrapping, efficiency, and product-led growth.

On bootstrapping: "A big part of it was obviously just that we didn't need the money. I think we were very, very lucky to be profitable from from the start." [1]
On avoiding venture capital: "You move slow. You have a lot of politics. You have a lot of bureaucracy." [10]
The power of the MVP (Minimum Viable Product): "I've always been a really big fan of MVPs. And so I literally just built myself or be one in a couple weeks." [11]
Advice to founders: “For 90–95% of startups, there's no excuse. Just build the MVP. See if anyone cares.” [3]
On staying lean and efficient: "You can build a completely different kind of company with 10% of the resources and 10% of the people, but you're still moving 10 times faster and building a 10 times better product." [8]
On big tech inefficiency: "Ninety percent of employees at tech giants are working on useless problems." [12]
Building for the right customers: "You almost want customers who share the same overall vision... your early customers will shape the kind of product that you're building." [8]
Product over hype: "I didn't want a sales team going out and selling our product like I wanted people to buy us precisely because they understood the value of high quality data." [8]
On company culture: Surge AI operates with lean teams, no standing 1:1s, and asynchronous communication to foster speed and autonomy. [3]
The problem with internal incentives at large companies: "You're simply building them to impress someone like hey I need to impress my VP, I need to impress my manager, I need to impress my director so that I can get promoted." [11]
On selling the company: "I mean I definitely wouldn't sell for 30 billion or even 100 billion. I mean if you if you think about us as a company, I already have everything I want... we're profitable, I have complete control over our destiny." [11]
Focusing on unique ideas: "I actually do believe in it... the idea of a startup as something that is a place where you can take big risks." [8]
The focus on complex problems: "We always wanted to focus on this area of more complex, agentic behavior." [9]
Surge as a technology company: "A lot of the other companies in our space they're just not technology companies at the end of the day they are either body shops or they are body shops masquerading as technology companies." [8]

On the Future of Work and the "100x Engineer"

Chen has popularized the idea that AI will not just augment, but exponentially amplify the productivity of top performers.

The emergence of the 100x engineer: “AI makes the '100x engineer' possible and disproportionately favors people who are already the '10x engineers.'" [3]
AI as a drudgery remover: "AI today as something that isn't necessarily coming up with the greatest ideas... but it often just removes a lot of the drudgery of like your day-to-day work." [8]
Unlocking ideas: "Good people have so many ideas that they just don't have time to implement. And if you think of AI... it just helps you put them to paper." [8]
The potential for single-person billion-dollar companies: "Already you have a lot of these single-person startups that are already doing $10 million in revenue... If AI is adding all this efficiency, then yeah, I can definitely see this multiplying 100x to get to this $1 billion single-person company." [13]
The importance of talent matching: “How do you match this person with the right problems that they are actually an expert on? Even your average PhD in English literature is not going to be able to write good poetry.” [9]
Credentials aren't everything: "Holding a doctorate in a certain field isn't a perfect predictor of being the best model trainer." [9]
The human-AI partnership: "In general, we think of ourselves as a “human/AI company” where humans and AI work together to improve each other." [2]

On a Personal Philosophy

Chen's approach is deeply rooted in a focus on substance and independent thinking.

On his motivation: "In many ways Surge is an embodiment of me and my interests. And what I've always loved doing is analyzing data and figuring out how to how to use that data to make models better or to make products better." [8]
Avoiding the hype: "I'm glad I'm not surrounded by the default ways of Silicon Valley thinking.” [3]
On his personal work style: "I actually have no one on one meetings... I try to avoid filling my meetings all day." [14]
The joy of insight: "I just think it's so cool that a lot of the data we're providing it's just so insightful and it like helps people build models in ways that they just wouldn't know how to otherwise." [8]
A mission-driven approach: "He is deeply passionate about AGI and quality data... a rare breed of OG founders who are technologists at heart and genuinely committed to solving extremely difficult problems." [11]

Learn more:

Lessons from Edwin Chen of Surge AI

On Data Quality: The Bedrock of AI

On AI Development and the Future

On Entrepreneurship and Building Surge AI

On the Future of Work and the "100x Engineer"

On a Personal Philosophy

Written by Antoine Buteau

Lessons from David Duffield

Lessons from Ray Peat