Interview, Fireside Chat
Surge CEO & Co-Founder, Edwin Chen: Scaling to $1BN+ in Revenue with NO Funding
- The company, profitable with $1 billion in revenue and started in 2020, is viewed as having a valuation potential of $30 billion to $100 billion; however, the speaker rejects offers around $40 billion to $100 billion to maintain full control for the primary goal of achieving Artificial General Intelligence (AGI), with a prediction of reaching a $1 billion valuation within approximately two years.
- Economic forecasts anticipate a 10x increase in global GDP and productivity within the next 10 years, generating $10 trillion in value, driven by AI replacing "50 of the things" currently done on Google and automating tasks currently performed by average L3 or L4 software engineers by 2028.
- Specific timelines predict that the average engineer role will be automated by 2028, whereas complex achievements like curing cancer, designing new philosophical systems, and sending rockets to Mars are expected by 2038.
- The market is expected to evolve to include three additional frontier AI companies alongside existing players, featuring distinct personalities, focuses, and trade-offs similar to how current models differ in coding versus consumer optimization.
- Future development will prioritize "10x improvements" over incremental gains, utilizing AI to amplify "10x engineers" and generate serendipitous creative breakthroughs, though current models are noted as solving narrow "homework problems" rather than deep, meaningful issues.
- Significant risks include the potential for powerful models to accidentally maximize toward wrong objectives, leading to unknown outcomes in critical sectors like insurance or trillion-dollar companies, which poses a threat greater than current benchmark hacking.
- Reliance on synthetic data is expected to cause model collapse in real-world use cases due to narrow similarity scopes, necessitating the eventual replacement of synthetic data with high-quality human data after months of filtering.
- Current capabilities are limited to handling tasks involving writing little features or narrow academic problems, whereas meaningful code generation (50% of meaningful code) and idea generation for deep problems remain beyond current reach.
- Future progress will be measured by the variety of projects created, with an emphasis on building a company that addresses massive problems requiring significant headroom beyond current achievements.