SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

When we started OpenAI, I was convinced that deep learning was genuinely working and that it improved with scale. This realization was crucial because it seemed like others weren't recognizing its potential.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

Brad's adaptability is a key strength; he transitioned from finance to managing a fast-growing business, demonstrating a willingness to take on new challenges and figure things out as they arise.

BL

Brad Lightcap

04/15/24

@ 20VC with Harry Stebbings

The ability to focus on one to three critical priorities at any given time is essential for maintaining velocity at scale. This focus helps align the team and drive progress effectively.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

The biggest challenges for OpenAI include maintaining a strong research culture and ensuring sufficient compute resources to meet the growing demand for AI capabilities.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

The expectation versus reality of AI adoption is often skewed; while expectations are high, the actual capabilities of current models may not meet those expectations, leading to a rapid adjustment in perception as models improve.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

The commoditization of AI models is inevitable, but the enduring value will come from a small number of providers who can deliver highly personalized and context-aware models.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

The long-term differentiation in AI will not be in the base models but in how personalized and integrated they are into users' lives, providing context and utility that meets individual needs.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

There are two strategies to build on AI: one assumes the model won't improve, while the other assumes it will continue to get better. Most startups are currently built on the former assumption, which could lead to them being steamrolled by OpenAI's advancements.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

To avoid being steamrolled by OpenAI, companies should ask themselves if they are excited about a 100x improvement in AI models. Those that are should have a clear path to how better intelligence will enhance their products.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

Iterative deployment is vital; we want to avoid developing AGI in secret and instead allow society to engage with our models as they are released, which helps in understanding and shaping their use.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

I believe that scientific progress is the highest order bit of progress for society, economic growth, and quality of life, and if AI can help increase the rate of scientific progress, it will be a triumph.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

The biggest barrier to increasing scientific progress with AI is that the models are just not smart enough yet.

BL

Brad Lightcap

04/15/24

@ 20VC with Harry Stebbings

I think enterprises often underestimate the importance of giving people access to technology, which can lead to significant improvements in productivity that aren't easily quantifiable.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

I believe that the best thing we can do to sell more product is to make the product better, and the best way to improve the product is through better research.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

I believe that the best ideas often come from unexpected places within the team, not just from the most experienced members.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

I try to keep a team where great ideas can come from unexpected places, not just from the most experienced members.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

I don't think about hiring for more or less experience; it's about finding the right person for the role, regardless of their background.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

I think the biggest challenge for OpenAI over the next 12 months is doing the best research and productization of innovation.

BL

Brad Lightcap

04/15/24

@ 20VC with Harry Stebbings

I think the rate of adoption in the enterprise will be faster than people realize, and we will break the convention that enterprises are slow adopters of technology.

BL

Brad Lightcap

04/15/24

@ 20VC with Harry Stebbings

The scaling of models has worked consistently, and it's incredible that making these models bigger leads to predictable improvements in their performance.

BL

Brad Lightcap

04/15/24

@ 20VC with Harry Stebbings

I wish I had appreciated the order in which the technology would have an impact; it surprised us how important it is in creative industries compared to more traditional sectors.

SA

Sam Altman

04/15/24

@ 20VC with Harry Stebbings

I think it's important to have a supportive partner who understands the demands of this job and is flexible about it.

Sam Altman & Brad Lightcap: Which Companies Will Be Steamrolled by OpenAI? | E1140

Takeaways