AWS for Software Companies, Interview, Suger Inc | Amazon Web Services

Key Takeaways

I'm John Yu, a co-founder and CEO of sugar.io. We work with many AI startups um who are trying to commercialize their AI products after years of you know R&D and we help them access millions of users that are you know live on AWS. I think there's a lot of misconception about generative AI and and rightfully so because it's such a new industry or new kind of paradigm shift and the biggest misconception that you see on LinkedIn or social or you know on on speeches is the concept of the death of software or SAS. What generative AI is is really these large language models that people can access via an API call. It's only going to supercharge the SAS industry as companies are now leveraging it its capabilities from a you know what I call hosted in you know the sellers cloud to now host it in the buyers environment where these buyers can now leverage all this first party proprietary data that exists uh internally. One of the most interesting things about AI is the concept of uh like pricing. you know there's an underlying variable cost which is the number of tokens that you have to access you know via these large language models and so there's a natural price floor there's a lot of companies that are now leveraging AI for customer support tickets but even if there's a discount curve based on the number of tickets resolved there's still the underlying cost there's a cap or there's a floor at which they can price and so I think cost is one of the biggest you know factors that people are considering when they're leveraging AI the other side of it of course is data security I talked to so many you know CTOs or companies that are very concerned about how their you know internal employees are actually leveraging AI in a way that is incredibly secure and that you know their own internal data does not leave their environment uh which really exposes the or increases the amount of threat vectors that exist. AWS helps my software company uh sugar.io be a lot more agile and efficient in how we you know operate our own internal processes.