SVP of Product at Vercel | The Velocity Advantage: How AI-Native Teams Ship Fast and Safe

Well, hello San Francisco. We are so fortunate  to be here at the epicenter of AI where all these companies around us are shipping new models,  new agents and new autonomous systems at a breathtaking pace. Versel provides a developer  experience and an AI cloud that's actually

powering many of these ships. So today we're going  to take a look at how to ship at the speed of AI. Building in AI is like building in an earthquake.  I guess that's appropriate for San Francisco. But the ground is always moving. Um what that means  is, you know, the best model changes at any time,

sometimes multiple times a day. What this means  for us is that your release infrastructure has to be flexible so that you can try out new technology  and see what works best for your product. You also want to be really really fast and that means under  the hood that you need a fast CI/CD pipeline. You need to be able to feature flag features to  certain users and you need to be instant be able to instantly roll back when things don't  work. That can often be the difference between

a product that stays at the forefront of  the user's mind and one that falls behind. Oftentimes, it's actually security that holds  AI back, particularly in large enterprises. When you think about AI, it's actually very very  expensive and that attracts a lot of malicious actors and there's a growth of malicious actors  on the web. bad bots, you know, security exploits that are just waiting to use these AI expensive  workloads. So to move fast, applications also need

to be secure by default and against the latest  threats that are always dynamically changing. Now, most engineering teams are not equipped for this  velocity. their CI/CD and their workflow is not set up to iterate rapidly or to test globally  with select users and in many cases even at the most cutting edge companies developers  are setting up this infrastructure DIY

um that means there's a steep learning curve and  sometimes talent shortages in the developer ranks meanwhile users are eager to move to the latest  technology and we've never seen hunger like this for AI. Business-wise, not having the right tools  means that bugs are not easily detected, security is bolted on, and ultimately you either ship a  sub-optimal user experience or you ship a product

late. It doesn't have to be this way. And part of  the reason why developers like Versell is because it eliminates that manual toil. Versel understands  your application. It recognizes the framework automatically and it configures the infrastructure  automatically as well. And so when builds are fast and rollbacks are instant, it means that every  developer can ship small iterative changes. And

that means that your team can move forward. You  can deploy on Friday, you can deploy on Saturday when that new model comes out. Also, security is  baked in by default and so your deployments can be preview deployments that you share only with  select users. Uh they can be obviously password protected and when you do roll out to production,  your application is protected by a firewall that