Transcript
Hey, how's it going? Hope you're having a great day. Um, I want to talk very quickly about this list. This is the AI ship releases and I very specifically want to go through everything as a whistle stop tour and just tell you a little bit about it because I think it's important.
Um, AI SDK number one. Uh, the AI SDK is basically an easy way to be able to talk with AI models through using TypeScript. That's that's basically it. Okay. Um, now there is a beta that's just come out and this does a whole bunch more to be able to stop things from like toolings from running on forever. But the main thing here is that you should have a go with just playing with it. It's really, really cool. It's really powerful.
Number two is the Vercel marketplace. And the Vercel marketplace, you might be thinking, hold up, this is this has been around for a while. Um, that is correct. So, there used to be the ability, well, I say used to be, there still is the ability to be able to just install a database, something like Neon or Upstash or any of these tools, um, by pressing a button. Well, one of the things that's kind of interesting, it's a bit of a paradigm shift, is we're seeing more and more kind of very specific niche AI models. Um, and as you can see, one of the big things here is that Vercel is thinking about the same process of where you just press one button and you immediately get that AI model or that AI tooling or potentially any number of things that might work with the AI SDK by just pressing a button. It means you spend less time on infra and more time actually building the thing you want to build.
Um, number three, and this is the sleeper hit, this is our favorite inhouse is the whole idea of workflow. You're going to hear the word durability thrown about a whole bunch. Um, what durability basically is is it's a way to be able to do something over a long period of time and it reliably work. Um, it doesn't have to necessarily be a long period of time, but even with things like flaky models or failovers, sorry. Um, that's the kind of thing that workflow does. So, it sits on one side and it basically goes step one, has it been completed? Cool. Step one completed. Step two, has it been completed? No, it failed. Okay, let me retry it. Step three, well, we don't do step three until a week's time. So, wait a week, then do step three. That's what workflows does.
Now, if you're a developer, this is a nightmare. It takes a long, long time, and you have to do things like cron jobs, and you have to hope to god that things don't break. Um, but this is now so much simpler and amazing for teams like to be able to do some incredible stuff. Um, maybe even just being kickstarted from a form.
Number four, Vercel agent. So this one's interesting and this one's actually solved a couple of bugs inhouse already from us. So Vercel agent basically sits on top of the observability layer and rather than you having to manually go through and go oh you know what happened at this 500 error how many people did it happen to what was the cause where did that come from is it a vendor yada yada yada Vercel agent basically tells you and it's pretty accurate we've we've seen it kick off about six or seven times and every single time it got the answer the cause of what caused the issue. Um, so really really cool tool. Just enable it. It's it's really cheap for what it does. If you compare this to certain people, data dog, um, it's it's really good.
So number five is Vercel for Python. So for the three developers out there that still use Python, really really good tool. So feel free to use that.
Number six, um, an agent on every desk. Okay, this is this one. This one this is a thinly veiled marketing ploy. So definitely click that one. Definitely click on the uh the contact us form. No, I think it's a really good thing. So, basically, Vercel is trying to encourage people to build agents inhouse and I can see this being a very very defining thing between kind of mediocre agencies or mediocre uh companies where they don't really touch AI. They don't use it in any internal workflows and they fall behind. Um and so, you know, they're telling you this and then they're giving you these on seven and eight.
So, seven and eight. Seven is a lead agent. So, it's a way of being able to take things like forms and then enrich them with AI doing R&D in the background. Um, and there was a great example saying it takes about 15 minutes to do this each time. Can you imagine how much time you're going to save where if an AI goes off and does all of this and it's already there, you can just literally run it with one button. Um, this does it where it can integrate into your forms and research that information.
And then finally is the uh data analyst agent. And this one is also very cool because you can throw this in Slack and you can actually start using it like oh uh out of interest how many people came to the website last month? How many of those people converted? What did they convert from? Which pages did they go to? You know all these kind of questions. Um and it will create the SQL to read your data warehouses and things like that to get that information.
Um I hope this all helped. I'd love to hear what you're building at the moment. Get in touch. Thanks a lot. Take care. Bye-bye.
