ChatGPT podcast mentions

Quote

Chat interface that was a model. Yeah, yeah, yeah. Yeah. And then it became you know, more sophisticated, you could upload files and PDFs. And so for my personal use case I use LMs mostly uh for like actually it sounds weird, but like proof So just the um before table of contents and then I just uploaded the PDF to the ChatGPT interface and say Can you give me the headers? Uh so I don't have to pull that um out myself. Making work a bit simpler like these um tedious things. Um but then Like you said, there was also the new opost model and then um ChatGPT released um Um codex five point three and uh mac OS app with that. So like yet another leap in terms of what um these models are capable of. I mean before there were also coding L LMs and it became more popular to use LMs for coding. But it's always you know more and more and getting better and better. And so uh before I used um Visual Studio Code, uh I mean because I well I uh Uh just used Visual Studio Code, the code editor, for like years, like maybe five

Quote

So just to kind of kick things off, I just want to talk about the landscape of Agentec workflow. There's I think, you know, at first um glance and we've talked about different automations. automation have kind of gone from being completely separate things to now most of the AI tools existing out there are having, you know, we saw scheduled tests originally. With things like ChatGPT, but they become much more usable, right? It's not just like just running a prompt at a certain time every day. It's running a prompt that actually has connection to different tools. in the same way that an automation would naturally work and N A D N make.com or otherwise, Zappy or otherwise. Um, and what we're going to show you today is how JAT form has kind of brought to you a really nice agentic workflow um and how other tools have kind of built upon where I mean you can do this kind of stuff uh with products like you know cloud code, codecs. and um different tools. Um but I'm you know try Trying to f like some tools are people are actually scared.

Quote

being announced to folks kind of adopting it. But yeah, I think I think effectively uh okay, I'll say this. If you're a large model lab and you want uh you know to build uh you know sandbox environments inside of your that people are not using their model as they would want to. They built ChatGPT, but I would say that Chat GPT now defines open I know they're doing a lot business from their their APIs. But still is this like how you think like isn't bold. um uh new is is your Your business now? Like why why don't you like put focus on that instead of the advice? Well well you're you're right. And so so well going into it we candidly. We were like, Okay, but yeah, Bolt I knew this thing is super cool. I'm like, we think people are stoked. Well but we think people will be stoked. But we're like maybe this will add, you know, uh best case scenario after month one, we'd be mind blowing. on if we added a couple hundred K of er R or something, you know? And and we were like, but we think there's probably gonna be a you know, an immediate huge business.'Cause we there was some early poll on uh you know folks wanted to put web container into their product offerings, you know, kinda similar to what Bolt is do doing or whatever. We were actually prepared for the the inverse outcome here. But uh well I guess

Quote

And so application volumes went up again. So you had these predecessors. of just like economic and job market changes, they were already making it. more, you know, less of a candidate market and more of a hiring market uh to start with. And then of course 2022, Chat GPT comes out and everything accelerates. Since then basically we're all using AI more and more in all aspects of our life. this week I I pointed ChatGPT at my fridge and asked it what I should make for dinner. And so it's a it's it's impacting everyone at home, everyone at work, and it's impacting how people look for jobs. And so on the back of that, what's what's happened is job seekers. feeling the softening job market and sensing that they need to up their game and apply. more jobs are using AI to do just that. And so they're really doing two things. One is is they're automating the application process. It's going from a a a world where People used to think about which jobs to apply for and then put some effort to applying into them to having AI just kind of like find as many jobs as you can, apply me to all of them. And then secondly, if is AI can also automatically customize every single job application. according to that post. So it'll read the job description and it'll try to make sure that your resume and your coverage

Quote

As far as like how quickly things are changing and how They don't feel like they've changed that much, right? It's been like four plus years. Um and there has been like a pretty coherent arc of like first we did autocomplete then And we were like, um we want to have chat conversations, ChatGPT came out. out. Let's ask agent, let's ask not even agents. Let's ask asked LLMs about our code bases. Um and then we gave them tools. Right. And they started writing code for us. And then we moved to the CLI and now we're kind of in this like uh let's go back to the UI and apps arc. But like over four plus years, that's not that much. I do think it's accelerating so much. For like the first three years it was like, Okay, so what are we doing? Are we actually changing that much and now it feels like oh yeah we're like yeah. But like it i I it is important to take a longer view of it because day to day it feels

Quote

That's much closer to an agent. It really is an agent at that point. than it ever has and like these reasoning models where it gives itself a what type of information it should return. That's closer to an agentic workflow. I would almost argue any chat bot you're interfacing with, like ChatGPT, Anthropic. Claude, Gemini, those are agents now. I don't think they're just chatbots anymore. Interesting. And I and as far as the risk, I the risk is much higher for agents because it's being thrown at people. Everyone's saying These agents can save us a ton of money and do a bunch of work for us. Let's experiment get this going fast and get it out in the wild so we can start saving that that that money and getting this free work as as done as fast as possible. I think a lot of the risk is coming from speed of delivery. There are definitely companies. We've had it sure sounds like that's getting pushed to the side right now. Security is We saw this with web app.

Quote

boot camps that have a job placement guarantee or have some kind of demonstrated career your help. That's kind of the first thing that I look at. Then it's also really important to see that the curriculum aligns with what you're actually interested in because there's a lot of boot camps or actually programs in general right now that call themselves AI, but then they're actually Actually one of two things. So they'll either be just teaching you how to prompt engineer on ChatGPT. which is not really what anybody's looking for, or they will be machine learning. And so they'll have several weeks on data cleaning and EDA and So if you're interested in the AI world, I would focus on the other side. on programs that talk about prompt engineering, fine tuning. Those would be kind of like the big buckets that I would be looking for so that you could at the end have the skills that you need to be building these AI apps. There's a lot of different programs right now. There's a ton. There are many programs. programs coming from universities and I have found just

Quote

The thing that frustrates me the most is something you actually said earlier in this conversation for a different time. reason it's sort of that I have to start from scratch every time I call up a new window for I I use AI all the time, like constantly, constantly, constantly, right? For all different purposes. But every single single time I feel like I'm starting from scratch. And I think that the ways that we have of combating that, whether it's your your agents.md or your call dot MD for your written instructions uh for certain agents or Or you know, um ChatGPT has this great memory feature. But what they're really doing is they're recording like Facts. And I find that the thing that I spend a lot of my time doing when I'm trying to jumpstart an AI. conversation is I'm trying to teach it more stylistic things, right? So if I'm trying to iterate on a draft and edit something, I need to really quickly get my preferences about style into the the brain of this thing into its context engineering. Uh into its context and I need to do this context engineering, right? This is a place where I don't see great solutions in the market. tend to be sort of just bigger versions of that glorified key value version of memory and i i'm being silly there that's not fair but that's kind of what it is right a fact gets retained as memory and it gets recalled later. And so my best case of jump s…

Quote

into the generative AI side first. Yeah, so this is our first scenario that we're really are seeing a lot of interest in. So I mean no one can Ways that we can leverage not just Chat GPT or Dali or those other tools that have become really famous, but how can organizations make use of general generative and autonomous AI to make their life easier. So some of the things we're looking at at the moment is how can we use ChatGPT and other technologies and our algorithms to actually harness your corporate knowledge, harness, understand what files and information and stuff that you have around your organization, bring it together using natural language processing to actually really help capture what the essence of that knowledge is, where you are, and actually help mitigate a lot of risks that a lot of organizations seeing with an aging workforce, with uh uh a remote workforce, uh we've seen that things like generative AI is gonna be a real uh game changer here. Yeah, there's a lot of um generally uh in corporates in in large organizations or even even small and medium ones, there's always a challenge about knowledge Knowledge management and having you know, being able to access the right information at the right time. Um, people who are knowledge workers spend a lot of time looking for information.

Quote

filters and then because the a lot of that data was multimodal we had to build And so I think it was when we built all of that we then talked about um integrating with the ecosystem. So we build integrations into DuckDB, for example. And so Initially we got a bunch of community interest pre chat GPT because of the duck db integration. But then when ChatGPT came out and everyone started building RAG applications and personalized recommenders that were based on you know semantic similarity, the open source community notice, hey, like I can store the data here. There's a vector in Guess what? Like you're a free vector database. And so We kind of saw a lot of interest from the open source community that way and we kind of realized w where that opportunity was. Right. And so um it's all about like, okay. Okay, you have all this data. What is the thing that people want to do with the data and what is the value they want to get out of that data. And that first use case that came around was was um was the vector search. And so I think for the past couple of years, the community probably knows in the market thinks about Land C B as like a vector database and and that's sort of like the primary reason is that's the slice within the AI development

Quote

Learning approach that is not based on the whole new generative AI side. And they work really well. And those systems, by the way, are not trans like you know, they're not moving to using like, oh let's replace that with like a chat GPT because Fun fact chat GPT doesn't really work that well with numbers. Like not chat GPT and it's uh only Chat GPT, but like this whole transformer architecture doesn't really work quite well. with numbers out of the box, at least like not in the way that they're applying it in ChatGPT. Let's circle back real quick to the use case you mentioned before for for the dashboards and you mentioned I it didn't work quite well like For me as a customer, right, if the patterns, right, of my product, say, you know I offer widgets right that are very popular, you know, during the holidays. here and there, maybe you know, my company gets featured on Hacker News and I get you know bombarded with traffic, whatever. the case may be. But you know, if if I'm if I'm a limited staff operator or perhaps Maybe I don't even have an SRE or somebody who really understands observability, but at least they can describe According to to previous sales data. Can they provide enough information to you?

Quote

check. You know, should we check after generation? Should we check before generations? Should we do both? Uh, you know, should we be Integrating other searches as well. Yeah. So uh on the documentation we do have have a way to connect our MCP server as a connector. to Claude and directions for um ChatGPT as well. Um at generally these large language models, when you ask me why They haven't uh lowered their price per API call for the same. Search. One of the things that's uh pretty well known is that uh the the GROC models, XAI models, um called Brave and uh anthropic calls brave if you're in clawed code it even tells you Hey, I'm calling Brave. So they're kind of uh you know stuck with that pricing. Um you know, and even if they get a discount, you know, it's they are really stuck with the brave pressing and then um the overhead of calling etc. So I think that's one of the reasons why the price hasn't dropped. Um and the other one is you know building uh some