Robert Ta and Jonathan compare how Google and Palantir deploy AI engineers differently. Google's approach makes databases queryable for AI agents, while Palantir builds custom ontologies and knowledge graphs tailored to each customer's business, giving engineers ownership of actual outcomes.
Stratechery by Ben Thompson came out with an article around the deployment company. They titled it Deployment Company Back to the '70s: Apple and Intel. The deployment company meaning, uh, OpenAI's new venture to help deploy AI into enterprises and win the enterprise landscape. We talked about this on the last podcast where OpenAI and Anthropic are both investing in partner ecosystems. Essentially catalyzing a lot of forward-deployed engineer work. Jonathan shared with me of what happened and what nobody's really talking about. And he had this take that I thought was really interesting around the difference between forward-deployed engineers to Google and forward-deployed engineers to Palantir. Jonathan, do you want to talk about that real quick?
Yeah, I found this a little surprising. This is Ben Thompson of Stratechery. He was interviewing one of the Google program leads who's rolling out the FTE program. And they were comparing like, you know, the Palantir FTE model of building like ontology and understanding the application and integrating software and AI into it versus the Google forward-deployed engineer model where they really just gave an example of like, you know, you're trying to solve a problem, you have data in a database and you need that data queryable and in your, in your Gemini sessions. And with Ford, deployed engineer is there to get that data and to make it queryable and to like, you know, make sure that any of the range of questions that you need to ask the agent, it has indexed data for. So it is not building an ontology, is not trying to understand the business. It's really just like a brute force, like let's get the data in the system. It's almost like, you know, it's like a bitter lesson FTE for Google versus an ontology FTE at Palantir. So it's an interesting distinction.
