Human Centric AI – A Fireside Chat

Automation, do you think it competes against the quality? So we set up the quality rules and make sure that we're conforming to quality rules, but I continue to believe that a human in the room continues to be needed. Somebody who has a good understanding of the business, and a good understanding of the data.

If a hallucination was manifested, they'd recognize it right away. Because the thing with hallucinations is you don't recognize them as hallucinations. So I think that human-centric AI is going to continue to be important, where you've got humans that have domain expertise and have the ability to identify when things have gone wrong.

And then you've got to track back to how through lineage, right? How do we get to this point where we're making some bad decisions?

Tooling plays a big role because one of the challenges that we have always seen, and the industry has always lacked, that there was just, there was just not enough tooling for data engineers or any kind of data personas where they can do these things. So investment in data warehousing and data lakes, but it didn't really actually help us getting the data in the right place. Tooling is hugely important and I think building the right tools for the job, right?

So understanding what are the jobs they need to get done and how do we build tools that allow them to do that very efficiently, but gives them the flexibility that they need. I think engineers, they want to have a control as part of the tooling. They're exercising that control in making better decisions as opposed to doing mundane jobs that could have easily been automated.

I think automation of repetitive tasks is incredibly important because repetitive tasks are boring, so you're gonna make mistakes because you're bored by it. Automating the repetitive stuff having the engineers focus on where they can really deliver more value to the business. I think it's what satisfies everybody.