Sushanth (Shushant) Raman

speaker

542 appearances 1 recordings 1 series first heard Jul 2026 last heard 22 Jul

Sushanth (Shushant) Raman’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.

Appearances

newest first · ▶ plays the moment

How I Invest with David Weisburd · E406: Why AI Won't Transform Most Enterprises for 10 Years · 22 Jul 2026 podcast

Like maybe the leader's wrong.

Maybe the system in place doesn't make sense and you need to work with your team to re-architect the system.

Like whatever it is, figure out the structural solution versus don't figure out the bandaid.

What you have to do is you have to make sure your product benefits from every sort of foundation model, like evolution, right?

So like the way we think about Palette is we have all this context about these businesses that's stored, right?

Like for example, what are their SOPs for running quotes?

What are their standard operating procedures for tracking order shipments?

What are the ways that they handle orders from Walmart versus Target?

So there's all this institutional context.

There's an execution layer built on top of the models.

So every model release,

should make our product better because we have our context we have our execution layer we have our data layer we have our data integration layer so in many ways we're like a suit a next generation system of record that contains context but also does these actions so every model release arguably makes our product more performant

but we're not at the risk of the models because they don't have all the institutional context, ingredients, the data that we have, so they can't come and displace this overnight.

And I think that's the right way to architect a next generation AI company.

And the wrong way to do it is to basically go and write all the deterministic code

And then when the model upgrades, all that code and that logic you wrote is kind of useless.

You have to rely on the models to kind of go and do these computational intensive tasks.

But you have to set up the architecture and the execution engine and the context layer that takes advantage of them.

Yes.

The rules of the game have changed in a lot of ways and you have to rethink everything from first principles in terms of team building.

Showing 421–440 of 542 · page 22 of 28