The Aarthi and Sriram Show

The episode has a handful of genuinely useful observations - ElevenLabs being a third-place player at Series A, the LLM data-exhaustion vs. video data-abundance split, and the reversal of business-school/engineering-school dynamics - but these are surrounded by long stretches of platitudes ('great time to build,' 'focus on one use case,' 'models are getting better') and extended personal anecdotes that add no transferable insight.

“some spaces like LLMs, we are now like running out of data and looking at reasoning and architectural improvements. But I think in other spaces like video and 3D and music, there's still a ton of room to run on like the core data side”

“Consumer businesses used to not monetize for like 10 years and now a lot of them are like making real money with strong retention from day one”

The 'cosplaying as a founder' framing and the Stanford dropout credentialing point are genuinely fresh angles, but the bulk of the conversation rehashes standard 2024 VC talking points - wave 1/2/3 of AI, narrow-focus-wins, product velocity as moat - that have circulated widely without adding a first-principles argument for any of them.

“we have too many young people, like cosplaying as a founder, like, it's easier than ever to like look cool on Twitter and know the terms and like be in the in network”

“the business school students are going over to the engineering school and begging to be like, led onto this amazing product as like an intern on strategy”

Justine and Olivia Moore are legitimate AI-focused partners at a16z with eight-plus years in VC and a verifiable portfolio including ElevenLabs; they bring real deal experience and pattern recognition across hundreds of AI companies, though as investors rather than operators, their claims are observational rather than hard-won from building at scale.

“When we invested in 11 labs in the text to speech space. I think they're now doing amazing. Have a ton of huge customers are, are doing really well in that space. But they were like the number three player or something at the time that we did the Series A”

“there was one at CRV Justine, where he pitched us four or five times on five different businesses. And the fifth time we actually invested”

There is a reasonable volume of named companies (ElevenLabs, Granola, Gamma, Kreya, Bolt, Flux, World Labs) and a few concrete situations (ElevenLabs as third-place player at Series A, ten Stanford classmates dropping out together, 200 vertical-B2B companies in a YC batch), but quantitative evidence is almost entirely absent - revenue figures, growth rates, and market sizes are referenced only as vague gestures like 'millions of dollars' or 'quadrupled their run rate.'

“they were like the number three player or something at the time that we did the Series A. Yeah. But um, and, and for a long time text to speech had been considered like it's way too crowded”

“I've seen cases where companies are already paying millions of dollars a year for one AI product”

The host asks a reasonable set of thematic questions covering funding, moats, enterprise, and founder backgrounds, but almost never follows up on a specific claim, pushes back on a vague assertion, or asks guests to defend a position; she frequently takes over the airtime with her own extended personal stories (bolt.new, Claude writing coach, resolutions coach), and the session ends with a pure softball favourite-product question.

“Yeah, yeah, totally”

“What's your personal favorite AI product? Something that you use all the time or you can't live without, or it's like a sort of specific use case that you're obsessed about”