Discover/Grit/The Grittiest Conversations of 2025: AI,...

Would the academic community have created a different architecture under a different name that looked very similar and felt very similar? Maybe Google had its proprietary transformer model, which nobody could use because it was IP and it didn't become very popular.

Grit · The Grittiest Conversations of 2025: AI, Business & Beyond

“And so maybe Google had its proprietary transformer model, which nobody could use because it was IP and, you know, it didn't become very popular.”

24:25 / 24:45

Or would have the, you know, academic community created a different architecture under a different name that looked very similar and felt very similar that carried the weight instead. And so maybe Google had its proprietary transformer model, which nobody could use because it was IP and, you know, it didn't become very popular.

Why this clip

Counterfactual speculation about alternative AI timelines that challenges the narrative of Google's mistake. The contrast between open vs proprietary development models resonates with current tech debates about open source AI.

24:25 - 24:4520sBold/Contrarian

Share

LinkedInX

More from this episode

Grit

Bold/Contrarian

I don't think Google had any idea then. Otherwise, they probably wouldn't have given it away. Would it have blown up like this if it's been a secret? Would Google have created the ChatGPT moment?

This is a provocative counterfactual about one of tech's biggest strategic mistakes - Google giving away the Transformer architecture that enabled OpenAI's dominance. The hook questions whether AI would have exploded without open access, challenging the narrative of inevitable progress.

24:03 - 24:2219s

Grit

Domain Expertise

Most of the data we produce, if you were to read it, you could lock me in jail for thirty days and ask me to solve one of these physics problems. I can't even barely read the type of data that we create because it's really experts in PhDs.

Vivid metaphor about the complexity of AI training data that makes technical concepts accessible. The 'lock me in jail for 30 days' imagery is memorable and the self-deprecating admission creates relatability while highlighting expertise gaps.

23:23 - 23:3917s

Grit

Consequences

If the AI crawlers go crawl the web and get all the content, when you go to your Perplexity or ChatGPT, do you go back and look at the original source? No. You just read it right there. The original source isn't getting compensated.

Identifies a fundamental business model disruption in clear, accessible terms. The rhetorical question format engages viewers while highlighting an economic consequence most users haven't considered.

30:05 - 30:4035s

Grit

Consequences

We've never seen dollars spent like this. What if they're wrong? What if this is maybe not even enough investment? What if Microsoft decided not to do it, and Meta and Amazon did? They missed the whole thing.

Series of escalating 'what if' questions that capture the high-stakes nature of AI investment. The contrast between 'never seen dollars like this' and 'what if it's not enough' creates tension that drives engagement.

8:51 - 9:0716s

From the blog

FrameworkOriginal Research

The 5 Clip Archetypes for VC Podcasts

10 min read

Want clips like this for your podcast?

We find your top 5-8 clips, write the hooks, and deliver ready-to-post content. First 2 episodes are free.

Get 2 Episodes Clipped Free