A Beginner's Guide to AI

The episode offers one usable framework - a three-layer AI stack (infrastructure, model, app) with a rough directional capital allocation claim - but the rest is philosophical meandering, hope messaging, and fantasy book discussion. Platitudes heavily outnumber actionable insights for a B2B operator.

“I would say the investment trend is huge boom in application layer for sure. A lot's still happening in the model layer... near term investment trend is m like 60% at least. An early stage venture”

“AI is a tool. And like a hammer. A hammer, you could use it to build a house or as a murder weapon.”

The guest explicitly flags her own analogies as circulating ones ('sometimes you hear this analogy'), and the remaining takes - AI as hammer, Internet comparison, alignment problem, hunter-gatherers worked less - are all well-worn. The somatic/ecological/communal intelligence framing is mildly interesting but not developed with rigor.

“we're in inning one of a nine inning baseball game. Sometimes you hear this analogy, right? Like we are so early”

“Our somatic intelligence, our connection with our physical body and the wisdom there. Our, our ecological intelligence, how deeply we used to live with the land”

Tallulah Le Merle has genuinely multi-angle experience - FTSE 100 digital transformation consulting, CEO roles at AI scale-ups, and an allocator role in an investment firm - but named firms are absent, the investment practice size is unclear, and the conversation skews heavily toward thought-leadership and a forthcoming book on 'hope' rather than hard-won operational specifics.

“I'm a partner in an investment management firm and I lead our AI strategy. So I'm looking at how to allocate capital into early stage AI startups and into venture funds”

“I was working for a large strategy consulting firm. My clients were big corporates, like FTSE 100 companies... Johnson and Johnson and Mars and HSBC”

The episode names a handful of real companies (OpenAI, Anthropic, Fireflies, Otter, Canva), provides ballpark valuations, and cites a WEF report - but the WEF figures are immediately disclaimed ('don't quote me on this'), valuations are approximate, and no deal specifics, fund sizes, or outcome data are offered.

“OpenAI is like $500 billion valuation. Anthropic is like 200 billion probably now or something”

“if you look at the World Economic Forum report where it does say something like 90 million jobs, don't quote me on this, but something in that order of magnitude will be replaced by 2030”

The host asks broad, poorly formed questions, frequently agrees and affirms rather than probing, and derails a significant portion of the episode with a personal digression about fantasy fiction and Harry Potter. There is no meaningful pushback or follow-up pressure at any point.

“taking a look now, now the people can wake up that, that tuned off. Um, because we always talk about investments and stuff. Where do you see is the, I mean you are at the pulse of the, of, of startups”

“for me, what happens, I started my life as reading science fiction and a little bit of fantasy. Now that there comes more and more tech, I read more and more fantasy. What do you think to that?”