HD in HD

There are genuine operational gems - the jet-ski innovation model, city-level isolation for testing, and the AI agents replacing 5,000 FTEs - but they're embedded in stretches of biographical filler, broad platitudes about culture, and repetitive framework-naming. The insight-per-minute rate is uneven.

“We have today 20,000 agents and we are running agents, for example, to talk to restaurants, partner customers. And we have today we calculate around 5,000 extra people working in pros just because our agents are behaving as regular employees”

“iFood has 2,000 cities, it's quite common that we take 10 of them and we say these jet skis run an app that only runs on this city and then they can do whatever they want on that city”

A few genuinely fresh framings - capital scarcity as a competitive moat outside the US, the European competition authority creating 'local losers' instead of global champions - but these are interspersed with well-worn founder-hero narrative, Jim Collins name-drops, and standard Silicon Valley-vs-rest-of-world takes.

“The competition is not between Italy and France, is between Europe and us, Latin America, India and China”

“The Chinese are not copying, they are innovating. To operate a big business with thousands of people and millions of orders per month, they have to be super good to do that”

Bloisi is a genuine practitioner who built iFood from 25,000 to 160 million orders per month against Uber and SoftBank-backed competition, and now runs a publicly listed global tech group with $15B+ revenue and 40,000 employees - not a thought leader, a real operator at scale.

“This month iFood did 160 million orders. So we grew from 25,000 to 160 million”

“iFood has more than 90% market share in Brazil”

When numbers appear they are compelling - $20B GMV, 0.7-0.8% of Brazilian GDP, sub-$0.01 per Rapido ride, 20,000 agents - but the guest repeatedly declines to supply specifics ('I forgot the name,' 'I don't remember their number now'), and financial projections like 'a few billion dollars in profits' go completely unchallenged.

“They charge like the number is...it's less than $0.01 of dollar for sure per ride. And they are the double of the size of Uber”

“Selling $20 billion per year only in Brazil...It's like 0.7, 0.8% of the Brazilian GDP”

The host surfaces some genuinely useful follow-ups on the jet-ski dependency problem and the ambidextrous tension, but far too often reacts with 'that's awesome' or 'very interesting' and lets vague financial claims, forgotten company names, and a self-serving investor pitch pass without challenge.

“Oh, I didn't even know that.”

“But then how do you deal with the dependencies of the uh, I guess the test because I imagine you guys are going to build some new thing to test on ifood some other team needs to coordinate for dependencies”