Joe Seager-Dupuy

Early stage Consumer VC in London | Post mostly about VC, tech, start-ups | Worship at the altar of Jiro 🍣

[Joe Seager-Dupuy @jsdxyz]

General principles:

[Joe Seager-Dupuy @jsdxyz]

As per the eternal wisdom of Michael Porter, it’s the companies stuck in the middle who are buggered when markets bifurcate… in other world: be amazing, be cheap, or be ready to go bust.

[Joe Seager-Dupuy @jsdxyz]

This feels obvious. There’s a pattern.

[zerohedge @zerohedge]

"the share of tokens used for US models on OpenRouter has collapsed": Bloomberg

[Joe Seager-Dupuy @jsdxyz]

The big AI labs probably will end up having a product portfolio addressing both segments (after all Volkswagen owns Lamborghini and Skoda) - the cheap stuff just isn’t the focus/race right now.

[zerohedge @zerohedge]

"the share of tokens used for US models on OpenRouter has collapsed": Bloomberg

[Joe Seager-Dupuy @jsdxyz]

Hours in => value out is the business equivalent of calories in => calories out in nutrition. Yes, it’s sort of true, but most of the value is understanding the nuance.

[Joe Seager-Dupuy @jsdxyz]

The 996 debate misses the point entirely.

It’s not about self-flagellation. It’s about velocity of iteration.

[Joe Seager-Dupuy @jsdxyz]

So should everyone hustle?

Yes if: you’re working on a high value, unsolved problem with a steep slope of learning and high marginal value per unit of new learning.

Not necessarily if: the above does not apply.

[Joe Seager-Dupuy @jsdxyz]

$300bn+ of market cap lost in 1 day in 1 stock.

Wild times.

[Joe Seager-Dupuy @jsdxyz]

@a16z General mental model: delegate the information gathering, own the decision