Trade-offs

Also, I need your China recommendations

Next week I am travelling to Beijing, Shenzhen and Hangzhou. I haven’t been to China in over a decade so looking forward to seeing everything in person (which may lead to either more or less frequent posts). Recommendations for things to do and people to meet very welcome.

I’ve been reading a lot of 1920/30s history over the last few months (btw, I maintain a reading list for founders here) and I keep coming back to the idea that it’s entirely possible for seemingly likely things not to either not happen or take far longer to happen than anyone believed.

Dwarkesh and Dylan Patel had a very interesting conversation about an impending set of (capacity and materials-related) trade-offs which will have to be made between continued consumer electronics progress and improved AI capability. It’s really worth listening to. This is clearly going to play out in very interesting and not entirely predictable ways. You can have an abundance of intelligence but you might have to trade an abundance of device choices in order to access it. The war effort of the 1940s becomes the token effort of the 2030s? We shall see. AI as the next technology paradigm is a given but the timeline for its diffusion may be surprising.

This is not a surprising position to take for anyone who’s been around startups their whole life. Jerry Neumann just wrote a (predictably) good essay about how, for the literal industry of startup advice which exists, the success rate of startups hasn’t actually shifted in ten years. Although outlier returns have increased, venture backed companies now have a higher rate of failing than a decade ago. Building a vision is hard and nothing is a given, no matter how inevitable it might seem.

Coming back to trade-offs for a second. Asking about trade-offs is one of the more useful questions a board member can pose to a founder/management team. A lot of startups (and their boards) make decisions without realising they are in fact choosing trade-offs. A lot of unplanned scenarios happen because strategic decisions can have non-immediate impacts elsewhere in the company. It is like a slow-acting set of Newtonian business physics.

Other things

Aibl are building an AI Enablement directory for the UK and Ireland, which I think is much needed.

Max wrote a very good post on the new kind of holding company financing that’s needed for AI services. A few people have quoted this back to me.

History is not always written by the winners, it’s written by the historians, who can be a much more nefarious group of people. Ada Palmer’s book on the Renaissance is an excellent and dense combination of history and the history of the history.