• What a modern LLM does during training is, essentially, very very quickly skim the textbook, the words just flying by , not spending much brain power on it.


• Rather, when you or I read that math textbook, we read a couple pages slowly; then have an internal monologue about the material in our heads and talk about it with a few study-buddies; read an

... See more

75 percent [of our] acquisitions come from approaching celebrities, politicians, athletes, the “celebrity adjacent,” etc. That way, we can control the content…. We are approaching authors and celebrities and politicians and athletes for ideas. So it’s really we are on the look out. We are scouts in a lot of ways…

For someone to be successful in starting a company and then turning it into a multigenerational business, they need to have good taste in at least these two things: ideas to pursue and people to recruit.

Ramp keeps finding itself in the right place at the right time because the company is built to be in the right place at the right time, executing against a mission that will never go out of style with a structure and team that allows it to take advantage of new ways to fulfill it.

For the same reason these systems will be helpful with anything, we should expect that these systems will be able to help with the problems that they create. If we believe there are solutions to unhappiness or a feeling of a loss of purpose, and that these solutions can be found with intelligence, then we should expect these systems to be able to h... See more

There is a potentially important source of variance for all of this: we’re running out of internet data. That could mean that, very soon, the naive approach to pretraining larger language models on more scraped data could start hitting serious bottlenecks.

Without personal moats, we may find ourselves in a perpetually evolving, seemingly more meritocratic game.