Person
1 episodes · 19 quoted claims
Susan Li says she started kindergarten at age four and graduated from college at 19, giving her 15 years of formal education.
“I started kindergarten when I was four and I graduated from college when I was 19, that having 15 years of formal education is, you know, I'm woefully undereducated”
Susan Li's first day at Morgan Stanley was on the trading floor at the company's headquarters at 1585 Broadway.
“When I showed up at Morgan Stanley for my first day, I was on the trading floor in the big Broadway headquarters at 1585.”
Susan Li joined Meta in April 2008 as an IC4, a junior individual-contributor level, in finance.
“I joined as an IC4 in finance. So really actually pretty far away from the center of the”
Susan Li began her Meta career mostly doing revenue forecasting, which she describes as the mathiest part of finance.
“So I started off my career doing mostly revenue forecasting, which was kind of like the mathiest part of finance.”
David Ebersman, then Meta's CFO, spoke with Susan Li as she weighed moving into a News Feed product management role or broadening her finance duties.
“I remember sitting down with David Ebersman, who was our CFO at the time, and he looked at me and said, look, I know you're considering these options.”
Susan Li says Mark Zuckerberg has held all-hands meetings throughout her 17 years at Meta and has become an excellent public speaker.
“Mark has done all hands for all of those 17 years and he clearly has become now a truly excellent public speaker”
Early in her Meta career, Susan Li was invited to a meeting of the five heaviest SQL users, who were consuming too much infrastructure power.
“the meeting organizer tells us that we have been called because we are the five users of SQL who consume too much power.”
Susan Li says Reality Labs' annual revenue over the next 20 years is very hard to forecast, so Meta evaluates such bets by what they must be worth to pay off.
“It is very difficult to pencil out what the annual revenue forecast for Reality Labs is going to look like over the next 20 years.”
Susan Li says Meta went from an e-commerce-fueled peak in 2021 to negative year-over-year revenue growth for the first time.
“we really went from this e-commerce-fueled heyday in 2021 to now negative year-over-year growth for the first time, which is obviously very alarming.”
Susan Li says Mark Zuckerberg once gave her an EBITDA hat as a gift.
“the story is that Mark at one point gave me an EBITDA hat which was a very kind gift from him”
Susan Li decided Meta should emphasize free cash flow rather than EBITDA, and now gives out free cash flow hats.
“I realized pretty quickly that we actually as a company shouldn't be wearing free cash flow hats instead”
Susan Li says Meta and its peers are making large capital investments to train frontier AI models.
“And then now you have, among many of our peers and ourselves, this big investment to train what we all aspire to be frontier models.”
Susan Li hopes research advances will allow AI training to be distributed, easing the need for one giant cluster and reducing energy constraints.
“hopefully, there will be different kinds of research innovations along the way that will unlock things like being able to distribute the training so you don't need sort of one extremely large cluster”
Susan Li says that after Meta's October 2022 earnings call, some investor callbacks had no questions and only feedback.
“And in October 2022, for the first time, there were sometimes no questions.”
Susan Li says headcount is easy to account for at Meta because org charts show who works on what.
“And headcount's really easy to account for because you have org charts.”
Susan Li says Meta keeps GPU capacity fungible so it can be redirected to where demand spikes, such as in India.
“you need to divert capacity to where suddenly something has happened in India and you want a lot of compute to be available to be used there.”
Susan Li says that when Meta budgeting allowed trading headcount for compute, people almost never chose compute.
“people almost never trade for compute for exactly the reason I described, which is that if they get allocated 100 new headcount”
Susan Li jokes that she acts as an expensive machine learning model by approving routine employee expenses.
“I say that as a person who is like a very expensive machine learning model for approving expenses”
Susan Li says account lockouts are a problem where AI could make human reviewers far more productive and justify more investment.
“everyone knows someone who has gotten locked out of their Facebook or Instagram account.”