I completely agree with this
AI allows depth of research at scale and hyper-efficient idea generation to reduce time spent on bad ideas. Both can contribute to real acceleration in idea velocity.
I think a few things will start to happen:
1) At tight coverage firms, coverage will expand. It is typical for a multi-manager to cap analyst coverage at 50 names. The manual steps required to be deep on 50 names: models built, updated, data work done, sell-side check ins, IR check ins, earnings previews, etc, necessitated these tight coverage areas. When I covered 300 healthcare stocks, I did so with a team of 9, because each analyst had infrastructure and coverage on 30-50 names and we were ready to pounce on any idea at any time (this low latency reaction time has been a key advantage of the multi-manager model...it doesn't take 3 weeks go go through IC to react to a stock dislocation). That 50 stock cap could *easily* today move to 60-80 and I could see it going higher without a lot of comprehension drift.
2) At sectorized but less tight strict coverage firms, the best investors on the team will pick up new coverage areas. At a Tiger Cub and your Consumer Sector Head leaves? Give that coverage to your highly talented Industrials Sector Head, and now he covers both. Would have been very hard to do, and now that combined team definitionally spends more time on the road (BOTH industrials and consumer conferences and HQ visits), but it's possible because there is less infrastructure maintenance to do.
3) Quant firms will, finally, start to build successful Systematic Fundamental businesses, tapping into the same alpha pools that human teams reach today. Many have tried and failed over the last two decades (it's much harder than it seems, having been a part of a few of these), but the technology is finally close enough to attempt this. This will still require a human team, but it's not inconceivable that you will be able to recreate the $ alpha impact of a multi-manager L/S equity business line with one tenth of the headcount. In typical AI parlance, moats everywhere are melting, which, as Chandler points out, is better for idea velocity for humans. But this moat of talent & friction kept machines out of these alpha pools, and that moat is going away. I believe multi's who don't adapt will have a really hard time over the next 3-5 years (and also observe that multi's and the talented, incentivized individual pod teams are some of the most adaptive, creative, hard working people in public markets).
I loved this interview with
@PitchThePM and Chandler Bocklage (former PM at Point72 that sat directly next to Steve for a decade and current Head of Business Development at Point72 in charge of hiring analysts and PMs)...
“Usually, you have to be narrower to try to get deeper. But [Ai] might actually allow people to be as deep but broaden out, which means that you can have a wider universe of potential alphas. It’s pretty exciting.”
“Everything can’t be a two-year idea based on compounded earnings growth…Those are great ideas and the thing you can anchor your book on, but you have to have other things that might work in the intervening two years. Very few places are only looking at the P&L over a two-year block.”
“You can’t have the same type of ideas…you have to have different themes, different durations, different catalysts…”
“To be a long-term successful PM, you can’t be a one-trick pony.”
"Alpha decay is compressing. There are more and more people chasing the same alphas.”
“Can you teach what you do and what makes you special to somebody else and are you willing to teach what makes you special to somebody else? There are a lot of people that aren’t.”
“When a war breaks out, things change, and the PM might have to think about that a little more holistically than the individual analyst.”
Thoughts
@pmje73,
@DanielSLoeb1,
@davidein? 🙏💙
(Not investment advice).