MFOHome About Commentary Discussions Funds The Best Resources Support Us!

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Register

Categories

In this Discussion

Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.

    Support MFO

  • Donate through PayPal

AI is Coming For Your Fund Manager

beebee
edited December 6 in Fund Discussions
Interesting:
If all this works, we may see the end of the investment management business as we know it. Instead, each person could have a personal AI that combined personal information — tax situation, financial goals, income prospects — with superior market knowledge, and that never slept or had its attention wander. These AIs could trade with each other with no need for human intervention.
asness-ai-end-human-fund-managers

Comments

  • Sven
    My counter-argument is that if Ai is so great, why are self-driving cars keep hitting pedestrians and other accidents? Thought the Ai algorithm can perform more calculations simultaneously than human can. Driving should be quite straightforward even in urban areas.
  • hank
    Maybe they’ll grab Hussman.
  • Old_JoeOld_Joe
    edited December 8
    @Sven- I hate to admit this, but I chanced to follow a "Waymo" vehicle around in SF traffic for a while the other day, and I was really quite impressed with the way that it handled itself. Very quick to recognize exceptions to normal traffic flow and to reposition itself accordingly, without causing problems for other vehicles in the immediate area.
  • Sven
    @Old_Joe, in the last several years we have had mixed experience with Waymo taxi mostly in San Francisco area. I have to admit this year’s ride was smooth comparing to prior ones. The car nearly hit a pedestrian when it did not slow down at the intersection. This year is better. I think the improvements are likely to be Ai are the generative learning type where it continues to recognize the inputs and refine its algorithms.
  • Derf
    AI comes in different versions, say #1 to #5 with 5 being the highest. For investing & driving I'd take a #5. I have a #2 & 1/2 in my Subie & didn't get what I paid for in my estimation.
    Good luck with your AI, Derf
  • ShostakovichShostakovich
    edited December 9
    I've wondered for some time if active management will die out pretty rapidly or be replaced by a smaller number of active managers driven by AI wars. Perhaps the major quant hedgies are already doing this. Perhaps it really is time to go all-in on index funds.

    And @Sven, I have to wonder what a reasonable expectation is for self-driving cars to get the kinks worked out. I gotta think AI-driven portfolio management is a tad easier (and had a head start) over self-driving cars. IDK.
  • Observant1
    "I've wondered for some time if active management will die out pretty rapidly
    or be replaced by a smaller number of active managers driven by AI wars."

    Here's my opinion on what may occur to firms who actively manage funds.
    https://www.mutualfundobserver.com/discuss/discussion/comment/183896/#Comment_183896
  • BaluBalu
    AI will eventually do tasks better than humans in all tasks. If the task is being evil, it will be more evil than humans can. E.g., if the task is to deny insurance claims, it will deny even more claims and perhaps can come up with even more B.S. reasons for doing so. If the task is being a safe driver, it will be a better driver than humans and reduce accident rates. If the task is killing, it will be a more lethal killer. If the task is to propagate extreme ideas, it will do a better job of that too.

    You got the idea.

    Good luck to all of us.

    With the above in mind, where and what to invest in?
  • AnnaAnna
    edited December 10
    I guess I am slow. I just realized that artificial money is the currency of an artificial intelligence world. So, it appears that investments have already started and, as far as good and evil go, I think the evil has the edge since the edge goes to the most motivated powerful. So, investing in the area that targets powerful but good is the way to go but... Are there any investments that put good above more power and more wealth? I would think that sucessful investments and investments in good stuff is probably a contradiction. So, investments in powerful companies that drive the future independence of the programming and decision making tools of AI, independent of anything other than concentration of wealth might be the way to go.
  • FD1000
    No need AI to take over. We already know thru history that the SP500 index simplicity beats most funds over the decades and why VG is so big.
    High tech is coming after other job for decades and AI will take over additional white collar jobs.
    The gap will continue to grow no matter what any government may do.
  • ShostakovichShostakovich
    @FD1000 - what if all market participants have AIs that converge (say, 98%) on the same trades / stock picks? Does the market become perfectly efficient and the market trend go flat?
  • FD1000
    The theory is that if AI figures out that NFLX is a great company, it will generate buys for many people.
    Since the stock market movements are random, AI will not guess them properly.
    As someone who worked in IT over 35 years. We used to say garbage in, garbage out. That means you got to feed it the right data.
    What is the right data? That is what everybody looks for. Companies with the best analysts are still looking for it.
    Final found it and why it's mostly unknown. (https://blog.grainstonelee.com/insight/who-is-final-israel)
  • ShostakovichShostakovich
    @FD1000 - everything you say makes sense, but the sheer volume of data that is capable of being managed by companies (of all sorts) is increasing exponentially. And, I have to imagine that in short order the types of data that do much of the heavy lifting in forecasting and prediction (i.e., account for 98% of the variance) will be settled upon. So, gaining access to that last 1.5-2% isn't going to move the needle much for many investors. Or maybe imperfect data can be overcome with a super-rich model (as Bayesian maven Gelman says "Big Data Need Big Model").

    Now, I guess the counterpoint would be that if JPM, DE Shaw, GS, and kindred competitors spend gazillions of dollars shaving 1/1000000th of a nanosecond off a trade execution time, they might find a way to convert that 1.5% information edge into some real money. I don't know.

    I guess I can just see a point where there are fewer meaningful asymmetries to leverage. I think Charly Munger or Tweedy Browne (or etc.) said at one point that you used to have pour through pages of out of date trade journals to get a sense of companies' profitabilities and obligations, but now that can be sifted through readily by computers and databases.

    So, again, while I was never all-in on indexing, I'm guessing that this spells the end for large chunks of the biz, with only a few super brains-in-beakers types of firms like Shaw or Guggenheim or etc. I don't see how First Eagle or Harris Associates or Wellington or FPA or Osterweis or etc. last another five years.



  • sma3
    Interesting re Israeli AI firm. I guess the IDF wasn't using them Oct 7th?

    I am always bemused bu the multiple attempts over the years to replace MDs and their diagnoses.

    AI may very well be able to read X rays and EKGs better than humans, but if I had a dime for every time "something" about they way a patient looked or the tome of their voice prompted me to ask the question that made the diagnosis I would be very rich.

    Patients want human beings taking care of them, not a machine. AI may eventually stop disasters with poorly trained RNNPs ( like the one after an online course and 500 hours of preceptor-ship sent a patient with a blood sugar of 600 home to die per Bloomberg's report) but that will not solve the issues in medicine today- too much corporate profit.

    Will only make them worse
  • FD1000
    edited December 11
    Indexing has done most of the lifting already. More investors accepted it. Most are not interested in fast trading because it doesn't produce better results.
    Sure, the top 1% have a chance to use AI correctly. They are going to need a lot of brain and machine power.
  • Old_JoeOld_Joe
    @FD1000- Thank you for once again reminding us what a fantastic genius you are. With all of the other important stuff that we have to keep track of we sometimes lose track of what you and your fantastic genius are up to. Hoping to hear more about you and your fantastic genius in the future!
  • hank
    edited December 12
    Seems to be having a beneficial effect on inflation. Bloomberg today headlined lower portfolio management costs as a reason today’s PPI wasn’t higher.


    ”US Producer Prices Jump on Eggs But Fed Categories Come in Soft”

    • PPI increased in November by most in five months on food costs
    • Drops in portfolio management, airfares suggest soft PCE index

    (Above excerpted from Bloomberg Media LLC website on 12/12/2024)
  • yogibearbullyogibearbull
    CPI +2.7%, core +3.3%, both in line with expectations (Wednesday)
    PPI +3.0%, core +3.5%, both above expectations (today, Thursday)

    Remember that PPI is based on wholesale prices and when those are rising, the next CPI readings (based on retail prices) may also be higher.

    This doesn't look good for progress towards +2% average inflation target.

    The next PCE release is on Friday, 12/20/24, a couple of days AFTER the FOMC Statement & Powell's presser next week.
  • hank
    edited December 12
    Thanks @yogibearbull for providing the PPI details. Agree it doesn’t look promising for inflation. I mentioned the ”lower portfolio management costs” somewhat in jest. Reminds me of how lower prices for calculators, computers, flat-screen TVs reduced the official inflation rate over the decades.

    Skeptics often argued: ”You can’t eat computer chips.”
  • hank
    Old_Joe said:

    @FD1000- Thank you for once again reminding us what a fantastic genius you are. With all of the other important stuff that we have to keep track of we sometimes lose track of what you and your fantastic genius are up to. Hoping to hear more about you and your fantastic genius in the future!

    It’s hard to be humble.

  • gman57
    One thing being overlooked... ALL that data being fed into AI is human produced. SO.... if the auditors, financial people at companies make mistakes or enter garbage data to make the last quarter look good, last year look good, current conditions look good as often happens what is AI going to tell you? Maybe over the long term but short term it's still garbage in garbage out.
Sign In or Register to comment.