Shortly fund managers and trading firms will need to adopt AI in order to maintain their edge. The question will be how it will fit into the overall set of tools used to perform trading, investing, and planning/executing strategies The issue at this point is that most of the AI market investing studies are based on back-testing. How much of this is effectively curve-fitting performed by the AI? How will the AI perform in actual market forward-testing with live information? Stanford's 'Terminator' AI Just Beat 93% Of Human Fund Managers Using Just Public Data, Professor Warns These Entry-Level Number Crunchers Risk Extinction https://www.benzinga.com/markets/eq...managers-using-just-public-data-professor-war Stanford's "Terminator" AI fund manager beat most Wall Street pros and now threatens junior analysts' jobs, researchers say. What Happened: The bot, built by Stanford Graduate School of Business professor Ed deHaan and Boston College colleagues, retro-tested more than 3,300 diversified mutual-fund portfolios from 1990 to 2020 and outperformed 93% of their human managers after adjusting each fund every quarter. Researchers fed the algorithm only public data — financial statements, analyst forecasts and daily prices—mirroring what real managers saw in the pre-Internet era. Using what IBM defines as a “random-forest” model, the AI repeatedly split the data into new patterns, swapped out risky positions and steered more cash into broad index funds, a move that boosted risk-adjusted returns across nearly every test case. The bot never relied on "secret signals," deHaan told Fortune. Instead, it uncovered "information hiding in plain sight." DeHaan warns that the real losers are entry-level number-crunchers. "Sitting around, crunching Excel spreadsheets is a job that will barely exist in five years," he said, arguing that automation can do a week's work in hours. That said, it is worth adding that the study is a "thought experiment" and not live trading. The bot never faced rival algorithms or the compliance limits that constrain real funds. DeHaan says its edge would shrink once the technology spreads, echoing other research showing AI gains erode when everyone uses the same signals. Why It Matters: The Stanford study suggests funds would have needed to quintuple fees to match the AI's performance, highlighting the pressure on managers to adopt smarter tools or vanish. Or as DeHaan puts it, the shake-out could create a market that favors the "clever human who thinks like a human and can ‘out-human' the AI.” Goldman Sachs believes AI automation could disrupt up to 300 million jobs worldwide, a scenario that would land hardest on entry-level finance roles. Early adopters contend the shift is already underway, with trading desks integrating specialist models for analysis, risk checks and execution, shaving days of work to mere minutes. That backdrop is fueling a boom in robo-advisors and other low-cost digital platforms, while a separate report shows AI systems are already parsing market data with great speed and accuracy.
I'll make it big so you get a feel for what I am say... THEY DID NOT GO UP AGAINST MARKET MAKERS!!! Yes, jobs with be displaced...No question about it. But their data is flawed... Apples and oranges... PS "I have run (tested) my program and I have a huge win rate"... Cool..."Have you run it live"??
That doesn't say much...historically, fund managers and stock traders, pickers, in general have performed rather shittily. Financial statements, analyst forecasts and daily prices can only semi predict the future vaguely. There's still alot of market ambiguity and noise to read through. I'm guessing that AI investor robot only barely marginally beat the average, maybe just out of sheer luck even. Or editing the hindsight conditions to make it shine a bit
People are going to lose so much money trusting AI BS and others will make a fortune tricking it. For example: I reasonably smart person might notice a correlation between X and Y events, the might find in interesting and try to study whether the two are really linked. The AI will start treating that correlation as a law of nature and will lose horribly one the rug gets pulled from under it. I found this quote to be a particularly moronic way of looking at things. Since they're backtesting against real funds those funds obviously were willing to operate as they were. There's no "need" to raise fees. The notion that raising fees would just let them earn more money is dumb. If they charged 4x fees the would have virtually no business. Their competitors would get all the customers. That huge number of portfolios makes this much LESS impressive. Their numbers mean that 231 human managed portfolios still beat the AI. How many people are out there going "I don't want to make too much money, I'm going to find the 2000th best fund manager an put my money with that guy? The reality is those 3300 funds had different prospectus and investment goals. It sounds like they either: They just ignored the stated goals of the funds and put the same AI in charge of all of them. (Unlikely) The made unique AIs that actually tried to follow the public offering documents so the funds wouldn't be sued. Which means they could train each one uniquely against the known performance they were trying to beat. Finally the idea that large scale professional trading would really only operate on backward looking, widely disseminated, easy to get information is stupid. It's literally the notion that "past results guarantee future returns". I have to believe that at least a few real fund managers actually pick up the phone and call people with questions. The might even leave their offices and go look at the actual operations of a company they're about to invest the GDP of a small country in.
TLDR: technology is helpful but you'll need to know how to use it. This shouldn't be a surprise to anyone.
My largest holdings are in three of these close end funds... For them to invest in their companies, they need to meet with upper management and have them open "all" their books. https://www.royceinvest.com/ PS I walked into the local coffee shop. I ask the waitresses if they would give up their Apple phones or watches...A strong "NO". Common sense to invest in Apple...My own simple survey. No AI...Eyes on the ground.