> > > Tying this thread with a previous one on retirement financial planning, by > silklister Josey John: > > http://factordaily.com/ai-big-data-machine-learning-funds-fintech/ > > I'd be interested in thoughts from folks like Shyam Sunder and Deepak > Shenoy (and others too, of course) > > Udhay >
Sorry, saw this late. In that article: 1) medallion is only open to Rentech current and ex employees. They hare 5 and 45 (5% mgmt fee and 45% of profits) and are insanely profitable. 2) Accura's growth of 185% in four years on small caps is not a big deal. Reliacne Small cap fund - one of the funds that does only small cap investing - has made a ludicrous 216% in the same time. ( https://www.valueresearchonline.com/funds/fundperformance.asp?schemecode=16182) So has DSP Micro and Small cap fund, which is also a very famous one in the space. Effectively, if this fund was AI, it has been beat by two non-algo fund managers by a fairly large margin. 3) REntech and Accura couldn't be more different. Medallion works on ultra short term trades for the most part. Their other fund - which is open to external investors - has lower performance and that might have a few stocks held. But Accura essentially holds for the longish term and deals with illiquid stuff. 4) I think much of what Accura does is not AI at all. You just can't get enough info about smallcaps. 5) Gupta and Nagpal must be from seriously rich families. And they seem to have 400 cr. iunder "advisory" for this money, and only the rest is in portfolio management (about 350 cr.) Yes of course fund managers can be replaced by an algo...I have personally built stuff like this :) WE deployed stuff that I think continues to make money...on Indian markets. Most of the forex trading that happens worldwide is between computers. I can guarantee that most trades will be algo-executed at some point in the future. I also think that fund management will be an excellent money making opportunity because of the ludicrous amount of information that's not easy for computers to figure out - and they'll never figure it out, IMHO. Markets have forever been inefficient and while they are, there will be ways to beat them consistently. But this Accura thing is not anywhere close to this kind of algo, IMHO. There's more to this story but you can't refute what the founders say, though I think their "success" is not phenomenal and isn't reallly related to AI. I think they're very smart folks. Cheers, Deepak