Are you at the risk of being sued by companies that you give a "fraudulent" score to (correctly or incorrectly), maybe a class action? Is this one of the reasons why there is no public version of your tool?
Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
81–90 of 109 posts
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#82Congrats on getting out into this space. I suspect you will be very successful.
Have you considered training a model to use the quarterly reports to identify companies likely to see a significant upside or downside over the next N months? Seems like it would be a very straightforward task and could pick out some real diamonds from the rough. Not sure how well it would work though... Could be a very simple way to chase returns.
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#83Have you read Eliot Peper's novel Uncommon Stock? It is about a startup doing automated analysis of accounting information. "I'm a CPA and a developer" immediately reminded me of the book.
I must say the real thing is so much cooler than the story, even if the story itself is also very cool.
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#84I love symmetries. Is the reverse of this also true? Could there be disclosures missed by investors that if unearthed would positively impact the company?
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#85Out of interest - why didn't you start a hedge fund instead?
I think big part of a strategy like this is long term. If you want to short a stock you need to make sure the fraud is discovered by other investors. So you probably need to have a strong marketing department that basically thrashes other companies, this may not be a goal a lot of people aspire to. With a short you're also exposed to lose an infinite amount. And finally "markets can stay irrational longer than you ca…
What you are saying is true, but it's a little bit superficial. There are many successful long/short funds and even a few short only funds. The risks are largely manageable risks.
The upside to starting a hedge fund is astronomical. And if what they claim they can do they can in fact do, the best business model seems to be start a hedge fund and manage the risks doesn't it?
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#86I've read many SEC filings (mostly 10-Ks, 8-Ks and 4-Qs), and I led a project to develop an UIMA-based analysis platform for them.
What is shocking is the lack of diligence that companies can get away with. The "risk" section (§6 of Form 10-K) is often token info copy-&-pasted from other filings, notably from competitors, instead of having an entity's actual risks in it.
Good luck with the business!
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#87I've read many SEC filings (mostly 10-Ks, 8-Ks and 4-Qs), and I led a project to develop an UIMA-based analysis platform for them. What is shocking is the lack of diligence that companies can get away with. The "risk" section (§6 of Form 10-K) is often token info copy-&-pasted from other filings, notably from competitors, instead of having an entity's actual risks in it. Good luck with the business!
Thanks! Yes, we keep track of risk factors that are just copy-pasted across companies. There is so much of it!
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#88It seems to me all you have to do to fool a ML model is rephrase sentences and just use verbiage it has never seen before. So why isn't your tool just the latest item on a pre-IPO company's checklist that is looking to commit fraud to scrub against? Also if Muddy Waters can do something similar to this (if not better?) doesn't that mean you should just hire them and throw away your ML model?
It isn't that easy to reverse engineer our models. We test our models for a certain degree of robustness (same concepts expressed in different ways will be generally picked up). Plus, if companies go out of their way to express risks in cryptic terms, our models picks up needlessly convoluted text as well. We also constantly retrain our models so I don't foresee this becoming a problem in the foreseeable future.
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#89I love symmetries. Is the reverse of this also true? Could there be disclosures missed by investors that if unearthed would positively impact the company?
Yes! That is something we are looking forward to working on in the future.
Re: Launch HN: Bedrock AI (YC S21) – Using ML to identify red flags in SEC filings
#90Out of interest - why didn't you start a hedge fund instead?
I think big part of a strategy like this is long term. If you want to short a stock you need to make sure the fraud is discovered by other investors. So you probably need to have a strong marketing department that basically thrashes other companies, this may not be a goal a lot of people aspire to. With a short you're also exposed to lose an infinite amount. And finally "markets can stay irrational longer than you ca…
My guess is they also have plans to move in on other sorts of fraud detection, for example bond issues or privately held firms. That’s my good faith interpretation.
My bad faith interpretation is they don’t trust their algos enough to trade on it themselves.