top of page
Writer's pictureKris Bennatti, CEO of Hudson Labs

Hudson Labs in 2022

Timeline View, Blackline/Diff, AI-generated News Feeds, and so much more

Welcome to 2023. The year 2022 was an eventful one for Hudson Labs. The popularisation of ChatGPT in 2022 meant that new customers suddenly understood the powerful technology underlying our product: large language models. The collapse of FTX and the bear market served as a necessary reminder

of the importance of risk management and early fraud detection.

This past year was the first full year of our product’s launch, and we could not be more proud of what we’ve accomplished. It has been a pleasure to have you all along for the ride.

Highlights include:

  • New features including the Timeline View, Blackline/Diff, AI-generated News Feeds, and so much more

  • Machine learning breakthroughs leading to better models, better content and better predictions

  • Notable new clients, including some of our personal heroes

  • Hudson Labs analysis featured in the WSJ, Bloomberg, Financial Times, and Reuters

  • Conferences, speaking engagements, and team milestones


A few of our many new features released in 2022…

Automated News Feeds


We launched the world's first AI-generated news feed, where our models comb through SEC filings in real-time and surface price-moving, decision relevant news. The news feed is powered by our updated models.


The news feed allows customers to monitor the highest impact events in their area of focus. Most of the news we surface doesn’t make headlines in traditional financial media, because just like equity analysts, journalists aren’t reading every SEC filing as it gets published. For example, this year we broke the news about the Moderna CFO getting suspended after just a few days on the job.



Bedrock AI's automated news feed


Blacklining/Diff/Document compare


This year we added blacklining also known as ‘diff’ or document compare. This feature allows you to easily compare a company’s annual and quarterly reports to previously filed versions. Customers use this functionality to discover emerging risks and opportunities.


Blacklining(document compare) feature


In addition to blacklining, we also show exactly when red flags were first disclosed, if they are not new. We compare current disclosure to all of the company’s previous filings, not just the prior quarter so that you know exactly when information first became public.


Example of  when red flags were first disclosed.


The Timeline View


Timeline View showcases our new, more powerful red flag ranking models. The Timeline View uses a smart algorithm to order red flags and present them in a collapsible format, reducing information overload. Easily filter for red flags based on red flag categories like governance, accounting policy, and credit risk.


Bedrock AI timeline view




Coverage expansion


In 2022 we added coverage of registration statements (prospectuses), SEC comment letters (extracted and processed from the original PDF), non-timely filing forms, and earnings transcripts.


Timeline view w SEC comment letters and non-timely filings


Notable New Clients


In the past year, we proudly gained a number of notable clients and a few famous (infamous?) customers…none of whom we can disclose here due to the confidentiality clause in our terms of service. Be happy for us anyway.



News and Media Features


Our analytics caught the eye of some of our favourite financial journalists. It was an honour to see our name in such venerable publications as the Wall Street Journal, the Financial Times, Bloomberg, and Reuters.


Crypto Coverage


When the crypto markets collapsed, our team looked into publicly traded crypto mining firms and discovered that accounting red flags were not unique to FTX.


Our research indicated that 84% of identified crypto mining firms had internal control failures. This analysis was featured in the Wall Street Journal article “Accounting Red Flags Are Common Among Public Crypto Companies” by Jean Eaglesham and Ben Foldy.


WSJ article regarding Crypto firms

None of these publicly traded crypto mining firms were audited by a Big Four accounting firm. This was featured in "SEC Heightening Scrutiny of Auditors’ Crypto Work" by Jean Eaglesham of the Wall Street Journal.


SPAC Coverage


Throughout 2022, our team researched SPACs and de-SPACs and the related accounting and disclosure trends.


Approximately 40% of de-SPAC filings since 2020 disclosed serious cash flow issues, which was featured in "Why More Than 40% of Ex-SPACs Are Running Out of Cash" by Bailey Lipschultz of Bloomberg.


 Bedrock in Bloomberg regarding SPACS


Our Co-Founder and CEO, Kris Bennatti, also added colour commentary to our SPAC and de-SPAC findings, which were included in the above-mentioned Financial Times article as well as another "SPACs Trip Up on Accounting Requirements, Reporting Deadlines" by Nicola M. White.


Other coverage


Approximately 60% of 10-Ks and 10-Qs are filed between 4 and 5:30pm. When companies deviate from normal filing schedules, it often means big news. This finding was featured in Herb Greenberg’s article at Empire Financial Research.


We were featured in '“Lifetime value” is Silicon Valley’s next buzzword' by Karen Kwok at Reuters. We observed that the phrase “lifetime value” or “LTV” was mentioned by executives during their earnings calls 47x more frequently in late 2022 than in early 2019.


We made the Bear Cave Christmas List at the very beginning of 2022 which was a great way to start the year. Thank you, Edwin.



Conferences


In July, our CEO Kris Bennatti spoke on a panel at UC Berkeley’s Fraud Fest 2022 with Dan Taylor from Wharton and Hane Kim from the SEC. Kris spoke about forensics analytics and how Hudson Labs leverages machine learning to identify hard-to-find qualitative red flags.



 UC Berkeley's Fraud Fest 2022

In September, Hudson Labs co-hosted a well attended investing mini-conference with Vision Research in New York City where we first shared our research on the quantification of qualitative signals and the impact on price.


Our CTO, Suhas Pai was the co-chair of the Privacy working group of Big Science, which published the world’s largest open source multilingual language model (larger than GPT-3) called BLOOM. His work at Big Science was featured in 2 publications including 1 at NeurIPS, a top machine learning conference.


In September, Suhas chaired the Toronto Machine Learning Summit (TMLS) on NLP. He was also the Chair of the Toronto Machine Learning Summit’s 6th Annual Conference & Expo 2022. Suhas also recently spoke about ChatGPT with Aggregate Intellect.


 Suhas Pai at the Toronto Machine Learning Summit which he chaired..


Popular Hudson Labs newsletters in 2022:

Product features upcoming in 2023

  • Impact assessments: Ever wonder how many small cap technology companies report material control weaknesses and whether it correlates with ROI over the next 6 months? Soon you’ll never have to wonder again.

  • Boilerplate highlighting: Save time by safely ignoring text that repeats or rephrases information that isn’t company-specific.

  • Earnings transcript coverage: Easily identify core business drivers with AI generated topic guides.

And so, so much more! Stay tuned for the results of our game-changing new research breakthroughs.



Thank you


We couldn’t have achieved what we did without the help of our incredible investors, advisors, customers, employees and cheerleaders. Thank you to everyone who offered advice, feedback and words of encouragement over the past year. We are deeply and genuinely grateful. While there are too many people to thank, we would like to highlight a few individuals who’ve been instrumental in our story.

Canaccord Genuity: We were accepted into CG’s advisory program this year. Thank you to David Kassie, Aravinda Galappatthige, and Kelly Bourque for your wisdom, support and effort in helping push us towards our goals.

Michael Seibel and Y Combinator: Thank you for saving us from some potentially disastrous decisions.

Our favourite customers (you know who you are): Thank you for your witticisms, feedback, and for never taking anything Kris says very seriously. We should be paying you…just kidding…very much just kidding.

Xiao, Ray, Di, Andre, Sheida, Laura and Fran: We’re honoured to get to work with you every day. Hudson Labs owes our success to your commitment, ingenuity, and hard work.

FinTwit Thought Leaders: Thank you to Michelle Leder, Edwin Dorsey, Dan Taylor and many others for your support and guidance this year.


Comments


bottom of page