Alternative Data Weekly #256

Theme: will there be enough humans to do all this work AI is creating for us?

Sep 19, 2025

Special thanks to our sponsor, StockTwits.

For more information, please contact Michiel Maaskant. Related: The StockTwits deal with Polymarket was announced this week.

QUOTES

“To unlock data’s full value, it must function like a utility: flowing seamlessly, priced transparently and available on demand.“ – Kirk McKeown, Carbon Arc

News
Pods
Charts
Final Thoughts (good advice)

#1 – Caitlin Moorman published AI is changing how our data team works, and that’s great. September 2025.

My Take: Implementing AI analytics will help data teams, but we still need to ask the right questions at the right time for the business. Human translation and judgment are difficult to replicate & fully trust. The key is to focus on seamlessly turning information into intelligence that will help the business accelerate forward.

“this (AI Analytics) is a technology that’s genuinely helping us get closer to the goal of helping our organizations make better, faster decisions with data. It will, however, require us to evolve how we work and to think differently about where we add value.”

#2 – Kirk McKeown published From Electricity And Oil To Data: Why Every New Age Demands A Liquid Commodity. September 2025.

My Take: Selling data is hard. Making data trade more like a commodity will expand access and likely have a Jevon’s Paradox effect. Total consumption (use of data) will vastly increase once data can be traded like a liquid commodity.
Efficiency increases → cost falls → demand rises → total consumption increases.

#3 – DataActionMentor published The only AI analyst setup that actually works. September 2025.

My Take: Nuance matters. Many will come out and talk about extremes, as that can drive attention (“agents will replace all data analysts”). But like just about everything, there is nuance involved.
I am a believer that AI will likely change the day-to-day of many jobs, but there might not be enough humans to do all the work. People with domain knowledge and a working understanding of both the business & the information needed to run the business, will be in high demand.
Add in the ability to communicate and work with people, and you’ll be an important part of most organizations.

BONUS: Seeking Alpha published President Trump calls to do away with quarterly earnings reports. September 2025. h/t Daniel Goldberg.

My Take: This is good for Alternative Data sources, right?

What else am I reading:

Matthew Bernath of Alternata is seeking test users for the Alternata Exchange. September 2025 (related LinkedIn article from Alternata here).

Pedro Aguayo published Financial Agents…. Dawn of a New Era. September 2025.

ChatGPT is Eating the World published Shocker: Judge Alsup “postpones” preliminary approval of class settlement in Bartz v. Anthropic. Updated with order. September 2025.

Sydney Nielsen published Introducing Agent Observability: Making AI Work in the Real World. September 2025.

David Brunner published What does AI really mean for the future of knowledge work?. September 2025.

Dan Goldberg published Private Company Alternative Data. September 2025.

Peter Baumann published Discussing Master Data Management. September 2025.

Hugh O’Connor published The 20% Rule: How Small Levers Drive Big ROI in Data Strategy. August 2025.

Alex Izydorczyk published Why do Startups Power Perplexity Finance?. August 2025.

McKinsey’s Financial data and markets infrastructure: Positioning for the future. January 2025.

Source: Ben Lorica of The Data Exchange Podcast interviews Josh Pantony, CEO of BoostedAI.

My Take: Alfa is BoostedAI’s flagship product. This was very interesting, as what they have built is similar to what (my former employer) ModuleQ was attempting to build. I remain confident that this is how knowledge professionals will engage with data in the future.
This idea that your AI assistant will anticipate what you want is interesting. Similar to social media, you show up, it knows who you are, and it will be constantly suggesting things that are of interest to you.
The big key is the ability to trust the answer. They discuss “auditability” and the importance of being able to recreate the information that was presented.
Lastly, I thought the multimodality was interesting. Increasingly, important information lives in non-traditional, non-text formats (video, podcast, etc). How to capture all of this? Humans just don’t have the capacity.
Note: The Spotify podcast episode is 40 minutes, while the YouTube version is 48 minutes. I’m not sure why there is a difference.

Highlights (40-minute run time)

Minute 00:45 – intro to Alfa (BoostedAI’s product).
Minute 02:45 – building what’s next.
Minute 06:00 – trust the output? “Auditability”.
Minute 08:30 – multimodality; reasoning.
Minute 12:30 – context beyond the math … get to know the person.
Minute 15:45 – buyside vs sell side (primarily fundamental buyside users).
Minute 21:00 – discussion of proprietary models.
Minute 24:00 – aligning user expectations.
Minute 26:00 – philosophy on data (what happens when you become reliant on non-reliable data).
Minute 28:00 – Bloomberg’s moat.
Minute 34:00 – vision of AI’s evolution.
Minute 37:00 – the power of an AI companion.
Minute 38:00 – other industries where Alfa is applicable?