Alternative Data Weekly #128
Thanks for being here!
Things to check out:
Our transaction data is now available at more granular “row level”. 4+ years of history. Daily data.
90 West is re-launching Cross Country Insights.
If you care about RBLX (the stock), you should be checking out our Gift Card Data.
Contact me to learn more.
Theme that emerged in this week’s email is … will we run out of data?
QUOTES
“There are no “data” best practices. The future of the data space is completely unpredictable.” - Sven Balnojan
“Obtaining rich proprietary training datasets will be the key challenge for startups looking to create AI models for industry verticals.“ - David Sacks
News Articles
Podcasts
Cool Charts
Final Thoughts (Delayed Gratification)
#1 – David Sacks published The Give-to-Get Model for AI Startups. March 2023.
My Take: This really drives home the point that good, large data is essential to developing the best AI models. Quote: “In many industry verticals, obtaining a rich proprietary dataset will be the key challenge to producing a differentiated AI model.” Value will flow to the data owners.
#2 – Sven Bolnojan published Here’s why you need to change your work with data, at your whole company; Thoughtful Friday #28. March 2023.
My Take:TL;DR: It’s not too late to start working with data. The tools for working with data are better than ever. The article provides a helpful framework for how to think about transforming your organization to use data more effectively. The most interesting to me was the focus on data products. Moving from “service oriented” to “product oriented” is much more efficient. Most data teams today are scrambling to answer ad hoc questions from business leaders. The key is to anticipate key important levers & build products to deliver the key pieces of information in an ongoing fashion (“productize your data”).
#3 – Sean Cao, Wei Jiang, Junbo Wang and Baozhong Yang published in IR Magazine From man versus machine to man plus machine: The art and AI of stock analysis. March 2023.
My Take: A rare article that focuses on the positives AI will deliver for humanity. One of the many finding in the study is that post-alternative data availability, analysts covering affected companies improve their performance relative to the AI-only forecast model we build. Humans can still add value! We ain’t done yet!
BONUS: Wesley Bray from The Trade published Two thirds of investors believe alt data adds alpha despite growing demands for its application to be improved. March 2023. “Data from Coalition Greenwich shows that alternative data continues to grow rapidly with its perceived value increasing, while firms work towards improving how best to utilise these alt data sets.”
BONUS 2: Revelate published Quick Facts Data Monetization Banking Experts Should Know. February 2023.
BONUS 3: April Miller from Inside Big Data published How AI Helps Prevent Human Error In Data Analytics. March 2023.
BONUS: Check out Data Is Plural (The Podcast), the new podcast from the creator of one of my favorite weekly emails: Data is Plural.
BONUS 2: Check out a new podcast from Omri Shtayer & Evan Schnidman: I See Data People.
#1 – Tobias Macey interviews Matt Turck Mapping The Data Infrastructure Landscape As A Venture Capitalist. April 2023.
My Take: Check out Matt Turck’s MAD Landscape. It is good to think about the landscape & ecosystem from the perspective of an investor. Most interesting to me was the idea that the leaders in data (Airbnb, LinkedIn, etc) are years ahead of the common problems in the market…pay attention to them, but don’t move too early. Also, remembering that you can be right on the analysis and wrong on the outcome.
Highlights (61-minute run time):
Minute 01:00 – Matt Turck intro
Minute 03:00 – Discussion of MAD landscape (Machine Learning, Artificial Intelligence & Data)
Minute 12:00 – shift away from Hadoop (end of era) to modern data stack (next big phase)
Minute 14:00 – how Matt thinks about categorization
Minute 17:00 – impact of VC and investment on the evolution of data landscape (Snowflake IPO)
Minute 22:30 – impact of tighter capital & attention being paid to AI
Minute 29:30 – interplay between data infrastructure & AI development
Minute 33:00 – categories ripe for consolidation & which will remain stand-alone
Minute 38:45 – impact on “de facto architecture”
Minute 45:45 – unexpected entrants into the market
Minute 49:15 – most interesting lesson Matt’s learned personally
Minute 58:00 – Matt’s perspective on what he sees as the biggest gap in the tooling space today
Source: Matt Turck
Source: I came across this idea from this podcast: Exhaustion of High-Quality Data Could Slow Down AI Progress in Coming Decades. March 2023.
Related: Pablo Villalobos published Scaling Laws Literature Review (citation)
Bottom line: there is a relation between the size of a model/dataset and it’s performance. This analysis measures how well a model will work as the data gets bigger or smaller.
This is in line with the theme for this week’s email that, despite the vast amount of data being created every day, there is concern among the AI community that we will run out of good data to train all these models.
This chart:
The ability to delay gratification is an underappreciated skill. A skill that increases the odds you will create a sustainably happy life.
In Jawad Mian’s book Stray Reflections, his passage from January 2016 cites French writer & philosopher Pascal Bruckner.
"Credit changed everything. Frustration became intolerable & satisfaction normal. To do without seemed absurd."
Pascal Bruckner
I laugh at that quote because it reminds me of comedian Louis CK (1:00 minute mark) who said, when I was young “when we ran out of money we just stopped doing things”.
This is no longer the case. The Bruckner quote highlights the way in which credit has played a big role in making instant gratification the norm. It is increasingly difficult to go without the things we want. This is a fundamental change.
"To enjoy was once forbidden, from now on, it's obligatory."
Pascal Bruckner
In addition to credit, we now have a host of technologies that deliver to us everything immediately. This further ingrains in us that idea that instant gratification is the expectation. Pain becomes intolerable. Pleasure obligatory. Immediate results become a habit. This impacts our decision-making process & impulsive decisions are less likely to be good decisions.
The Stanford marshmallow experiment is most famous among the multitude of studies to conclude that ability to delay gratification is a key to success.
This is a skill you can develop & improve.
“Learning to delay the impulse for immediate pleasure in favor of long-term satisfaction is a skill required for incremental and long-term growth.”
Call it suffering, call it fasting, call it patience. Whatever tool you use to practice patience, know that you are developing a skill that will serve you well.