Alternative Data Weekly #276
Theme: stop collecting data, start understanding systems.
Special thanks to our sponsor Sequentum.
Sequentum’s Web Scraping Agents: AI-Augmented Creation, Web Unblocker and Built-in Pagination with Visual Editor for Atomic Customization – Sign up for your FREE TRIAL of Sequentum Cloud.
QUOTES
“…we’re expanding from a world where data systems primarily serve dashboards and analysts, toward one where they will also power intelligent, autonomous systems operating in real time.” – Virna Sekuj, The Postmodern Data Stack: Where AI and Data Become One
News
Pods
Charts
Final Thoughts (o11y = observability)
#1 – Joe Reis published Parkinson’s Law and AI: Does AI Mean...More Work? January 2026.
My Take: I think AI will create more jobs than there are humans to do the jobs. Investment bankers working 80-hours-per-week are not going to start working 30-hours-per-week if they are given new productivity tools. They will “allocate it (JF: newfound free time) to new work or higher-scoped projects.”
I am recognizing the value of workflows and “the amount of tacit knowledge locked up in workers’ heads is likely an existential impediment to the effective deployment of AI agents in the real world.”
I would be very curious to see the minute-by-minute (second-by-second?) workflow of the most productive people in the world. Have AI train me in those habits. Have AI replicate that person’s productivity for me and everyone in my company.
This would all be driven by data inputs. Have AI watch every move you make through the course of a week and tell you how to do everything better. Data collection is a hassle and could be creepy, but if the value is obvious, people will want this data collected.
#2 – Seattle Data Guy published 5 Key Predictions for the Data Industry in 2026. January 2026.
My Take: My favorite was #3. Please use “AI foundations” as the key descriptor of anything you are trying to sell to senior management.
Teams might question the value of what has been built. The goal will be solid pipelines that add demonstrable business value. Make the value obvious to all.
The Predictions:
Microsoft Fabric Will Rebrand..Again.
1% Of Companies Will Continue To Cry For AI While The Other 99% Are Still Trying To Export ERP Outputs To Excel.
Modern Data Stacks Will Be Shaken.
AI POCs Will Start To Build Actual Foundations.
Snowflake Will Rediscover Themselves.
#3 – Razvan Cicu and Giovanni Pepe published From Monitoring to Observability: Our Ultra-Marathon to a Cloud-Native Platform. January 2026.
My Take: I am tracking this conceptual shift from “monitoring” to “observability” … as systems scale, raw telemetry gets overwhelming. The opportunity is preserving statistical structure and relationships over time, so correlation comes from comparing system states, not replaying history (raw data overwhelms current systems).
Said differently, it becomes about watching how your system is working rather than having your raw data send alerts off static triggers. Related: The Postmodern Data Stack: Where AI and Data Become One
BONUS: Matthew Bernath published 2026 Data Monetisation Trends. January 2026. “The winners in 2026 will not be those who run the most data projects, but those who build compounding data ecosystems.”
What else I am reading:
Shane Evans of Zyte published The age of fast forward web data. January 2026.
Dan Entrup published Data Influencers. January 2026.
Sarah McKenna published 6 Core Principles of Web Scraping. January 2026.
Dan Goldberg published Prediction Markets Data Monetization. January 2026.
Scott Hamilton published The Big Blur. January 2026.
Ravenpack published AI in Finance 2026: The Autonomy Era. January 2026.
Asymmetrix published Anthropic’s legal plugin wipes off $300bn in value but investors miss the point (also, everything is a data business). February 2026.
Source: Richie Cotton of Datacamp’s Dataframed podcast published Our Data Trends & Predictions for 2026 with DataCamp’s CEO & COO, Jonathan Cornelissen & Martijn Theuwissen. January 2026.
My Take: I like these predictions. Most interesting to me is the AI app builders seeing tailwind. From a data perspective, we need to find a way to connect more and better data to drive these apps.
Prediction 1 - junior hiring crisis gets better. Senior people need to retool. Being AI native is a huge benefit.
Prediction 2 - personal shopper by end of year.
Prediction 3 - everyone gets their own tutor. High agency people already have one.
Prediction 4 - new “GPT 3 moment”. Meaning a big leap in tech. New models on better infrastructure. This will be the unlock. This increase in value will be obvious to everyone. Grok will be the big beneficiary.
Prediction 5 - AI app builders’ subtle shift toward app builders going mainstream. Interesting comment that user growth is flat, but revenue is exploding. So a few people are seeing a ton of benefit. Lovable / Replit benefit.
Prediction 6 - quality as a differentiator. AI slop reduced.
Prediction 7 - successful open weights model in 2026.
Prediction 8 - Open AI social login. Bring your own AI to applications.
“Pushing in more data will lead to better models” minute 22:00
Highlights (51-minute run time)
Minute 01:00 - interview starts
Minute 02:00 - junior hiring crisis gets better for juniors, worse for veteran employees
Minute 06:00 - personal shopper by the end of the year
Minute 17:00 - everyone gets their own tutor
Minute 21:00 - better hardware leads to “GPT3 moment.”
Minute 29:00 - AI app builders get momentum.
Minute 35:15 - quality is the new differentiator
Minute 39:00 - new successful open weights model.
Minute 44:00 - Open AI will launch social login.
BONUS: James Worthington and Eric Evans of Selling Signals podcast interview Ed Lavery: Selling Alternative Data Across Funds and Corporates. January 2026.
Source: BCG published As AI Investments Surge, CEOs Take the Lead. January 2026.
Jessica Apotheker, Sylvain Duranton, Vladimir Lukic, Nicolas de Bellefonds, and Christoph Schweizer.
BONUS: Dan Goldberg published Prediction Markets Data Monetization. January 2026.
o11y - observability
This is what I am trying to communicate to the world.
Imagine If You Could…
Analyze all your telemetry (real-time system behavior data) without exploding costs.
No dropping metrics. No aggressive sampling. No blind spots created by budget constraints.
CONTROL COSTS
Detect change before it becomes an incident.
Streaming anomaly detection, structural drift, and multivariate correlation that reveal how systems are actually changing, not just when a threshold breaks.
EARLIER NOTIFICATION
Cut through alert noise and focus on what truly matters.
Behavior-aware intelligence that reduces false alerts by understanding normal system behavior, not reacting to every spike.
REDUCE ALERT FATIGUE
Compare today’s system behavior to last month, or last year, on demand.
True long-horizon analysis and “time travel” comparisons without storing or replaying raw telemetry.
COMPARE ANY TIME FRAME
Understand system behavior across distributed, hybrid, and edge environments, without centralizing data.
Learn from telemetry generated across clouds, regions, and on-prem or edge systems while keeping data local, private, and lightweight.
SEE THE WHOLE SYSTEM
Do all of this with flat, predictable pricing and a simple plug-in.
No volume-based surprises. No extended retention tax. No cost explosion as systems scale. No rip & replace.
COST TRANSPARENCY
This is what we do at SymetryML.
john@symetryML.net