Aleksander Molak
@alxndrmlk.bsky.social
"The Causal Guy" http://causalpython.io
Author || Advisor || Educator
Host at http://CausalBanditsPodcast.com
Causal ML Tutor @ Uni of Oxford
CausalSky: https://bsky.app/profile/did:plc:imz3rf35poonl7yxt7bogui4/feed/aaamrclcu3tfa
Author || Advisor || Educator
Host at http://CausalBanditsPodcast.com
Causal ML Tutor @ Uni of Oxford
CausalSky: https://bsky.app/profile/did:plc:imz3rf35poonl7yxt7bogui4/feed/aaamrclcu3tfa
I'm curious about the original study.
Does it really show "cancer prediction 5 years in advance"?
Btw., I think there's a great potential in ML in medicine, but reality is often more complex than scenarios typically considered in theory.
Does it really show "cancer prediction 5 years in advance"?
Btw., I think there's a great potential in ML in medicine, but reality is often more complex than scenarios typically considered in theory.
November 30, 2025 at 3:14 PM
I'm curious about the original study.
Does it really show "cancer prediction 5 years in advance"?
Btw., I think there's a great potential in ML in medicine, but reality is often more complex than scenarios typically considered in theory.
Does it really show "cancer prediction 5 years in advance"?
Btw., I think there's a great potential in ML in medicine, but reality is often more complex than scenarios typically considered in theory.
- Black Friday pre-sale of Justin Bélair's "causally aware" intro to biostatistics course
We'll start sending today's issue at 9am PT / 12pm ET / 6pm CET
Register here (FREE): causalpython.io
3/3
We'll start sending today's issue at 9am PT / 12pm ET / 6pm CET
Register here (FREE): causalpython.io
3/3
Causal Python || Your go-to resource for learning about Causality in Python
A page where you can learn about causal inference in Python, causal discovery in Python and causal structure learning in Python. How to causal inference in Python?
causalpython.io
November 30, 2025 at 11:22 AM
- Black Friday pre-sale of Justin Bélair's "causally aware" intro to biostatistics course
We'll start sending today's issue at 9am PT / 12pm ET / 6pm CET
Register here (FREE): causalpython.io
3/3
We'll start sending today's issue at 9am PT / 12pm ET / 6pm CET
Register here (FREE): causalpython.io
3/3
- New notebook for double machine learning with TabPFN (tabular foundation model)
- Alberto D. Horner on a NeurIPS 2025 paper introducing deconfounding generative model (DeCaFlow) by Alejandro Almodóvar et al.
- Causal evaluation of modern AI systems
2/
- Alberto D. Horner on a NeurIPS 2025 paper introducing deconfounding generative model (DeCaFlow) by Alejandro Almodóvar et al.
- Causal evaluation of modern AI systems
2/
November 30, 2025 at 11:22 AM
- New notebook for double machine learning with TabPFN (tabular foundation model)
- Alberto D. Horner on a NeurIPS 2025 paper introducing deconfounding generative model (DeCaFlow) by Alejandro Almodóvar et al.
- Causal evaluation of modern AI systems
2/
- Alberto D. Horner on a NeurIPS 2025 paper introducing deconfounding generative model (DeCaFlow) by Alejandro Almodóvar et al.
- Causal evaluation of modern AI systems
2/
... you can get lifetime access to the course here: https://buy.stripe.com/14A8wQcpVcPWayA8wT7N606
Use DARKSTATS2026 to get our Black Friday discount
We go live on Feb 27, 2026.
And you get a 💯30-Day Money-Back Guarantee if you don't like it.
5/5
Use DARKSTATS2026 to get our Black Friday discount
We go live on Feb 27, 2026.
And you get a 💯30-Day Money-Back Guarantee if you don't like it.
5/5
November 29, 2025 at 10:50 AM
... you can get lifetime access to the course here: https://buy.stripe.com/14A8wQcpVcPWayA8wT7N606
Use DARKSTATS2026 to get our Black Friday discount
We go live on Feb 27, 2026.
And you get a 💯30-Day Money-Back Guarantee if you don't like it.
5/5
Use DARKSTATS2026 to get our Black Friday discount
We go live on Feb 27, 2026.
And you get a 💯30-Day Money-Back Guarantee if you don't like it.
5/5
In our upcoming course on biostatistics, we wanted to show the students the "dark" side as well.
If you want to learn or refresh your statistical knowledge in a way that does not hide the "dark side" from you, and shows you how causality fits in the picture...
4/
If you want to learn or refresh your statistical knowledge in a way that does not hide the "dark side" from you, and shows you how causality fits in the picture...
4/
November 29, 2025 at 10:50 AM
In our upcoming course on biostatistics, we wanted to show the students the "dark" side as well.
If you want to learn or refresh your statistical knowledge in a way that does not hide the "dark side" from you, and shows you how causality fits in the picture...
4/
If you want to learn or refresh your statistical knowledge in a way that does not hide the "dark side" from you, and shows you how causality fits in the picture...
4/
At each of these steps we need judgment and expertise.
In learning this subtle art of modeling, positive examples (how to do something right or correctly) are important, but negative ones (what can go wrong, how to break something) are equally -- or perhaps even more -- important.
3/
In learning this subtle art of modeling, positive examples (how to do something right or correctly) are important, but negative ones (what can go wrong, how to break something) are equally -- or perhaps even more -- important.
3/
November 29, 2025 at 10:50 AM
At each of these steps we need judgment and expertise.
In learning this subtle art of modeling, positive examples (how to do something right or correctly) are important, but negative ones (what can go wrong, how to break something) are equally -- or perhaps even more -- important.
3/
In learning this subtle art of modeling, positive examples (how to do something right or correctly) are important, but negative ones (what can go wrong, how to break something) are equally -- or perhaps even more -- important.
3/
... a density plot can save us, but in fact it typically relies on even stronger assumptions.
In data analysis and in predictive modeling, we make tens or even hundreds of micro-decisions that can impact the outcomes and, subsequently, the decisions we make based on our analyses.
2/
In data analysis and in predictive modeling, we make tens or even hundreds of micro-decisions that can impact the outcomes and, subsequently, the decisions we make based on our analyses.
2/
November 29, 2025 at 10:50 AM
... a density plot can save us, but in fact it typically relies on even stronger assumptions.
In data analysis and in predictive modeling, we make tens or even hundreds of micro-decisions that can impact the outcomes and, subsequently, the decisions we make based on our analyses.
2/
In data analysis and in predictive modeling, we make tens or even hundreds of micro-decisions that can impact the outcomes and, subsequently, the decisions we make based on our analyses.
2/
Get access here: https://buy.stripe.com/14A8wQcpVcPWayA8wT7N606
---
Course release: 27 Feb 2026
Programming languages: R, Python
**A quote from one of our live cohort students
3/3
---
Course release: 27 Feb 2026
Programming languages: R, Python
**A quote from one of our live cohort students
3/3
November 28, 2025 at 10:14 AM
Get access here: https://buy.stripe.com/14A8wQcpVcPWayA8wT7N606
---
Course release: 27 Feb 2026
Programming languages: R, Python
**A quote from one of our live cohort students
3/3
---
Course release: 27 Feb 2026
Programming languages: R, Python
**A quote from one of our live cohort students
3/3
This course is for people who want to learn biostatistics or refresh their statistical knowledge.
- Learn statistics in a causally-aware way
- Learn about real-world challenges
- Learn in a way that is fun *and* deep
Get $36 off with the code ALEXBLK25 (no strings attached)
2/
- Learn statistics in a causally-aware way
- Learn about real-world challenges
- Learn in a way that is fun *and* deep
Get $36 off with the code ALEXBLK25 (no strings attached)
2/
November 28, 2025 at 10:14 AM
This course is for people who want to learn biostatistics or refresh their statistical knowledge.
- Learn statistics in a causally-aware way
- Learn about real-world challenges
- Learn in a way that is fun *and* deep
Get $36 off with the code ALEXBLK25 (no strings attached)
2/
- Learn statistics in a causally-aware way
- Learn about real-world challenges
- Learn in a way that is fun *and* deep
Get $36 off with the code ALEXBLK25 (no strings attached)
2/
If you want to learn or refresh (bio)statistics in a way that goes beyond standard ways of teaching, or you know someone for whom this could be helpful, leave your email here or share it with them: https://decisionacademy.io/
We'll be sending more details on the presale on Friday.
5/5
We'll be sending more details on the presale on Friday.
5/5
November 26, 2025 at 9:37 AM
If you want to learn or refresh (bio)statistics in a way that goes beyond standard ways of teaching, or you know someone for whom this could be helpful, leave your email here or share it with them: https://decisionacademy.io/
We'll be sending more details on the presale on Friday.
5/5
We'll be sending more details on the presale on Friday.
5/5
We'll be releasing the self-paced version of this course in February 2026.
But this week, we're opening a Black Friday presale!
It will only last until Monday.
4/
But this week, we're opening a Black Friday presale!
It will only last until Monday.
4/
November 26, 2025 at 9:37 AM
We'll be releasing the self-paced version of this course in February 2026.
But this week, we're opening a Black Friday presale!
It will only last until Monday.
4/
But this week, we're opening a Black Friday presale!
It will only last until Monday.
4/
The course also does not shy away from discussing challenges, failures, and difficulties analysts and statisticians face in their daily work.
We learn from failures, so why not discuss them when we want to teach others?
3/
We learn from failures, so why not discuss them when we want to teach others?
3/
November 26, 2025 at 9:37 AM
The course also does not shy away from discussing challenges, failures, and difficulties analysts and statisticians face in their daily work.
We learn from failures, so why not discuss them when we want to teach others?
3/
We learn from failures, so why not discuss them when we want to teach others?
3/
Over 6 months ago we met with Justin Bélair, and we decided to embark on a journey together to build a course on biostatistics that does not shy away from discussing causality.
2/
2/
November 26, 2025 at 9:37 AM
Over 6 months ago we met with Justin Bélair, and we decided to embark on a journey together to build a course on biostatistics that does not shy away from discussing causality.
2/
2/
In general, partial identification might serve this purpose in certain cases.
Whether the particular approach described by Uriah could be useful in a case like this, I'd need to think about it
Whether the particular approach described by Uriah could be useful in a case like this, I'd need to think about it
November 26, 2025 at 7:39 AM
In general, partial identification might serve this purpose in certain cases.
Whether the particular approach described by Uriah could be useful in a case like this, I'd need to think about it
Whether the particular approach described by Uriah could be useful in a case like this, I'd need to think about it
In the described case, yes
November 26, 2025 at 7:38 AM
In the described case, yes
Enjoy!
-------------
This is an excerpt from Causal Python Weekly.
To get it in your inbox every Sunday, subscribe here (FREE): causalpython.io
5/5
-------------
This is an excerpt from Causal Python Weekly.
To get it in your inbox every Sunday, subscribe here (FREE): causalpython.io
5/5
Causal Python || Your go-to resource for learning about Causality in Python
A page where you can learn about causal inference in Python, causal discovery in Python and causal structure learning in Python. How to causal inference in Python?
causalpython.io
November 24, 2025 at 7:34 PM
Enjoy!
-------------
This is an excerpt from Causal Python Weekly.
To get it in your inbox every Sunday, subscribe here (FREE): causalpython.io
5/5
-------------
This is an excerpt from Causal Python Weekly.
To get it in your inbox every Sunday, subscribe here (FREE): causalpython.io
5/5
Part 1: Causal Bounds for Observational Data
Part 2: Causal Bounds for Experimental Data
Part 3: Causal Bounds for Observational Data with Pure Mediators
I gathered the links to all three parts for you here: causalpython.io#causal-bound...
4/
Part 2: Causal Bounds for Experimental Data
Part 3: Causal Bounds for Observational Data with Pure Mediators
I gathered the links to all three parts for you here: causalpython.io#causal-bound...
4/
Causal Python || Your go-to resource for learning about Causality in Python
A page where you can learn about causal inference in Python, causal discovery in Python and causal structure learning in Python. How to causal inference in Python?
causalpython.io
November 24, 2025 at 7:34 PM
Part 1: Causal Bounds for Observational Data
Part 2: Causal Bounds for Experimental Data
Part 3: Causal Bounds for Observational Data with Pure Mediators
I gathered the links to all three parts for you here: causalpython.io#causal-bound...
4/
Part 2: Causal Bounds for Experimental Data
Part 3: Causal Bounds for Observational Data with Pure Mediators
I gathered the links to all three parts for you here: causalpython.io#causal-bound...
4/
In his brand-new series of three tutorials, Uriah Finkel introduces us to the notion of causal bounds -- bounds on causal effects that may serve as a useful alternative to basic ways of answering causal questions under incomplete knowledge or unmet assumptions.
3/
3/
November 24, 2025 at 7:32 PM
In his brand-new series of three tutorials, Uriah Finkel introduces us to the notion of causal bounds -- bounds on causal effects that may serve as a useful alternative to basic ways of answering causal questions under incomplete knowledge or unmet assumptions.
3/
3/
In practice, we often want to answer causal questions under incomplete knowledge and/or unmet assumptions.
Causal bounds offer an alternative that allows us to relax some of these assumptions.
We pay for this relaxation in less precision, but we still hope to be able to get useful answers.
2/
Causal bounds offer an alternative that allows us to relax some of these assumptions.
We pay for this relaxation in less precision, but we still hope to be able to get useful answers.
2/
November 24, 2025 at 7:29 PM
In practice, we often want to answer causal questions under incomplete knowledge and/or unmet assumptions.
Causal bounds offer an alternative that allows us to relax some of these assumptions.
We pay for this relaxation in less precision, but we still hope to be able to get useful answers.
2/
Causal bounds offer an alternative that allows us to relax some of these assumptions.
We pay for this relaxation in less precision, but we still hope to be able to get useful answers.
2/
- Some exciting news from the Causal Python ecosystem are coming! (community, courses anyone?)
We'll start sending today's issue at 9am PT / 12pm ET / 6pm CET
Register here (FREE): causalpython.io
3/3
#CausalSky
We'll start sending today's issue at 9am PT / 12pm ET / 6pm CET
Register here (FREE): causalpython.io
3/3
#CausalSky
Causal Python || Your go-to resource for learning about Causality in Python
A page where you can learn about causal inference in Python, causal discovery in Python and causal structure learning in Python. How to causal inference in Python?
causalpython.io
November 23, 2025 at 1:49 PM
- Some exciting news from the Causal Python ecosystem are coming! (community, courses anyone?)
We'll start sending today's issue at 9am PT / 12pm ET / 6pm CET
Register here (FREE): causalpython.io
3/3
#CausalSky
We'll start sending today's issue at 9am PT / 12pm ET / 6pm CET
Register here (FREE): causalpython.io
3/3
#CausalSky