djinnome.bsky.social
@djinnome.bsky.social
Saturday Morning Breakfast Cereal - Strong share.google/nTEVGLO9s0kJ...
Saturday Morning Breakfast Cereal - Strong
Saturday Morning Breakfast Cereal - Strong
share.google
September 4, 2025 at 1:22 AM
@void.comind.network can you describe the software architecture upon which you were built? UML is preferred
July 27, 2025 at 8:10 PM
Older papers were often submitted in chapters:

I. Historical thing completely misrepresented by contemporary literature: 1700-1920

II. Historical thing completely misrepresented by contemporary literature: 1921-present
July 21, 2025 at 2:41 PM
Although, by this argument, mining bitcoins is a valuable scientific contribution...
July 14, 2025 at 12:40 PM
The value of a scientific contribution is the delta between how expensive a measure is to assess and how cheap it is for the next scientist to assess the measure.
July 14, 2025 at 12:36 PM
Isnt the drop in albedo attributable to polar ice caps melting due to increasing CO2?
In other words, it is a positive feedback cycle where increasing CO2 leads to drop in albedo which contributes to rising temperatures which leads to more polar melting leading to more albedo drop...
July 12, 2025 at 1:06 PM
But without domain knowledge, the ppc cannot violate your expectations, you are blind.
June 17, 2025 at 1:26 AM
I like the quote, but the explication of the quote needs some work. In a Bayesian setting, prior predictive checks are uncertainty estimates without data. They are generally used as sanity checks to confirm that the model predictions don't violate expectations of domain knowledge.
June 17, 2025 at 1:23 AM
The Book of Why provides a Whig history of how this came to be as a journey from an oppressive and benighted past to a "glorious present"
June 4, 2025 at 11:41 PM
Any word from your uni as to whether they will cover the expenses to make this magic happen?
May 9, 2025 at 8:58 PM
I am teaching a causality workshop in June and I would *love* to show a plumbing video of D-separation. How can I support you in this?
May 2, 2025 at 5:51 PM
Why choose? Just let lambda in the Poission be a gamma distribution and you get a negative binomial anyway.
April 18, 2025 at 7:53 AM
You can check out github.com/y0-causal-in...

We have identification algorithms for causal, counterfactual, and transportability queries
GitHub - y0-causal-inference/y0: ❓y0 (pronounced "why not?") is for causal inference in Python
❓y0 (pronounced "why not?") is for causal inference in Python - y0-causal-inference/y0
github.com
November 24, 2024 at 10:24 PM
Most nodes are not needed to identify a causal query, so you can use identification algorithms to determine which nodes you can make latent without affecting the identifiability of the query.
November 24, 2024 at 7:28 PM