causal → callum

data scientist, @InfAtEd alum, causal inference, experiments and machine learning

[causal → callum @causalcallum]

I just finished deploying my first blog post at https://t.co/bJgkzQwiGt on the topic of causal inference and structural causal models! ➡️

[causal → callum @causalcallum]

@tayloramurphy how I sleep setting “dbt build” as a single airflow task

[causal → callum @causalcallum]

officially at my wits end with twitter - my algorithm these days just shows me musk, snake oil sellers, and public outrage

[We're just normal men @wejustnormalmen]

We’re just normal men

[causal → callum @causalcallum]

“Please output this as JSON”

[Mahesh Sathiamoorthy @madiator]

How to get LLMs to output JSON ⬇️

[causal → callum @causalcallum]

we're so back

[Jo Kristian Bergum @jobergum]

This is wild. With OpenAI's text embeddings, we can crush current legacy dating apps.

[causal → callum @causalcallum]

@jobergum I see this for larger analysis queries where being 100% precise has less importance, but I struggle to see how it’s easier to query at a row level granularity using NQL vs SQL

[causal → callum @causalcallum]

when will all clothing sites realise that “model is XXXcm and wears size X” is the best size guide

[causal → callum @causalcallum]

@jokull god save the king

[Erik Bernhardsson @bernhardsson]

Few things bother me more than when I commit code with a beautifully sorted list of strings/enums/whatever, and then the next person just adds their thing at the bottom of the list.

[Rémi 🌳 @remilouf]

"And then they started making programs that call our API in a `while True` loop"

[causal → callum @causalcallum]

because buildings are now funded with tax payer money instead of stolen empire money

[The Cultural Tutor @culturaltutor]

Why don't we make interesting drain pipes any more?

[causal → callum @causalcallum]

Had to make to some matplotlib charts today at work and discovered this is the perfect use case for GPT4

[causal → callum @causalcallum]

@Benjaming_G appreciated! 🍝

[causal → callum @causalcallum]

look at me, a data scientist, who's usually writing spaghetti in notebooks consuming a big boy API with TypeScript

[causal → callum @causalcallum]

@rabois @instrumenthull do it

[Demetri (is over at the other place too) @PhDemetri]

These two DAGs comprise 80% of my job

[causal → callum @causalcallum]

my thing this week is offsetting nit-pick PR comments with 😜 to soften the blow

[causal → callum @causalcallum]

@pdrmnvd I recently got a high-speed train from Madrid to Malaga, which took 2hr30mins. LA to SFO is a similar distance - what's not to like?

[causal → callum @causalcallum]

@pdrmnvd but what have you got against high-speed trains between major cities? Imagine a west coast line connecting san diego, sfo, la, and portland