Kyle MacRitchie

Senior Meteorologist somewhere focused on quantitative ag wx, formerly known as an NWS meteorologist // tech buff, scotch lover // PhD Atmospheric Science

[Kyle MacRitchie @wxscience]

@AlexandriaVATES @ManHeHim @AlexandriaVAGov missed both trash and recycle this morning on s payne st, judging by neighbors the entire alley was skipped?

[Kyle MacRitchie @wxscience]

@Drlipid Do you have a post somewhere about the current best intervention(s) for someone with genetically high Lp(a) 250ish in her mid 30s with very low LDL-C/APOB?

[Kyle MacRitchie @wxscience]

@AlexandriaNow Walked by Sprig's Cafe in Carlyle this morning and it looks like it's closed? Windows totally boarded up. Any news?

[Kyle MacRitchie @wxscience]

@Selwyyyyn Hello. Thanks for letting me know. I will look into it.

[Kyle MacRitchie @wxscience]

@ekstasis anyone who knows anything about pizza near the pentagon would be eating wiseguy instead of papa johns. just sayin

[Kyle MacRitchie @wxscience]

@skdh We should reward quality over quantity. Simply aren't enough hours in the day to both churn out papers to make the admins happy and to put in the thought and effort required for discovering the next big thing. And people wonder why science seems to be so incremental these days.

[Kyle MacRitchie @wxscience]

@tcrawf_nh Have there been any assessments of how well the AI models deal with tail events? Would be an interesting study to see.

[Kyle MacRitchie @wxscience]

@DrAaronHill i'd add in pytorch given the industry emphasis on ML/AI these days. Can't go wrong knowing how to build wx based ML models

[Kyle MacRitchie @wxscience]

@VaDOTNOVA @BikeGillian hello - also interested; did you ever find out?

[Kyle MacRitchie @wxscience]

@ZimermanRicardo @KristerKauppi can you share the thought process behind taking a break from rapa?

[Kyle MacRitchie @wxscience]

@cloudpr0phet interesting, although I'm not convinced that this deficiency matters much to their potential

[Kyle MacRitchie @wxscience]

@elonmusk @elonmusk you should use xAI to improve weather forecasting. It's a crowded field but there's substantial upside if its done right, especially if you focus on long ranges

[Kyle MacRitchie @wxscience]

Weather folks - does anyone know of a place that has CFS performance metrics, specifically for timescales >= 1 month?

[Kyle MacRitchie @wxscience]

@JeninYounesEsq @AllysiaFinley Don't think this is new. The new thing that stands out to me is this nonsense that consensus (ie. approved narrative) equates to good science. Science and consensus have no business being in the same phrase.

[Kyle MacRitchie @wxscience]

@shawnmilrad ask chatgpt or claude to convert it to python??

[Kyle MacRitchie @wxscience]

@SnazzyLabs Sounds like they are divisive but they also have a reasonable return policy so umm.... Who cares? Don't like it? Return it and hit up dbrand or whomever you prefer.

[Kyle MacRitchie @wxscience]

@robot @dbrand gotcha - do you have an ETA on those?

[Kyle MacRitchie @wxscience]

@dbrand your z fold5 case link goes to a 404 error. Any chance that's an error and you actually have them?!

[Kyle MacRitchie @wxscience]

@BackyardBeans really enjoying your 12 oz Spark nitro can. Do you know how much caffeine it has? thanks

[Kyle MacRitchie @wxscience]

That's more or less what I've thought also. Regardless of how it pans out, I think it's pretty clear that AI won't replace conventional NWP, but rather will augment it. It's a team sport.

Also, regarding your statement that probabilistic forecasts are useful... I think way too many consumers of them have trouble turning them into actionable insights (e.g. forecasts are probabilistic but decisions are deterministic). Maybe AI can help in that regard as well.