Forgot to post this the other day. Had some leftover yakiniku bbq grilled salmon, so flaked it and sous-vide heated it; served it along side a 100% local salad (incl. the honeyed goat cheese). julienned beets + salad turnips FTW
Bonus Drops are back! Many thanks to those that provide extra support for their patience during the upheaval of the past couple weeks. There's a bit of something for everyone in this edition.
Before I go too far down the rabbit hole with this, I'm considering re-upping my "21 Recipes" Twitter book (https://rud.is/books/21-recipes/) for Bluesky.
I suspect it is going to take decades to recover from the damage being done in ~20 🇺🇸 states all out of fear and hatred of “other”, and the daft need to control the lives of, especially, women.
This is a woefully incompetent take by @washingtonpost. 10% drop to the main ChatGPT site seems to make sense to me when the chat feature is baked into Windows (still the most popular desktop OS), is free on Bing, and available in many other forms (like my fav: perplexity.ai).
It's also June numbers. The start of vacation time. School is out for a large portion of humans.
Terrible, terrible analysis. Makes me sad I give them 💵 if this is what they churn out.
Seeing people on LinkedIn referring to R as a dead language (and also seeing that the latest edition of a seminal stats /ML book now has examples in Python).
Juxtaposed with someone who saw me demonstrate the power of R back in 2018, and has now decided to learn it.
So far, they have been shown projects, how to import from Excel, not to have spaces in file or field names, where to find dplyr help, and how to rename columns.
They've done alright so far
And, to think the boy billionaire charges, what, $40K/month for this same thing you get for free (for now) on bsky? Oh, and that firehose doesn't require auth.
Making an automated post bot thing is also super easy.
The atproto python module is really well thought out.
There. I've used up my “say something nice about python" quota for the year.
@rmflight I’m watching this one with a bit of caution. Their results haven't been replicated, yet, and it's just on historical data, so they could have gamed it a bit (we're in the snake oil/money grab stage of AI, after all).
Just learned that #2.2 will be a lad! Cannot wait to see him and #2.1 grow up together. Given how fast 2023 has been flying by, January will be here in no time, too.
@johnmackintosh@rmflight@josiah I’m back to {base} these days for reasons that will make me a pariah in some circles if laid bare. All core R needs are a couple more data-first functions to fill in some gaps and one can be as expressive and efficient as the otherverse.
@AlgoCompSynth@johnmackintosh@rmflight@josiah I can 100% grok how that’d be the case! Capricious (personal hot take) API changes + other things I won't discuss are my core issues. I rly wish VS Code was a better R IDE, but I'm trying to be JS-first for datavis these days, so it will ultimately not matter in a year or two.
