Sorry, I really wanted to make this presentation but I wasn't able to. Will there be a recording made available anywhere?

Is this meant to be tariff related, AI related or both?

I’ve worked with several LLM frameworks in Python. ellmer is so far and away better than any of those frameworks I would never consider using Python for AI work (and if I have to, I’ll use chatlas)

I’m not convinced that AI is bad for R. An obvious talking point is “post language data science”: it now matters less than ever what exact language you use to express your data science code. (But code is still an key part because you must have something that humans can review)

Most of my world is in protyping ideas with messy data and Python is such terrible syntax for wrangling. I agree with Hadley that LLMs are making languages more irrelevant. I really think soon even python syntax knowledge will be largely obsolete.

This looks interesting Matt. Will have to update my computer but running locally would definitely solve a lot of problems

These are actual AI companies, so I assumed they wouldn't be susceptible to their own hype. If you have insights/perspectives on other companies I'd be interested in hearing. Like I said, I still can't understand how vibe coding is taking jobs. But at the same time AI companies are reducing hires.

A year ago I would have agreed. The fact AI companies are scaling back staff (dramatically) suggests something internally is replacing engineers (not necessarily vibe coding as described here). Considering the markets rally on the back of AI I don't think these layoffs are because it isn't working.

No one from #rstats who has faced a similar conundrum of having proprietary code, not having funds for enterprise models, but being able to use these tools in a safe way?

Oh and here's the website: ellmer.tidyverse.org

Thanks Hadley for the suggestions and thanks for getting back to me