Here is a little teaser of the next version of SQL.
You’ll be able to ergonomically work on the result set to transform it, this is done in one parse, so no immediate list are created, so you get high performance no matter how you organize your code!
#elixir-lang #ecto #sql #postgres #tds #mysql
Jacob is a such a great interviewer, I’m looking forward to see one!
Yeah, but I think that is more confusing then just treating the emojis as atom when not quoted.
I wish there was a better syntax for this, ideally emojis should be treated as atoms if not quoted.
based on guess alone, I would think comma ai.
Yeah, as we’re a small community I don’t think that any of the bigger LLM players is going to help us out and make the experience better. With tinyboxes we could crowdsource LLM fine tuning in the community and make Elixir best in class for LLM workflows.
I’m not sure where tidewave is going, but if it was me then I would be spending the time fine tuning deepseek, it’s open source, and if tidewave has some funding I’m sure they could afford a tinybox to train it on!
I was thinking about doing everything at compile time, which is possible, but it might not be needed with these numbers:
They recently hired Devin. And fired their entire Indian offshore onsite dev team.
This is wild, [sql](github.com/elixir-dbvis...) is between 400-650x faster then Ecto at generating SQL groups.google.com/g/elixir-ect... [read more!](elixirforum.com/t/sql-brings...)
