Love this tool. Invaluable for figuring out what code actually got called after complicated dispatch.

Hey, that's not a fair assessment.

It would also dump greenhouse gasses into the atmosphere.

Anaconda distributions aren't exactly small. I'd also assume there's some sort of environment isolation, so you're getting multiple copies. Maybe one per workbook?

The embeddable package from python.org [1] is 20 MB unpacked. To install different sets of packages per workbook, just have Excel automatically set the module path as appropriate.

[1] https://www.python.org/downloads/release/python-3114/

No need to use Anaconda or any general purpose venv solution. Microsoft screwed this up.

kNN methods are broadly not O(n^2)[1], especially in practice where approximate methods are used.

[1]: https://en.wikipedia.org/wiki/Nearest_neighbor_search

how would you build an index over the gzip encoded data? seems quite different from building indices over vector embeddings.

> big acquisitions that, in hindsight, made Big Tech more competitive (the most common examples are Google acquiring DoubleClick and YouTube, and Facebook acquiring WhatsApp and Instagram).

How did these acquisitions make big tech more competitive?

I’m saying it made those companies more competitive (as in “better equipped to compete”), not that it made the various markets more competitive (as in “characterized by competition by multiple actors”).

“Big Tech” tends to refer to the companies.

Edit: To clarify further, I’m not casting those claims about older Big Tech acquisitions as wrong or misinformed. Nor am I trying to make any specific argument about the Democrats’ strategy / Neo-Brandeisian thought here. Just directly answering the question I replied to. I’m quite confident that this particular case against Microsoft, as argued by the regulators, is quite weak and that the political tail is wagging the dog here.

> Most elevators (at least in NYC) start at floor 1, not 0.

European buildings are generally 0-indexed: https://en.wikipedia.org/wiki/Storey#Numbering

The same applies to most former continental Europeans colonies in latin america and africa. 0 based indexing, sometimes with Floor 0 (argentina comes to mind) or simply ground (Terreo in Brazil)

Except people call it the "ground" floor which makes sense, not the "zero" floor which does not.

It’s not consistent but I would estimate that “0” buttons are as common as “G” buttons in elevators in Europe - along with -1, -2, etc for basement levels.

0 makes sense when you consider -1 to be the first basement level.

Which also gently introduces children to negative numbers : - )

Currently living in Munich, it is the quietest place I ever lived. Also super clean, like more than Singapore.

ibis in python is a really nice middle ground. Nice API + in a programming language, but executes on database backends (which could be polars or duckdb on in memory arrow tables).

> it's really solving the problems that hurt AI most

Isn't this a bit premature? Mojo doesn't tangibly exist for most people (we can't run it ourselves), and I am unaware of any ML/ AI applications built with Mojo.

Yeah, definitely early days. But they're saying all the right things: fast code off the bat, easy to vectorize and parallelize, efficient use of memory, easy to create abstractions that execute efficiently on various architectures. The demos are impressive and they have the track record to be credible. I've played with it and the stuff that's been implemented so far works well.

Swift for Tensorflow was such a success.

Oh boy, I remember reading about this for the first time and thinking it was April's fools.

I have tried it. Iirc, the company developed it, or its compiler, specifically for ML/“””AI”””.

If they successfully import Pythons libraries without a bunch of wrapping, its userbase will likely follow.

It was faster than Python and easy to do paralleization like go. But it’s not quite fully baked yet.

Is there a description of what exactly an organizational account entails?