1. kipukun Nov 21, 2025 parent

   XGBoost is just faster on the GPU, regardless of the file format. A sibling post also pointed out compression helping out on bandwidth.
2. kipukun Nov 20, 2025 parent

   The cuDF interop in the roadmap [1] will be huge for my workloads. XGBoost has the fastest inference time on GPUs, so a fast path straight from these Vortex files to GPU memory seems promising.

   [1] https://github.com/vortex-data/vortex/issues/2116

3. kipukun Aug 5, 2025 parent

   This game is awesome. I have no other feedback other than I had a lot of fun.
4. kipukun Feb 24, 2025 parent

   You'd be surprised. As long as your query is using Polars natives and not a UDF (which drops it down to Python), you may get good results.
5. kipukun Feb 24, 2025 parent

   To the second footnote: you could utilize Polar's lazyframe API to do that cosine similarity in a streaming fashion for large files.
6. kipukun Feb 9, 2025 parent

   For your wimsey library, using “pipe” to validate the contracts would seem to me to drastically slow down the Polars query because the UDF pushes the query out of Rust into Python. I think a cool direction would be to have a “compiler” which takes in a contract and spits out native queries for a variety of dataframe libraries (pandas/polars/pyspark). It becomes harder to define how to error with a test contract but that can be the secret sauce.
7. kipukun Jun 28, 2024 parent

   Most people want unlimited for the convenience. You can very easily go to Fi or a budget carrier for more fine-grained plans like the ones you described.

   EDIT: typo

8. kipukun Jun 26, 2024 parent

   Looking at your YC history, Rubbrband was initially meant to be "a way to train open-source Machine Learning models in just one line of code." I'm curious what this initial offering was and why you pivoted.
9. kipukun Jun 10, 2024 parent

   Gainax's financial troubles are well documented in an interview with Hideaki Anno, where he detailed how the company was pretty dysfunctional from the start. The company's president was arrested for tax fraud in 1999, for example. This has been a long time coming.
10. kipukun May 11, 2024 parent

    (2021)
11. kipukun May 6, 2024 parent

    On the Parquet topic, the Polars lazyframe workflow of operating on a Parquet file is so convenient for me when wrangling hundreds of millions of rows. You avoid loading the data into a SQL server if you don't need to.
12. kipukun Apr 13, 2024 parent

    Might've been overlooked, but "fastly" should probably be "quickly" on the main page.
13. kipukun Feb 27, 2024 parent

    Do you think you'll add local file support in the future? Also, do you have any plans on making the reading of a source parallel? For example, connectorx uses an optional partition column to read chunks of a table concurrently. Cool how it's abstracted.
14. kipukun Jan 10, 2024 parent

    I'll bite -- he discovered a vulnerability in a large company and responsibly disclosed it to them. How is that a felony? Why would you post a felony online?
15. kipukun Nov 10, 2023 parent

    1. I use it on all of my servers.

    2. FreeBSD offers a coherent base system with sane tools. pf is a godsend, and the ports collection is a good trade-off of configuration and simplicity. ZFS used to be a killer feature but is being integrated into Linux systems.

    3. FreeBSD is designed around "general purpose" use and has a lot of backing from corporations like Netflix.

    4. FreeBSD is missing a true container solution that is compatible with Docker (jails don't count). There has been some work on this as of the most recent BSDcon, so I'm hopeful.

    5. I never want to look at iptables again.

    EDIT: fix spacing

16. kipukun Aug 31, 2023 parent

    Why not explain that POV?
17. kipukun Mar 31, 2023 parent

    so you're saying Valve should release software that, if a bug appears in this now unsupported Chrome version, is now susceptible to this bug that will literally never get fixed?
18. kipukun May 2, 2022 parent

    Neat project. Any roadmap for cross-validation support (GridSearchCV and friends)?
19. kipukun Dec 21, 2019 parent

    You know, I really like this idea. It's simple, straight-forward, the website loads, and its demographic is mostly everyone. Good stuff. One qualm is that two scrollbars show up, probably from some overflow: scroll stuff. Would consider looking at that.
20. kipukun Dec 14, 2016 parent

    Reading the article, they analyze data they show on charts. You could make the same conclusions they're drawing by also looking at the charts. Sure it's a shameless plug, but it definitely doesn't discredit their analysis.

This user hasn’t submitted anything.