clircle

43 ms·

1. 511.▲
   by clircle 6y ago

   Here is a github page for an Emacs package manager: https://github.com/raxod502/straight.el It's a good page because it contains comparisons too all other emacs package managers[0], and carefully describes the pro
2. 512.▲
   by clircle 6y ago

   I recently moved my 400 GB music collection to nextcloud music. It is served from a Rpi on my bookshelf with an attached 1 TB SSD. So far, so good. It is streaming my music collection to computers on my LAN, but not outside my network. I&#x
3. 513.▲
   by clircle 6y ago

   What are the great successes of machine learning in finance?
4. 514.▲
   by clircle 6y ago

   I do as much data work as possible in data.table, which parallelizes things under the hood, and doesn't seem to require me to configure anything.
5. 515.▲
   by clircle 6y ago

   I like i3/sway on small screens, but on my large 4k screen all the windows are too big, too much white space. For me, the tiling paradigm does not work with big monitors. Linux on the whole doesn't work well with 4K. I don't
6. 516.▲
   by clircle 6y ago

   What is your suggested alternative solution?
7. 517.▲
   by clircle 6y ago

   Mostly agree with the author, but The best anti design would be to turn off the damn recommender system and liking, rather than putting pagination back and hiding follower counts.
8. 518.▲
   by clircle 6y ago

   Based on my readings of tech reviews, it seems like the macbook M1 is the way to go right now. Great battery life, great performance. Maybe a better keyboard than a dell, probably not as good as a thinkpad. What's the argument against
9. 519.▲
   by clircle 6y ago

   No halfway decent statistician has ever advised that one bootstrap a time series. Naive bootstrap obviously does not apply to that data. And the bootstrap is not for taking a small amount of data and making it a large amount, it’s an empiri
10. 520.▲
    by clircle 6y ago

    "The concept of Evince is one document, one window"
11. 521.▲
    by clircle 6y ago

    I mostly like Gnome. I have been using it for years. But the development team of Gnome is so closed minded that it makes my head hurt. I asked on the bug tracker if Evince could have a tabs feature, so I can have 10 pdfs open in one Evince
12. 522.▲
    by clircle 6y ago

    I've noticed this too. The trackpoint on my X220 is excellent, but something about about the later notebooks feels very wrong. For example, I'm using a T460, and the trackpoint is always drifting, and can't be calibrated to t
13. 523.▲
    by clircle 6y ago

    This is technical and interesting, but can anyone tell the difference between web browsing with FF/uBo and Chrome/uBo? I personally cannot, other than that the fonts render a bit differently. Webpages load fast and no ads get thro
14. 524.▲
    by clircle 6y ago

    Syncthing is great, but the lack of first party iOS client drove me to Nextcloud.
15. 525.▲
    by clircle 6y ago

    I have a similar setup, but I'm using a raspberry pi. Does anyone know a good iOS client for Nextcloud Music?
16. 526.▲
    by clircle 6y ago

    This is very interesting. Can you recommend some readings about these features? Especially as they may pertain to R.
17. 527.▲
    Cross-validation: what does it estimate and how well does it do it?(arxiv.org)

    9 points by clircle 6y ago|0 comments
18. 528.▲
    by clircle 6y ago

    I've never understood what makes R lispy. I think R is a very flexible and interactive language, more so than Python, but does that mean it is lispy? Pardon me, I don't know enough about languages to pin it down.
19. 529.▲
    by clircle 6y ago

    Outside of simple time series, I'm not aware of any good way to extrapolate.
20. 530.▲
    by clircle 6y ago

    The only reason that this may not be accurate for ML is because machine learners generally make no attempt to quantify their uncertainty in their predictions with e.g. confidence intervals or prediction intervals. And there is a whole field
21. 531.▲
    by clircle 6y ago

    Luke is the architect of xlisp-stat and current R code team member.
22. 532.▲
    by clircle 6y ago

    I'm intrigued. This looks like an update to xlisp-stat that interfaces with some newer CL libraries. There is a nice paper about why UCLA switched from Lisp to R in the 90s. https://www.jstatsoft.org/article/view&#
23. 533.▲
    by clircle 6y ago

    I prefer to call AI “random number generators”
24. 534.▲
    by clircle 6y ago

    Brings back fond memories of Diablo and warcraft.
25. 535.▲
    by clircle 6y ago

    Despite the downvotes, I'm still curious about the problem that this solves.
26. 536.▲
    by clircle 6y ago

    What is the point of this work? Other than that it is cool.
27. 537.▲
    by clircle 6y ago

    So how many times did they have to tweak their algorithm to get this nice result?
28. 538.▲
    by clircle 6y ago

    Is it faster though?
29. 539.▲
    by clircle 6y ago

    I think this is an important problem. With logistic regression or deep learning, at least one can compare (out of sample) calibration curves or discrimination measures. With a language model, what can we do?
30. 540.▲
    by clircle 6y ago

    Thank you for this thoughtful comment. I assume you are talking about use-package? I've been using it for several years. There is a long standing effort to have it merged into emacs/elpa, so we hope the situation will improve. ht