erikgahner

7 ms·

1.
by erikgahner 3mo ago
This was also discussed on HN a few days ago.[1] [1]: https://news.ycombinator.com/item?id=48552440
2.
by erikgahner 3mo ago
Got it, thanks!
3.
by erikgahner 3mo ago
What exactly is the innovation here compared to the press release from two years ago?[1] [1]: https://www.unsw.edu.au/newsroom/news/2024/05/Ultrasonic_col...
4.
by erikgahner 7mo ago
I wouldn't call this a minor detail (i.e., nitpicking), and it is worth pointing out again and again when these studies get public attention. We should encourage stronger research designs (including A/B tests) if we care about the
5.
by erikgahner 7mo ago
A lot of people might read this and infer that AI use causes depressive symptoms, but the study cannot say anything about causation at all. The study is also transparent about this fact: "Further work is needed to understand whether th
6.
by erikgahner 8mo ago
Great stuff! You can make minor adjustments to the R-script so you do not need to rely on {dplyr} and {tidyr}. For example, use merge() instead of left_join() and use the base pipe, |>, instead of the magrittr pipe, %>%.
7.
by erikgahner 8mo ago
https://erikgahner.github.io/
8.
by erikgahner 9mo ago
Not OP but I guess Gladwell-style can be understood as writing that is presented as nonfictional but has more in common with fictional writing.
9.
by erikgahner 10mo ago
https://archive.is/84dmE
10.
by erikgahner 10mo ago
They are already combined in dtplyr[1]. [1] https://dtplyr.tidyverse.org/
11.
by erikgahner 10mo ago
Interesting to see far-right terminology being used unironically at HN.
12.
by erikgahner 1y ago
100% agree. I am more than happy to see Astral taking steps in this direction. People can continue to use uv, ruff, and ty without having to pay anything, but companies that benefit tremendously from open source initiatives can pay for a pr
13.
by erikgahner 1y ago
Looking good! What does it mean that it is free forever? That the code is available on GitHub and we should not have to be concerned about the site going offline or moving to a different place (that is not free)? Or, if it takes off, that w
14.
by erikgahner 1y ago
Did you write the stand-up comedy set? This have all the telltale signs of AI (down to the use of em dash).
15.
by erikgahner 2y ago
The article is more than five years old (from February 2020).
16.
by erikgahner 2y ago
Most of these examples/walkthroughs look like they have been generated by LLMs. They might be useful for teaching purposes, but I believe they significantly underestimate the amount of domain knowledge that often goes into data extract
17.
by erikgahner 2y ago
Point taken! I should have considered a subtitle for the post to make it clear what my main point is. That being said, I tried to nuance the post and not make any bold predictions. I am happy to revisit the argument in five years and see wh
18.
AI will be dead in five years(erikgahner.dk)
41 points by erikgahner 2y ago|38 comments
19.
by erikgahner 2y ago
Fair point. My personal preference is still for blog posts to share this information in the URL, i.e., to communicate when a blog post was published even before I click a URL. Usually bloggers tend to write new entries if they change their
20.
by erikgahner 2y ago
Good point! There is a difference between blog posts and pages to take into account here. For blog posts, I would like to see at least the publication year in the URL, as well as the exact date in the post itself. For pages, I am fine with
21.
by erikgahner 2y ago
I agree with a lot of these ideas but I do not get the obsession with short urls. E.g., why not make it easy for readers to see the date of when something was published? In this case I cannot even see a publication date on the post itself (
22.
The CrowdStrike Incident(thezvi.wordpress.com)
2 points by erikgahner 2y ago|0 comments
23.
by erikgahner 2y ago
ggplot2 is pretty much an exception at this point, and if the package was created today it would use the pipe instead of add [1]. Most packages use pipes now (especially with the introduction of the pipe operator in R 4.1.0). [1] https:&#x
24.
by erikgahner 2y ago
I am not against AI in software per se (I like the ChatGPT desktop app), but I prefer not to have AI in tools where the primary use case is not related to AI. Specifically, I don't want to have to consider how the AI is being developed
25.
by erikgahner 2y ago
I have gotten into the habbit of removing any software with opt-in AI chatbots. I recently uninstalled iTerm2, and I fear Firefox will be next.
26.
by erikgahner 2y ago
Sure looks like it. Here is an archived version: https://archive.is/c4XbV
27.
by erikgahner 3y ago
https://archive.is/hLl2l
28.
by erikgahner 3y ago
Thanks! I had problems accessing the webpage (due to my adblocker).
29.
Twenty Years of Marginal Revolution(erikgahner.dk)
3 points by erikgahner 3y ago|0 comments
30.
by erikgahner 4y ago
That's fair, I should have elaborated a bit more on that point in my post. My thinking is that I need to use principles that will optimise the signal-to-noise ratio. In other words, just consuming everything in my feed is not optimal.

More ›