• 0 Posts
  • 74 Comments
Joined 7 months ago
cake
Cake day: November 17th, 2024

help-circle
rss


  • This is interesting, I would be quite impressed if this PR got merged without additional changes.

    I am genuinely curious and no judgement at all, since you mentioned that you are not a rust/GTK expert, are you able to read and and have a decent understanding of the output code?

    For example, in the sway.rs file, you uncommented a piece of code about floating nodes in get_all_windows function, do you know why it is uncommented? (again, not trying to judge; it is a genuine question. I also don’t know rust or GTK, just curious.


  • The image is taken from Zhihu, a Chinese Quora-like site.

    The prompt is talking about give a design of a certain app, and the response seems to talk about some suggested pages. So it doesn’t seem to reflect the text.

    But this in general aligns with my experience coding with llm. I was trying to upgrade my eslint from 8 to 9, and ask chatgpt to convert my eslint file, and it proceed to spit out complete garbage.

    I thought this would be a good task for llm because eslint config is very common and well-documented, and the transformation is very mechanical, but it just cannot do it. So I proceed to read the documents and finished the migration in a couple hour…










  • Took me a several click to get to the source: https://budgetmodel.wharton.upenn.edu/issues/2025/4/3/real-time-federal-budget-tracker It has detailed budget breakdowns so it is decently convenient to explore.

    I have done a very brief scrolling, here are some interesting findings. All the following data are year-to-date, comparing 2024 to 2025, adjusted for inflation:

    Ups:

    • despite mass firing, spending on federal employee salary has slightly gone up (by 4 billion, from 79 to 83 billion), which I assume is either current admin paying themselves, severance, or sign-on bonus to hiring back their fired employee.
    • DoD and DVA spending has collectively gone up more than 20 billions.
    • Unclassified spending up aroubd 20 billion, from 70 billion to 89 billion, a near 30% incease, curious what those are.
    • Federal Highway, Railroad, and Transite are collectively up couple billion.
    • DOJ, DOE, DOC has surprisingly gone up in spending. DOJ in particular gone up 1 billion, from 7 billion to 8 billion.

    Downs:

    • NIH and CDC are either slightly down or remain the same. It is worth noting that NIH and CDC collectively account for around 20 billions total spending, the same amount as the increase in DoD and DVA spending.
    • Department of Education down 10 billion from 90 billion to 80 billion.
    • USAID gone down 2 billion, even though it only accounted for total of 6 billion of spending in 2024.
    • Homeland Security spending, surprisingly, down half a billion from 5.5 to 5.
    • USPS down 1 billion, from 14 billion to 13 billion.



  • My conspricy theory is that early LLMs have a hard time figuring out the logical relation between sentenses, hence do not generate good transitions between sentences.

    I think bullet point might be manually tuned up by the developers, but not inheritly present in the model; because we don’t tend to see bullet points that much in normal human communications.