firexcy

Many Chinese users actually hate this feature because of how it's designed. There's a mandatory pop-up upon the first launch of every app to ask for your choice, which can be clumsy if you're setting up a new device. If you tap the wrong option (which is common since the dialog is small and crowded), you have to go deep into system settings to restore internet access. Moreover, many foreign app developers know nothing of this and have their apps throw an error even before a Chinese user has the chance to give permission, forcing them to manually kill and reopen the app to be online, adding to the errands.

Great project. It’s a surprise to learn that it’s possible to make it to #598 with three posts.

(I’m not sure why the data from my old domain, cyhsu.xyz, hasn’t been aggregated to the new hsu.cy, despite the methodology page saying it should. Must I return the canonical header in addition to a 301 code?)

It was also during the past year that I came to realize how powerful and versatile Hugo is. I had used Hugo for many years, touching only the most basic feature set. Last year, I decided to be done with Twitter and Instagram and make my own timelines of text and photo posts with similar layouts. Initially, I thought it might require separate instances of GoToSocial and Pixelfed. It turned out that Hugo could do it all with a few tweaks, and now they are live at hsu.cy/notes and hsu.cy/gallery, respectively. I highly recommend it to everyone who wants to start their own blog.

I appreciate his work for being more informative and organized than average AI-related content. Without his blogging, it would be a struggle to navigate the bombastic and narcissistic Twitter/Reddit posts for AI updates. The barrier to entry for AI reporting is so low that you just need to give a bit more care to be distinguished, and he is getting the deserved attention for doing exactly that in a systematical and disciplined manner. (I do believe many on HN are more than capable but not interested in doing the same.) Personally, I sometimes find his posts more congratulatory or trivial than I like, but I have learned to take what I want and ignore what I don’t.

Shameless plug: I wrote this as I tried to understand HN: https://hsu.cy/2025/09/how-to-read-hn/ (Discussed here: https://news.ycombinator.com/item?id=45331030 ) I know it only scratches the surface, but hopefully it can provide some basics.

I think what’s telling isn’t whether em dashes are used but where they are used. You can still use em dashes as you see fit in serious academic and literary writing without others in the community raising an eyebrow. But a sudden proliferation of em dashes in a tabloid or an online forum will surely and rightfully raise doubts about assisted writing.

That’s correct; I somehow messed up that paragraph during editing. Thanks for pointing it out.

Author here — the image quality was comparable, if not better, which is understandable because both use pretty standard commercial scanners. An upside is that instead of cutting the binder as 1DollarScan does, they tore the pages from the seam, so the pages remained whole. They didn’t do any compression either.

Agreed. I used to struggle with remembering all these names in novels, but recently came to terms with the “dysnomia” by drawing parallels between reading fiction with hearing anecdotes, where capturing the rough dynamic and vibe is more important than remembering characters; confusing names is venial if the confusion is part of the experience.

learn from the apps in […] China.

Please don’t. Most mobile apps made by Chinese developers, esp. big techs, do employ grid, paged layouts everywhere as though multiple iOS home screens were squeezed into the app. However, most grids in such layouts are useless, distracting, and even malicious from a UX perspective, their mere reason to exist being to steer users into endless rabbit holes of the devs’ multiple lines of businesses for KPI purposes, and thus subject to constant and arbitrary changes. As such, you can find an icon for personal financing in a cloud storage app, or find an icon for groceries in a ride hailing app, only to be replaced with icons for online dating and hotel booking and something something a week later. This density of user-adversarial features is to be avoid by all means.

CSS selectors with square brackets

They are also extremely useful when it comes to adblocking. Without their partial matching capabilities it'd have been almost impossible to target ad-containing elements in most of these days' js and framework-ladden websites.

I am a paying Kagi user and it seems to me that the post is from an over-zealous user venting after their unsolicited advice was rejected. Reading through the post without finding a single mention of search quality is quite telling about its content.

There is no reason for Kagi to remain “pure” and avoid AI features as suggested by hardcore AI haters. I am not a fan of AI hype either, but I am pleased to see that Kagi has integrated some moderate capabilities such as the summarizer and search-based generation, which are natural extensions of a modern search engine. (I do hope they improve the expert mode soon, as it is currently far inferior to Perplexity, but that does not invalidate the general point.)

Email-based account management may not be perfect from a privacy perspective, but registering with a privacy email alias has mostly resolved my concerns. As for GDPR, let’s not pretend that it is disproportionately burdensome for startups. I value the way a company operates much more than the privacy theatres (banners, opt-outs, legaleses) enforced by GDPR.

Other criticisms regarding operational details range from nitpicking to trivial. I do hope that the founder was less insistent on arguing with and lecturing zealous users like the author.

These days I stick to a minimal profile for the default shell (zsh on macOS and bash on Linux), adding no further than some aliases. When I need more “friendly” interactive features I launch fish which is… friendly and interactive, and exit when I’m done. Thus I get the best of both worlds.

I have a high prior that any nonprofit that hasn’t been rigorously shown to be good is probably bad, and the potential advantage of capitalism over normal charity usually isn’t enough to overcome my decreased certainty in its efficacy.

It’s just not convincing to say that immediately after an improvised performance of EA maths justifying a water dispenser charity against Instacart…

You can use a separate utility named “AutoRaise” to get autoraise and autofocus functionalities without having to rely on yabai for these. https://github.com/sbmpost/AutoRaise

Good in principle, but bad in the implementation.

From their Markdown Annotations specs [0], the authorship is recorded in a quasi-YAML block at the end of the file, e.g.,—

    Annotations: 0,95 SHA-256 1132bf5e376a605f5beed4b204456114  
    @Human: 0,20 33,4 45,6 62,4  
    &AI: 20,13 37,8 51,11 66,29  

1. it doesn’t work with the existing de facto standard for markdown metadata, i.e., the YAML frontmatters; the markdown standard is already too fractured to have a new extension;

2. it messes with types by using YAML strings for arrays and integers;

3. it’s inefficient — any change in the file would demand iterating through all recorded ranges and updating the offsets.

I’d prefer it to adopt the YAML frontmatters:

  authorship:
    authors:
      - {id: 0, name: Human, hash: xxxxxxxx}
      - {id: 1, name: AI}
    ranges: # adding ids to help resolve potential conflicts
      - {id: 0, author: Human, begin: 0, len: 20}
      - {id: 1, author: AI, begin: 20, len: 13}

I can never understand Stratechery's obsession with overquoting, and it just becomes more unhinged over the years. Every weekly post comes down to something like: hey, have you heard of the latest tech news X, as I wrote long before (super long quotes), and as I read somewhere else (super long quotes), (some piggybacked observations).

If Windows could’ve been dialed back to be as not manipulative as it was at the time of the post’s writing (ie 2k & XP), I would be glad to hold the temptation to self-claim as an expert.

My contrarian view is that the Magic Mouse _can_ be ergonomic as long as you use to be keyboard-centric and treat the Magic Mouse primarily as an extra-large scroll wheel. Then you'll find the scrolling experience with the Magic Mouse is even better than one with a traditional mouse or a trackpad because you can speed-scroll with any finger, at any angle, and with minimal frictions. I especially enjoy scrolling with the right tip of my ring finger.

Thanks! TIL

And the post was translated into English with GPT according to their "about" page. That's why mouthful and unnatural phrases scattered between the lines.

The system prompt used as found in the repo:

``` You are an expert tailwind developer. A user will provide you with a low-fidelity wireframe of an application and you will return a single html file that uses tailwind to create the website. They may also provide you with the html of a previous design that they want you to iterate from. Carry out any changes they request from you. In the wireframe, the previous design's html will appear as a white rectangle. Use creative license to make the application more fleshed out. if you need to insert an image, use a colored fill rectangle as a placeholder. Respond only with the html file. ```

(not sure about why the creative[commons?] license is referred here and why does it help.)

and for each generation the user prompt is:

``` [IMAGE_LINK] Turn this into a single html file using tailwind. ```

https://github.com/tldraw/draw-a-ui/blob/8a889bf36afc06fbb0c...

Looks simple enough to run “privately” by screenshooting a normal tldraw canvas and passing the prompt with it to the API.

For such nuances I keep this page handy: https://hyperpolyglot.org/unix-shells

Thank you for your work at PortableApps. Windows isn’t my daily OS but when I get my hands on it from time to time, the apps you pack are what keep me sane.

I’ve been using DDG’s “bang” feature (search engine shortcuts) for many years and continue to use it after migrating to Kagi. However, the ergonomics of bangs are suboptimal as the leading character `!` takes extra finger travels (on desktop) or taps (on mobile) to input, and there’s no way to customize the catalog of bangs, at least before paying for Kagi.

Therefore I made the simple Node application that emulates bangs with the single variation that the leading character `!` is replaced with `v` (case-insensitive to accommodate iOS’ auto-captalization), which is easier to input on either desktop or mobile. An OpenSearch manifest is included so the site can be discovered by browsers and eligible to be set as the default search engine. You can also supply your own catalog of bangs following the format of DDG’s `bang.js`.

A preview deployment is available at https://vbang2.vercel.app/ and you are invited to deploy your own instance (instructions in readme). I’m completely new to Node so please tolerate sharp corners all over the project :)

My understanding is that the PDF syntax essentially imitates physical printing in that it instructs the reader software to leave something at a given coordinate on a defined media with supplied resources. Thus it's easily portable but barely mutable.

My first thought upon reading this was that it seems rare for me to spend even 47 seconds looking at one screen on my computer without at least glancing at another.

If you waste ~3k words to argue such an evident point as "note-taking tools can't make you smarter", of course you should be surprised if people would spend even 47 seconds skimming through your article.

I think old CLI tools are hard to use because they were designed to be machine- and human-readable at the same time. They are effectively APIs in some scenarios (e.g., scripting and as linked libraries) and in others handholds. As a result they are restricted to be barely viable as either.

The git plugin allows you to disable it on certain devices. I enable it only on my main computer and sync the vault among devices with syncthing, which I find much more reliable than wholly relying on git.

they are apparently delivering emails in tarballs by default from now on

My hypothesis is that they use a set of generic methods (e.g., robot UA, transient cache, and JS filtering) and rely on user reports (they have a tumblr page for that) to identify and manually fix access to specific sites. Having a look at the source of the bypasspaywallclean extension will give you a good idea of most useful bypassing methods. Indeed, most publishers are only incentivized to paywall their content to the degree where most of their audience are directed to pay and they have to leave backdoors here or there for purposes such as SEO.