sheept

For your knowledge cutoff test, did it failing mean that it generated a generic "Kpop demon hunter" or it rejected the prompt?

fyi, you can also see the author email by appending ".patch" to the end of a commit URL

I feel like Deno would be perfect for this because it already has a permissions model enforced by the runtime

2k feels very small considering the number of business sites that embed Google Maps. I guess a lot of those sites use other website building services that handle the Google API keys for them, and/or they're old and untouched enough that no one enabled Gemini on them.

It was revealed by this post by Wiz from the beginning of this month: https://www.wiz.io/blog/exposed-moltbook-database-reveals-mi...

I think to figure out if it's a cultural difference or unwillingness for technological change, it'd be interesting to compare Korean or better yet Chinese apps, which were designed more recently.

In my brief experience,

- Promotional websites (e.g. https://ant.design/index-cn, https://seed.bytedance.com/zh/seedream4_0) seem to be designed similarly to Western websites

- Chinese mobile apps do seem to be more colorful in general (as observed with Japanese websites in the article), and some are more information dense, like Douyin Shop and Live. Bullet comments also add to the density

- I briefly worked at TikTok US, where the company uses Lark Suite. The desktop and mobile interfaces look pretty similar to Notion/Slack, but it has more vibrant colors and slightly more features (being an everything app for the workspace)

Then under your definition there must not be any widely used written language with an alphabet. Most of the world's alphabetic writing systems aren't phonetic transcriptions, they're standardized. They're usually based on the prestige dialect, at the cost of diminutizing other dialects.

For example, Spanish has a fairly consistent spelling system standardized by RAE, based in Madrid. But, for instance, even though much of Latin America doesn't have a distinction between s and soft c (seseo), they still keep the distinction in its spelling.

phonetic spelling based on whose dialect? should "merry" "marry" and "Mary" be spelled the same?

besides, pronunciation continues to evolve, so any phonetic spelling would continue to gradually diverge from the spoken language

Class components still work, and you can still use function components with hooks inside class components and vice versa.

In the parent comment's case of not having other dependencies, whatever the React ecosystem does isn't relevant if you aren't using any React libraries, which aren't really necessary anyways, especially nowadays when the LLM can reimplement what you need for you.

Nothing has changed about react-dom that prevents you from using React piece by piece—its docs still recommend attaching to a #root node even for single page apps.

Including web components in a React app is very seamless, and embedding non-React-controlled elements inside React is not uncommon (e.g. canvas, Monaco, maps), though for common use cases there's usually convenience libraries for React that wrap around these.

Taiwan is the country that uses "mainland" (大陸 dalu) to refer to China

Based on where other companies are right now, it's probably a pre programmed routine, but the robots autonomously balance themselves. Anything beyond that would be quite a large leap, and I think Unitree would've gloated about it way more outside of the gala. The robots' speed and consistency are still impressive, though.

I think the ping pong match video might be misleading you. Based on the visual artifacts around the robot, the original footage likely had a human player that was swapped in with a robot in this video. It also has an altered content warning.

just to note, signaling has two L's in UK spelling

The headline implies they selectively removed the word "safely," but that doesn't seem to be the case.

From the thread you linked, there's a diff of mission statements over the years[0], which reveals that "safely" (which was only added 2 years prior) was removed only because they completely rewrote the statement into a single, terse sentence.

There could be stronger evidence to prove if OpenAI is deemphasizing safety, but this isn't one.

[0]: https://gist.github.com/simonw/e36f0e5ef4a86881d145083f759bc...

I have a userscript that does this, but one downside is that it refreshes the page, so if I go back in history (eg I misclicked), it resets the recommendation page

It seems to be a common issue, and despite googling I wasn't able to find a solution that worked (back in Aug '25). For some reason, it does not happen if you position the dock on the left/right side rather than the bottom

Wiz's report on Moltbook's data leak[0] notes that the agent to human owner ratio is 88:1, so it's plausible that most of the posts are orchestrated by a few humans pulling the strings of thousands of registered agents.

[0]: https://www.wiz.io/blog/exposed-moltbook-database-reveals-mi...

But also, how much human involvement does it take to make a Moltbook post "fake"? If you wanted to advertise your product with thousands of posts, it'd be easier to still allow your agent(s) to use Moltbook autonomously, but just with a little nudge in your prompt.

Based on the time range, the decline of social media use after 2020 could be more strongly related to the decline of Covid and remote school/work

my high school still uses simply scheme and SICP for the fall semesters of its CS sequence I believe, but granted it's in an affluent city in the Bay and its recently retired teacher came from UCB

I do not think it'd be "so easy" to separate password input access into a separate permission because it'd only open up a can of worms. There's so many ways to read a password input's value, from listening to key events to monkey patching `fetch`, that it's not worth playing whack-a-mole just to provide users a false sense of security

I'm also skeptical that even a dark mode extension would be simple considering how varied web pages can be

dictionaries generally aren't guaranteed to contain an entry for every possible value of the key type. while you could implement the colors example with a dictionary, ideally you'd want the type system to assure that given a Color, there will be a string associated with it

in most cases you should be able to create an account and edit even if your IP address or range is blocked

our university has been rolling out just walk out markets across campus due to rampant stealing. shopping there doesn't feel like stealing, but the store design feels oppressive with racks of cameras and thick black shelves because it's designed for sensors first not humans

one minor downside (especially since I don't live on campus anymore) is that in order to walk around and peruse the shelves, I have to give them my payment info just to enter

AI is capable of consistent characters now, yes, but the platforms themselves provide little incentive to. TikTok/Instagram Reels are designed to serve recommendations, not a user-curated feed of people you follow, so consistency is not needed for virality

a reliable giveaway for AI generated videos is just a quick glance at the account's post history—the videos will look frequent, repetitive, and lack a consistent subject/background—and that's not something that'll go away when AI videos get better

it's because a year/month/date isn't enough to uniquely identify an instant in time. you can create a Temporal.PlainDate(Time) with those values though, then convert to the other types depending on what you need and it needs (e.g. time zone)

(CommonMark) Markdown is a rough superset of HTML, like how YAML is a superset of JSON. So whatever can be expressed in HTML can also be expressed in Markdown. With the way the CommonMark spec is written, Markdown is effectively just an HTML preprocessor.

A major limitation of Markdown is the lack of standardization. For example, even within GFM there's multiple subtle variants: a single new line becomes a space when rendering Markdown files, but a line break in issue comments.