HellsMaddy

The title accurately describes what's interesting about the submission, no hyperbole or clickbait. What's wrong with that?

Yeah, if you use the Bluetooth toggle in control center, it doesn’t disable Bluetooth at all, it just temporarily disconnects all Bluetooth devices. If you go to your Bluetooth settings, it will still be enabled.

For anyone who is still confused: https://en.wikipedia.org/wiki/Credit_card_imprinter

I'm curious: When updating packages, say NPM packages, what are your practices?

For us, we audit every update of every package. We read the changelog or release notes for every version from the current version to the new version (some packages make this very cumbersome, looking at you Vercel). Then, we use our judgement about the order in which to update packages; usually we apply all patch-level updates, run our tests; then minor updates, re-run our tests; and then we apply major updates one-by-one and re-run tests after each of them.

We've found plenty of semver violations in this process, to the point where we don't really trust the semver of NPM packages at all.

Is this a common approach, or do you just blindly update everything at once and fix things if they break?

Someone created a userscript to restore it, works great for me so far: https://github.com/Gerrit0/old-github-feed

Someone created a userscript to restore the old feed, it seems to work perfectly so far:

https://github.com/Gerrit0/old-github-feed

You were just being true to your name, by offering your revision history!

On a more serious note, this is a great example of how to handle a vulnerability report - fix it, change your processes, and say thank you! (geek_at could probably have done better by disclosing this in private first, though)

Give users a choice - do they want to draw a frame (will take a couple of minutes, many people don't have time/patience) - or do they want to vote on a few frames?

For each frame, get at least 4 different drawings.

For voters, show them 2-4 frames and have them pick the best one.

If there are more than 4 versions of a frame, do a bracket-based competition to determine the best frame.

This could be an ongoing process. Generate a version of the movie using the best available frames, but continue allowing people to redraw frames. The movie would continue getting better and better over time.

Then, when you watch the movie, maybe you could have a slider to control which "generation" of the movie you're watching. Slide all the way to the left and it will be really primitive, slide all the way to the right and it will be the best version.

I applied for access to Claude months ago, any suggestions on getting into the trial?

This isn't a LUKS bug. It's a bug in a completely different piece of software that most LUKS users have never heard of, which people sometimes use to unlock LUKS volumes.

Aaron, thank you for this course! I started watching it about a week ago, and I love your teaching style.

I've been using relational databases for years, but my conceptual understanding was never that strong. I hadn't touched MySQL in particular for several years. Your course not only got me back up to speed, but I learned a lot of things I'd never known before.

I love how you organize your lessons, how you make sure to answer the right questions at the right times, and how you prioritize helping the learner build a strong mental model.

I hope you make more content like this in the future.

I'm not a YouTuber, but I found Theo Browne's "How The YouTube Algorithm ACTUALLY Works"[0] to be pretty intriguing.

I watched it a few months ago and just re-skimmed the transript, so I'm not sure if I'm remembering it perfectly, but here's what I took away from it:

- YouTube's algorithm recommends videos based on audience viewing patterns.

- The algorithm doesn't care what the video is about, it only cares about who is clicking on it and watching it.

- Click-through rate and watch time are the most important metrics, they're how the algorithm determines who to recommend your video to, and whether to recommend your video to a wider audience.

- It first shows your video to a small group and if it's well-received, expands to a wider audience with similar viewing habits.

- It repeats this process until it finds the falling-off point, where the click-through rate and watch time drop off.

- Figure out who your niche audience is and make videos that they're likely to click on and watch.

- You're going to need to create enough videos in order for the algorithm to figure out where your audience is, and then you're going to want to make more videos for that audience.

- The average clickthrough rate you see in your analytics is not the whole story. You might have a 20% clickthrough in your target niche, but if the algorithm starts recommending your videos to a wider audience, your click-through rate is going to start to drop. That doesn't mean your video is not doing well, it could actually be doing well enough with your niche that the algorithm is trying to show it to more people outside of your niche.

[0] https://www.youtube.com/watch?v=58jv8WL6H9I

So was xrOS a red herring?

I have wanted something like this for years, and your take on it is way better than what I had in mind. I think the pricing is fine, I plan to pay for a month and see if it lives up to my hopes. Awesome job on the marketing page and demo video.

Is there anything like this for C/C++ applications?

Interesting:

    // we only keep unfollows in the past 90 days due to the huge size of this dataset,
    // and to prevent permanent "shadow-banning" in the event of accidental unfollows.
    // we treat unfollows as less critical than above 4 negative signals, since it deals more with
    // interest than health typically, which might change over time.
    val unfollows: SCollection[InteractionGraphRawInput] =
      GraphUtil
        .getSocialGraphFeatures(
          readSnapshot(SocialgraphUnfollowsScalaDataset, sc),
          FeatureName.NumUnfollows,
          endTs)
        .filter(_.age < 90)

https://github.com/twitter/the-algorithm/blob/main/src/scala...

For how long does the spray protect you? Is this something you'd use on a daily basis?

It’s really not that bad. You can sign up for a free Mozilla developer account and upload the bundle you’d like to sign, and install the signed extension a few minutes later on any edition of Firefox.

Yes, check out LangChain [0]. It enables you to wire together LLMs with other knowledge sources or even other LLMs. For example, you can use it to hook GPT-3 up to WolframAlpha. I’m sure you could pretty easily add a way for it to communicate with a human expert, too.

[0]: https://github.com/hwchase17/langchain

> Alfred Security also make a impressive, very inexpensive camera ($34 US, 1080P, night vision, 130 degree viewing angle and easy to mount).

I’m interested, but I can’t find anything about this on their site. Can you provide a link?

> When people stop and think about factory farming, they hate what they think about. The trick most people employ is to not think about it.

Exactly, this is known as "cognitive dissonance".

It might feel different to a human, but it’s the same for the animal either way. Similar to hiring a hitman, it’s no more moral to kill at an arm’s length. I believe it is less moral because there’s an additional victim (the person performing the killing).

> In some of my earlier examples, the model correctly predicts the make/model of cars that are only partially visible in the photo. To study how far this can be pushed, I panned a crop along a side-view of two different cars to see how the model’s predictions changed as different parts of the car became visible. In these two examples, the model was most accurate when viewing the front or back of the car, but not when only the middle of the car was visible. Perhaps this is a shortcoming of the model, or perhaps automakers customize the front and back shapes of their car more than the sides. I’d be happy to hear other hypotheses as well!

I would hypothesize this is an artifact of the training data being photos taken for the purpose of selling the cars.

As a seller, it doesn’t make a lot of sense to take a photo of the middle of the car, unless a) you’re trying to hide something or b) you’re being lazy (implying the car isn’t worth much of your time). I would imagine that better photography in general - composition, framing, lighting, setting - would be correlated with higher car prices.

I created a similar tool to help me synchronize my clipboard between my desktop and my laptop: https://github.com/b0o/clipshare

I love the pure function idea.

“How the dairy industry designs misleading studies” - https://youtu.be/VYHiT7wJFk0

As a privacy-concerned user, a bookmarklet may be preferable because it's only going to execute code when you explicitly request it, rather than an extension running all the time.

I developed an extension[1] after the Chrome Store stopped accepting MV2, but before Firefox support for MV3 in stable. I had to write some custom build scripts to build the extension with MV2 for Firefox and MV3 for Chrome. Pretty annoying.

[1]: https://github.com/b0o/aws-favicons-webextension

Seems lying them flat also makes their cleaning robot able to easily maneuver, meaning they don't need to leave any space in between panels for humans to perform maintenance. Pros: reduces land usage as you mention, but also less humans needed for maintenance.