codefined

We use Docker to deploy, do you know of any public forks that do Docker builds?

Looks fantastic! As someone who has done it before, if this gets popular you'll run into some abusive users that you'll want to deal with. Microsoft will often give out free access to it's PhotoDNA service for detection of explicit images of minors. VirusTotal will often do the same for malware in exchange for samples. You'll also want to have a structured retention process, e.g. size is inversely proportional to storage time.

Good luck, get in contact (see my profile) if you run into any issues.

> I only shared the solution and technique with one other server operator I fully trusted based in the UK

I think that was us! We ended up combining it with other fingerprinting indicators, but the whole 'use VGUI' was a surprisingly effective way at handling this. I believe they removed the web browser in ~2018, which was disappointing. Being able to have custom skill trees / fun integrations with servers was really powerful!

The table name has been changed to "hackernews", although it also hasn't been updated for 7 months.

One of the original devs here. The map is just html with position offsets!

It feels like having a predefined list of keys that are ignored is a "hack". Would a better solution be to simply wait until none of the _trigger_ keys were pressed?

E.g. if the sequence was "Ctrl" + "F8" then it would just ensure that neither Ctrl nor F8 were pressed.

Yeah! Whilst usually a side show to real events there's a bunch of virtual and in person events.

Two years there was an event in the UK dedicated to typing. Last year the US. Virtual tournaments happen frequently throughout the year. The most famous one is likely the yearly event hosted by Keymash[0].

Octahedron is also a famous player in the scene. He hosts a Discord server for proficient typists and regularly puts on events.

[0] https://keymash.io/

Most people typically speak in the 130 - 180 word per minute range, although this can vastly change based on context. Preprepared speeches, arguments and monologues generally have a higher pace than a usual conversation between two people.

> It’s difficult to type and think at the same time, so any time spent writing is a pause in thoughts, which can derail the flow.

I'm a competitive typist. I'm also a programmer. The main positive that I see is that typing no longer requires thought. I think of the code I want to write and it's transcripted onto the computer.

I would say that it's more advantageous to type _accurately_ than quickly. Typing competitions generally require 100% accuracy, which gives me a great amount of belief that what I'm typing is correct, without having to look at it / check it over and get side tracked from the code itself.

I made a very minimal[0] phone only score board. Scores are increased by scrolling, or holding to set it to a specific value.

Pressing the 'settings' tab enables modes for games like 'Whist' that involve betting, as well as configuring the number of players.

[0] https://score.feud.today/

This was really enjoyable, although I found it difficult to read all the letters when it got a little faster! Ended up scoring 170 after a couple of attempts. I looked at the code and it is fairly clean. I'd probably recommend using 'requestAnimationFrame' over the 'setTimeout' you currently use for the loop. I'd also make it clearer that bad accuracy lowers your score!

Vitest[0] is a drop in replacement to Jest that we've had exceptional success with. Faster test runs, more reliable, default TS and ESM support, etc.

[0] https://vitest.dev/

This was incredible feedback, thanks so much for taking the time to write it out. I've implemented 2, 3 and 4. You can now directly link to pages[0] and it'll auto-expand all boxes. The icons are intended to show whether the sequence has a code submission attached. There's now a popover attached to them to explain this.

1 will take some design work, and I've only got an hour or so to work on it now. I definitely agree however that a 'submit' button on a search is essential.

[0] https://oeis.femto.dev/A000045

Source code is available here[0]. Permissively licensed under MIT.

[0] https://github.com/popey456963/pretty-oeis

I keep on meaning to finish a redesign of the oeis site that I started a few months ago[0]. I've only spent a day on it so far, but I'm hoping it makes the site far more approachable to new comers.

[0] https://oeis.femto.dev/

Gitea is self-hosted, so Blender will be running their own instance.

I think it defaults to GMT + 0, aka Senegal / Guinea / United Kingdom.

`/dev/urandom` isn't a real file / stream. It's part of the 'everything is a a file' *nix mantra. Even if two users are reading from /dev/urandom simultaneously, they'll each get unique values. The CSPRNG keeps track of a sequence number and so you'll end up with something like [process 0 requests sequence 0, process 1 requests sequence 1, process 1 requests sequence 2, proceess 0 requests sequence 3...].

It seems only the 'https' site is gone, the 'http' site is still around[0].

[0] http://bugmenot.com/view/news.ycombinator.com

Actually looks to be tabled for TypeScript 5.0, release date of 'March 14th' next year.

The number also appears to include 'upstream' energy usage, things like how much energy was used to transport food to you. How much energy to produce the items you consume. I imagine raw household electrical usage is only a small fragment of this.

1.0.0.1 is owned by Cloudflare and is used for their DNS offering. It's likely a reliable 'ping' candidate for checking that your machine has internet access

[0] https://1.0.0.1/

All those 'kazillion' images are processed into a single 'model'. Similar to how our brain cannot remember 100% of all our experiences, this model will not store precise copies of all images it is trained off of. However, it will understand concepts, such as what a unicorn looks like.

For StableDiffusion, the current model is ~4GB, which is downloaded the first time you run the model. These 4GB encode all the information that the model requires to derive your images.

Nah, something is definitely wrong with the randomness occassionally. See this demo[0] of me reliably getting 30 satire results in a row >75% of the time. The chances of that happening are astronomically low.

[0] https://imgur.com/a/XTQsyNH

The thing that surprises me the most is that one tiny almond takes about 4kg of water to grow. That means that maximally about 0.04% of the water gets used (almond = 1.5g, water = 4000g, 1.5 / 4000 * 100). Avocados, for comparison, are at about 0.5% water usage (1 / 2000 * 100). Apples are at about 0.2% (0.25 / 125 * 100).

Have since found out there's a unified measure, called 'WUE' (Water Use Efficiency), measured in (kg m^-3) water. Some examples include: Almonds, 0.4; Legumes 0.42; Cereals 2.4; Rice 0.73.

Mattermost v7 was released on Thursday and includes video conferencing functionality! [0]

> Channels: Voice calls and screen share (beta)

[0] https://mattermost.com/blog/mattermost-v7-0-is-now-available...

Not seeing it in the article, but how are you handling `index` routing in the `app` folder? If you are putting it in the top directory (e.g. `/app/index.js`), is it not able to get a custom `layout.js` file (as that position is used for the global variant)?

If it's in an `index` folder (e.g. `/app/index/page.js`), how would one create a route for `example.com/index`?

> Clarity is built by Microsoft, one of the largest technology companies in the world. Microsoft processes a massive amount of anonymous data around user behaviour to gain insights and improve machine learning models that power many of our products and services. Clarity is one of the ways Microsoft gathers this important data – and why we've made it available for free.

Looks like this is a data gathering effort by Microsoft.

    let { value, weight } =
        if values.empty() {
            { value: 0, weight: 1.0 }
        } else {
            { value: values[0], weight: 1 / values.len() }
        }

Using JS style object expansion, instead of tuple expansion.

Looking at the Dutch site on my phone (Samsung S10) I noticed it took a little while to load compared to the nigh instant loading of the UK Gov variant. Looking at Page Insights [0] [1] tells a similar picture. Desktop time to interactive times of 0.4s Vs 3.5s and Mobile time to interactive times of 4.5s Vs 13.1s.

The Dutch website seems to spend a lot of that time running the Next JS framework stuff, which the Gov.uk variant does not. It might work quickly on fast computers, but even on modern phones it seems to visibly pause.