ahrjay

I had a go at this using the on-device models in edge and chrome, phi4-mini and gemini nano, worked surprisingly well for such small models.

https://ryanseddon.com/ai/how-to-build-an-agent-on-device/

Yeah the instructions are not clear. They're on the github repo[1] linked in the header.

1. Install Chrome Dev: Ensure you have version 127. [Download Chrome Dev](https://google.com/chrome/dev/).

2. Check that you’re on 127.0.6512.0 or above

3. Enable two flags: chrome://flags/#optimization-guide-on-device-model - BypassPerfRequirement chrome://flags/#prompt-api-for-gemini-nano - Enabled

4. Relaunch Chrome

5. Navigate to chrome://components

6. Check that Optimization Guide On Device Model is downloading or force download if not Might take a few minutes for this component to even appear

7. Open dev tools and type (await ai.languageModel.capabilities()).available, should return "readily" when all good

[1] https://github.com/ryanseddon/FFprompt

I built https://ffprompt.ryanseddon.com using the chrome ai (Gemini nano). Allows you to do ffmpeg operations on videos using natural language all client side.

Ah bummer I've been posting my earthin24 timelapses[1] to this for quite a while now.

[1] https://botsin.space/@earthin24

In my home state of Victoria Australia the government had a program to give out these powerpal[1] units for free that could measure your usage in realtime using the flashing led on our smart meters, we also require all energy grid operators (the people who own the poles and wires) to have an energy portal where users can get near realtime data to the nearest 30mins, soon to be 5 with some new legislation.

The former most people have no idea about but the powerpal has been a smashing success for consumers to understand what is using energy.

[1] https://www.powerpal.net/

I've got an oven and induction cooktop freestanding unit[1] that has big chunky knobs to change the induction power levels. Would never bother with gas again.

[1] https://www.fisherpaykel.com/au/cooking/freestanding-cookers...

I did something similar using filter_complex to create a 14x14 grid showing 196 days of earth full disc shots for my earthin24 Twitter bot. It's truly impressive what ffmpeg can do https://ryanseddon.com/javascript/an-earth-mosaic/

I combine this and 3 other satellites on my twitter bot into daily videos if you're interested.

https://twitter.com/earthin24

I have an original copy of this game in my garage! Including the vmu microphone.

Hopefully the next step is to take image decoding off the main thread.

According to a few people both IE11[1] and Firefox[2] disable hit testing on scroll if you don't move your cursor while scrolling.

[1] http://www.thecssninja.com/javascript/pointer-events-60fps/c... [2] http://www.thecssninja.com/javascript/pointer-events-60fps/c...

It's supported in WebKit nightly too.

Source maps recommend a similar technique to mitigate against XSSI attacks by prepending ")]}" to your source map http://www.html5rocks.com/en/tutorials/developertools/source...

I couldn't even use it on my phone, I think the viewport meta set to device-width makes it load zoomed in and the interaction with the content makes it impossible to zoom out.

I built something along the same lines a while back that allows you to drag and drop font files into the browser. http://fontdragr.com

Also has a bookmarklet so you can test them on your own site.

Don't forget source map support, that is a huge plus to have from day one.

I built a simple cli tool to do cross browser unit testing.

bunyip: http://ryanseddon.github.com/bunyip/

The ontouchstart issue can be avoided by using the ontouchmove/end to check of the user is scrolling. Infact zepto.js has a custom event called "tap" which does exactly that.

Nice source map support will be huge.

Firefox and Opera support the APNG format and there is a clever polyfill[1] to add support to all other browsers.

[1] https://github.com/devongovett/png.js

Of course web devs use them, but like you said it's a waste of bandwidth so the production version should be minified and combined.

There is the do not track initiative[1] but it's kind of a pretty please solution.

[1] http://dnt.mozilla.org/

It's not being "ultra-strict" it's a requirement of the spec[1] that the file be served with the correct mimetype.

[Edit] It doesn't work in firefox, the attribute triggers the permission bar but because it's the wrong mimetype it doesn't actually store anything.

Check it in Tools > Options > Advanced > Network

You'll notice that it has the domain referenced but has 0 bytes stored.

[1] http://www.whatwg.org/specs/web-apps/current-work/multipage/...

Valeri Vladimirovich Polyakov spent 437 days in the Mir space station in a similar experiment.

Looks good, I did notice that the appcache has the wrong mime type it's text/plain and needs to be text/cache-manifest.

Author here.

If you're interested in hacking on this project it's available on github to look at.

https://github.com/ryanseddon/font-dragr

You may be the same person who mentioned on twitter about this. That error should only happen when dropping some data that isn't in the correct format, it should never get to that part of the code if you're dropping a font file from your file system. Do you mind creating an issue on the github repo[1] with some info?

I've tested on ubuntu 11 with FF5 and all seems well, perhaps FF6 may be doing something to cause it.

[1] https://github.com/ryanseddon/font-dragr/issues

Block scoping is part of ES5 using let

https://developer.mozilla.org/en/JavaScript/Reference/Statem...

The selector engine runs right to left so even if you start with an id but end it with an element it will still be an expensive query. I think the key is to think about your selectors and not go overboard because it's easy.

Here's some interesting observations, this will work on any document that is in designMode or any element that has contenteditable attribute.

Firefox doesn't have the clipboardData interface like IE, Chrome and Safari have so it must be built into Firefox. From a quick look it seems like it doesn't expose the stream data to JavaScript which is unfortunate, I would love for this to be available.