FrostKiwi

The page looks Claude generated and ranks Opus at the top.

> On the other hand, looks like WebXR support for Safari will never come.

Thanks for checking. What a mess. I'm so sick of the workarounds that have to make.

Interesting, I had Fable generate me a summary using an identical style, with the same square design, bg and yellow highlights

FYI, setting `NvOptimusEnablement` and `AmdPowerXpressRequestHighPerformance` have been the canonical way on Windows, `__NV_PRIME_RENDER_OFFLOAD` + `__GLX_VENDOR_LIBRARY_NAME` on Linux. Though not an OpenGL feature per se, as you mentioned.

Same comment, same wording was had when going from H.264/AVC to H.265/HEVC and again to AV1. Across the board it's 30% better Size vs Quality and with an X amount increase in compute required to encode and decode.

Each time these standards were put into ink, they were years from being practical.

Self hosted gitea for many years with ~25 devs. Yes, it's essentially a FOSS carbon copy of GitHub. CI/CD is also intercompatible, uses the same syntax and pulls the original GitHub Actions packages. Now with the Forgejo split, I would prefer Forgejo, as it has way more steam behind it with Codeberg and Blender as the big use-cases.

The goblins evolved it seems

So So useful <3 Stabilization = you have to zoom-in, loss of FOV. Depending on Action Cam and undistortion parameters, this can be different, sometimes too little, sometimes too much. Gyroflow allows you to dial it in. Lot's Stabilization in a particularly shaky spot and widest FOV everywhere else, smoothed between to be unnoticeable.

Demo throws `can't access property "isInterleavedBufferAttribute", e is null` on firefox but work on chromium based browsers.

> But a compiler is deterministic — same input, same output. An LLM isn't.

Temperature 0 determinism is subject to active research. NVIDIA tried but failed so far, DeepSeek V4 seems to have done it. I hope judges won't be swayed by this an AI generated code will classified as uncopyrightable, just like Images are.

Actually blown away by how good that works with my 1920x1200 8-bit HP Z24n G2 monitor [1] and my Laptop's 1920x1080 6-bit Panel [2]

[2x5-HP-Z24n-G2] https://i.imgur.com/yLyrpfg.jpeg

[1x5-HP-Z24n-G2] https://i.imgur.com/Z7kH005.jpeg

[2x5-Innolux-N156HCA-GA3] https://i.imgur.com/F4Ypxwj.jpeg

[1x5-Innolux-N156HCA-GA3] https://i.imgur.com/etkot5o.jpeg

[1] https://jp.ext.hp.com/monitors/business/z_z24n_g2/

[2] https://www.panelook.com/N156HCA-GA3__15.6__overview_33518.h...

You are of course right. It's just that Modules inherently put a lot of responsibility on the build system. Among those, but not limited to: a "module registry" wasn't standardized and is in the hands of the build system.

Systems like ninja needs to know modules, which took time and then a stack further up systems like CMake needed to know modules, which took time. That's my answer to the parent "why are there so few modules projects". Because it took time for the ecosystem to catch up.

That's really interesting info, thanks!

It's the fault of built systems. CMake still doesn't support `import std` officially and undocumented things are done in the ecosystem [1]

But once it works and you setup the new stuff, having started a new CPP26 Project with modules now, it's kinda awesome. I'm certainly never going back. The big compilers are also retroactively adding `import std` to CPP20, so support is widening.

[1] https://gitlab.kitware.com/cmake/cmake/-/work_items/27706

But that fresh UUID is in the prompt.

Also it's missing the point of the parent: it's about concepts and ideas merely being remixed. Similar to how many memes there are around this topic like "create a fresh new character design of a fast hedgehog" and the out is just a copy of sonic.[1]

That's what the parent is on about, if it requires new creativity not found by deriving from the learned corpus, then LLMs can't do it. Terrence Tao had similar thoughts in a recent Podcast.

[1] https://www.reddit.com/r/aiwars/s/pT2Zub10KT

> is to spend at least as much as our salary

Reads as a very distopian "let's see how many people we can replace"

Same here (10s to 1s). The main reason for this is rolldown [1]. Already had it installed months ago, before it got merged into vite proper. Really awesome stuff.

[1] https://rolldown.rs/

I do not trust that, similar working was used when Sonnet 1M launched. Still not the case today.

> and it won't work without Steam running. Such a disappointment

I assume it will be like Steam Controller 1: Given no Steam and no special driver, the Controller is a Mouse + Keyboard, also referred to as "lizard mode".

I am also anti-DRM, but I don't think this can be solved easily. Consider the Dual Shock 2: Either it's explicitly supported or requires a custom diver to emulate into XInput or DirectInput. Even using XInput directly is cross platform a driver nightmare. Valve has done good work I think with their libSDL based Steam Overlay, becoming a kind of universal Input equalizer, going so far as to patch their games with updated tutorial input prompts based on controller like Dual Shock 2 vs XBox Controllers.

A firmware level solution is not really realistic at this point. Controller Manufacturer 8BitDo went this approach, with many device restart firmware modes per target platform. It's just not a good user experience.

XUASTC was just released, allowing for all ASTC Block sizes to be used on the web, once support lands in the big web 3D libraries like Three.js [1].

To me, this is huge. Relying on pre-calculated lighting when building stuff like [2], I constantly fight for VRAM, as older iPhones have very strict limits. XUASTC brings the all the remaining modern GPU compression schemes to the web in a standardized manner.

[1] https://github.com/mrdoob/three.js [2] https://mymachine.dmgmori.com/

No way, it's BPOS all over again [1]

[1] https://en.wikipedia.org/wiki/Microsoft_Online_Services#Prod...

This is astronomical. If I correctly understood, the full on compromise of Ubisoft happened because of this.

Thanks to libdav1d's [1] lovingly hand crafted SIMD ASM instructions it's actually possible to reasonably playback AV1 without hardware acceleration, but basically yes: From Snapdragon 8 onwards, Google Tensor G3 onwards, NVIDIA RTX 3000 series onwards. All relatively new .

[1] https://code.videolan.org/videolan/dav1d

A Valve employee using YouTube Playthroughs [1] to diagnose a bug is hilarious to me. Awesome story.

[1] https://mastodon.gamedev.place/@TomF/115589894339657055

Very Microsofty. I'm still furious with how they handled the global ban on updating CDN (Frontdoor) files until last week and not properly communicating it. https://www.reddit.com/r/AZURE/comments/1on0ung/azure_status...

Working fine in Japan as of 1:21 AM UTC

Running NixOS based Nextcloud for everything. Multiple family members getting their including me getting their Photos and Videos auto-uploaded and handled by NextCloud's Memories [1].

Awesome experience, but you really have to stick to the happy path. Even with a super powerful CPU, Videoplayback was unusable, until a dedicated AMD GPU handled the transcodes. Even though it's Fiber To the home, sometimes Upload speeds collapse for no apparent reason and it's unusable.

All in all impressive such a massive FOSS project runs at all.

[1] https://memories.gallery/

Ohh, just reread the incident report:

> "Customer configuration changes to AFD remain temporarily blocked. We will notify customers once this block has been lifted. [...] we are still working to mitigate this long tail."

Are you kidding me?! How can they claim this incident as resolved?

Surprised to see the situation getting worse, what the hell.

Had some Frontdoor operations timing out, but now I'm straight up denied with "Message: All Changes to Azure Frondoor Configuration are blocked currently."

What a mess.

Surprised to see the situation getting worse, what the hell.

Had some Frontdoor operations timing out, but now I'm straight up denied with "Message: All Changes to Azure Frondoor Configuration are blocked currently."

What a mess.