Google Open Source Blog

The amazing thing about JPEG is people are still squeezing out backwards compatible compression improvements 30 years later. Mozjpeg is well known by now, but Jpegli was just published last year and does even better at high bitrates[1]. It's hard to want to adopt newer formats when they keep getting squeezed by further improvements to good old JPEG. [1] https://opensource.googleblog.com/2024/04/introducing-jpegli-new-jpeg-…

Worth noting that production silicon is on the way for OpenTitan (https://opensource.googleblog.com/2025/02/fabrication-begins-for-produ… An OpenTitan chip in a future TKey can give both strong hardware security (you'll have real silicon resistant against hardware attacks, not just an FPGA) whilst remaining full open source hardware (you can see all the RTL here: https://github.com/lowRISC/opentitan/tree/earlgrey_1.0.0 available under an Apache 2.0 license).

Not exactly, although CUDA is a huge topic. See https://gist.github.com/shawwn/0e524d4a7a5d8fb152a86616559cc02a for some description of the process. Basically, jax or other program is converted to XLA, https://opensource.googleblog.com/2023/05/pjrt-simplifying-ml-hardware… then lowered to the specific architecture (which coudl be CPU, GPU, or TPU). Last time I looked, it was a horribly complicated stack with many parts changing rapidly, although with the switch to jax, things got cleaned up a bit. My personal favorite bits are the lower levels of jax, xla, and pjrt.

Would you recommend a different distribution mechanism? The Apple binaries are all signed (in accordance with Apple policies), and the team has historically invested significantly in supply chain security. e.g. (a now 2 year old article) https://opensource.googleblog.com/2022/09/flutter-slsa-progress-identi…

It would be great to understand what you mean by this -- we have a deep love for open source and the open developer ecosystem. Our open source team also released a blog today describing the rationale and approach for open models and continuing AI releases in the open ecosystem: https://opensource.googleblog.com/2024/02/building-open-models-respons… Thoughts and feedback welcome, as always.

Go was very much intended as a systems programming language. It's literally in the initial announcement[1]: "Go is a great language for systems programming with support for multi-processing, a fresh and lightweight take on object-oriented design, plus some cool features like true closures and reflection." The whole cloud bonanza came later. A lot of Go programmers I know also don't really like it (myself included). [1]: https://opensource.googleblog.com/2009/11/hey-ho-lets-go.html

>they should have two formats anyway Good thing JXL has the best progressive support of any image format available! [0, 1] Something which you can only find partially with incremental decoding in WebP and doesn't really exist in AVIF (just one of the many limitations of being bolted onto a video codec). You might often reconsider having both versions of a file with this capability. (which wasn't exploited much or at all with "dumber" and heavier progressive JPEGs) 0. https://opensource.googleblog.com/2021/09/using-saliency-in-progressiv… 1. https://jpegxl.io/articles/faq/#doesjxlsupportprogressivedecoding?

For maybe a single weeks worth of investment you could pick up Dart (the learning curve is tiny) and have that exact same value proposition of write once and run anywhere but with many other substantial advantages on top of it. Here’s a completely arbitrary one from just yesterday that focuses on supply chain security (https://opensource.googleblog.com/2022/06/Dart-and-Flutter-enable-Alls… but this kind of thing is happening on an almost weekly basis in the background around Dart and Flutter. I don’t think people understand what’s happening behind the scenes on these projects but they are developing some very solid fundamentals to build on top of into the future. Seriously take a look at this basic language tour, it’s incredibly familiar for anyone with an OO background out of the box https://dart.dev/guides/language/language-tour

Are people still getting hacked by Mad Gadget? I tried my best to patch everything. https://opensource.googleblog.com/2017/03/operation-rosehub.html You can do that too. We could point the finger at Java. But it's more impactful if we just pick up a keyboard and fix things.

For what it's worth, I reached out to the Open Source Security Foundation (OpenSSF) after their press release on "Securing Critical Open Source Projects" [0], because my open source project is on their "critical list" (Python top #200) [1]. After some clicking around, I was directed to the OpenSSF mailing list [2]. …where my request was left to rot, without any response. Made me wonder whether this is just a PR stunt / corporate power grab. [0] https://opensource.googleblog.com/2020/12/finding-critical-open-source… [1] https://news.ycombinator.com/item?id=25381397 [2] https://groups.google.com/g/wg-securing-critical-projects

Except original and fake-original is not hidden. Try opening https://opensource.googleblog.com/ https://security.googleblog.com/ Both opensource and security are shown. Disc: Googler but don't work on this project.

you're right, the following url holds a good writeup about the intensions of developing grumpy https://opensource.googleblog.com/2017/01/grumpy-go-running-python.htm…

I've met Mike Pall in person. I've worked with him some since I helped facilitate a sponsorship from Google (https://opensource.googleblog.com/2010/01/love-for-luajit.html). A lot of his work on LuaJIT has been supported by various sponsorships (see: https://luajit.org/sponsors.html). This gives him a lot of freedom to work the way he wants to, instead of working as part of a larger organization. Also note that the sponsorship page says that Mike is currently working on "unrelated projects", so apparently things besides LuaJIT are keeping him busy now.

Congrats on the milestone! I've been following for a while - awesome to see the progress you've made. Q: What are your plans for sharding / scale-out? At Notion, we run ~480 logical schemas spread across ~32 Postgres databases [1]. Our data model [2] (sorry for the blog post spam) has a recursive / graph-like structures, and we make use of columns or jsonb attributes like ` { table: Table, id: UUID }` which sounds like your polymorphic links feature. It seems like EdgeDB's model lines up well with how we already use our database. I did a quick cmd-f here, on the linked announcement, and in your docs looking for "shard" and "scale" but didn't find any relevant results. Postgres needs a Vitess! Q: Do you have plans to support EdgeQL embedding or SQLite? I am always looking for a way to compile "better than SQL" languages down to SQL. I like Datalog in this area because it's a composable way to define relationships/facts/derivations but no one is putting serious business effort to this idea (honorable mention to logica [3]). EdgeQL also fits the bill -- queries aren't logical, but they are composable -- plus looks easier to teach than most Datalog variants. I took a peek in the repo and saw that a few EdgeQL components are written in Rust [4]; are you considering porting more logic to Rust? That would make EdgeQL much more embeddable - it could run in WASM or linked into an Android/iOS app. My pie in the sky dream is to use a single composable query/logic language to…

See https://opensource.googleblog.com/2025/11/secure-by-design-firmware-de… This looks like it has a smaller code footprint at least. I'm not sure RISC-V is a very good target for this sort of thing. E.g. decoding the immediates in software is going to be very slow, whereas in hardware it's fast. But on the other hand it is a stable target and can be configured to be a lot simpler than WASM.

They’ve already written about it? https://security.googleblog.com/2023/09/scaling-rust-adoption-through-… https://opensource.googleblog.com/2023/06/rust-fact-vs-fiction-5-insig… And other posts.

I always thought Google had some interesting insights and ideas when they were playing around with the now abandoned service weaver project: https://opensource.googleblog.com/2023/03/introducing-service-weaver-f… More academic and language neutral introduction here if that’s your thing: https://arxiv.org/pdf/2404.09357

This article* makes the case for mozjpeg cleanly beating webp when we are above 500x500px image sizes. So. There's a lot more performance/compression to be gained within the jpeg container format than people generally argue for. * https://siipo.la/blog/is-webp-really-better-than-jpeg edit: https://opensource.googleblog.com/2024/04/introducing-jpegli-new-jpeg-… is likely the real GOAT when it comes to modern jpeg encoders in that it effectively breaks the 8bit color space "ceiling" within the format!

> New Relic did this for the Rails revolution, Datadog did it for the rise of AWS, and Honeycomb led the way for OpenTelemetry. I find this reading of history of OTel highly biased. OpenTelemetry was born as the Merge of OpenCensus (initiated by Google) and OpenTracing (initiated by LightStep): https://opensource.googleblog.com/2019/05/opentelemetry-merger-of-open… > The seed governance committee is composed of representatives from Google, Lightstep, Microsoft, and Uber, and more organizations are getting involved every day. Honeycomb has for sure had valuable code & community contributions and championed the technology adoption, but they are very far from "leading the way".

google announcement: https://opensource.googleblog.com/2025/01/see-code-that-powered-pebble… pebbleOS repo: https://github.com/google/pebble

also check out jpegli. https://opensource.googleblog.com/2024/04/introducing-jpegli-new-jpeg-…

Technically it also had a fourth :^) [0] but it was spun out into a separate project of its own, jpegli [1]: JPEG but it uses some tricks from JPEG XL. These include spatially adaptive quantization, quantization matrices that better preserve psychovisual detail, more efficient color spaces, and also HDR (10+ bit depth) support [2]. [0] https://github.com/libjxl/libjxl/tree/main/lib/jpegli [1] https://github.com/google/jpegli [2] https://opensource.googleblog.com/2024/04/introducing-jpegli-new-jpeg-…

Without any ARM specific optimizations, most apps wouldn’t even have to recompile and resubmit. Android apps are uploaded as bytecode, which is then AOT compiled by Google’s cloud service for the different architectures, from what I understand. Google would just have to decide to support another target, and Google has already signaled their intent to support RISC-V with Android. https://opensource.googleblog.com/2023/10/android-and-risc-v-what-you-…

There’s also open-source, standard cells for people interested. Here’s some links: https://www.vlsitechnology.org/html/libraries.html https://opensource.googleblog.com/2022/07/SkyWater-and-Google-expand-o…

> Qualcomm does not have any pull to make Google to make RISC-V Android a first class citizen. You can debate what "first class" means, but RISC-V was announced as an official Android platform in 2023, and runs in emulators today: https://arstechnica.com/gadgets/2023/01/google-announces-official-andr… https://opensource.googleblog.com/2023/10/android-and-risc-v-what-you-… https://github.com/google/android-riscv64

>is that even it's design purpose? It is not. The closest thing you'll get to it that has the features you mention, is Qubes OS [0]. When it comes to the L4 family, Google's adoption of seL4 [1] might be interesting to you, and possibly L4Re [2] [3], which is open source (not their tooling, though) with commercial support. Edit: there's a bunch more open source activity in this field, but I'm not familiar with it. [0] https://www.qubes-os.org/ [1] https://opensource.googleblog.com/2022/10/announcing-kataos-and-sparro… [2] https://l4re.org/ [3] https://www.youtube.com/watch?v=x8jxINR-q40

My recollection is that google code had been dead for a number of years when somebody remembered it was still a thing and then shut it off in 2016. Google was already using GitHub ("To meet developers where they are, we ourselves migrated nearly a thousand of our own open source projects from Google Code to GitHub.") and added an "export to GiHub" button on google code. Maybe if you were .net you went to codeplex but most everyone else (including Google) went to GitHub, if they weren't already there. https://opensource.googleblog.com/2015/03/farewell-to-google-code.html

The company behind Loongson has been very FOSS friendly. There was a 2009 Google Summer of Code project porting Debian (or was it Gentoo?) to Loongson CPU. https://opensource.googleblog.com/2009/05/2009-google-summer-of-code-c…

https://opensource.googleblog.com/2021/09/using-saliency-in-progressiv… can be interesting

[0] for those interested in Highway. It's also mentioned in [1], which starts off > Today we're sharing open source code that can sort arrays of numbers about ten times as fast as the C++ std::sort, and outperforms state of the art architecture-specific algorithms, while being portable across all modern CPU architectures. Below we discuss how we achieved this. [0] https://github.com/google/highway [1] https://opensource.googleblog.com/2022/06/Vectorized%20and%20performan… which has an associated paper at https://arxiv.org/pdf/2205.05982.pdf.