Google Fuchsia OS team affected by layoffs | Hacker News Reader

Google Fuchsia OS team affected by layoffs

9to5google.com

139 pointsSanjayMehta91 comments

Before anyone kills me and yells that “OS development is hard!”, remember that Fuchsia was running on a Pixelbook, with a basic GUI, 4 years ago, with the first source code being posted 6 1/2 years ago. Now look at what Andreas Kling has accomplished in 4 years…

One can reasonably wonder if Fuschia That said, Nest Hub type product 69394 text files. 66754 unique files. 22781 files ignored.

14 errors: Line count, exceeded timeout: ./scri Line count, exceeded timeout: ./src& Line count, exceeded timeout: ./src& Line count, exceeded timeout: ./src& Line count, exceeded timeout: ./src& Line count, exceeded timeout: ./thir Line count, exceeded timeout: ./thir Line count, exceeded timeout: ./thir Line count, exceeded timeout: ./thir Line count, exceeded timeout: ./thir Line count, exceeded timeout: ./thir Line count, exceeded timeout: ./thir Line count, exceeded timeout: ./thir Line count, exceeded timeout: ./tool ------------------------------------------ Language files ------------------------------------------ JSON 2186 Rust 13783 C++ 10232 Go 3227 C/C++ Header Markdown 3717 Assembly 447 C 1591 CSV 17 Dart 637 Python 424 Perl 41 diff 4 TOML 734 YAML 326 JavaScript 68 Bourne Shell 286 Bourne Again Shell 230 SVG 72 GLSL 57 HTML 29 JSON5 109 XML 42 Protocol Buffers 35 Bazel 37 reStructuredText 9 Starlark 20 Handlebars 38 Mako 8 make 16 JSX 2 vim script 10 CMake 4 Lisp 6 Dockerfile 18 Pascal 3 CSS 8 Jinja Template 3 Qt 1 Meson 5 Fish Shell 2 awk 1 Windows Module Definition 1 HLSL 2 Qt Project 1 DOS Batch 1 INI 2 Nix 1 ------------------------------------------ SUM: 46729 ------------------------------------------ is in development hell…

Eh, I think there's always the danger of lack of urgency in a large corp and instead enjoy tweaking forever, but it ships on millions of devices, so it's not vaporware by any means. I just hope Google doesn't end up strangling it organizationally because they can't bring themselves to actually pull the trigger and replace the core of Android with it.

Exactly. See Nokia and how they Maemo/Meego saga

I remember Google upfront said when they started development of fuchsia that it would take seven years before there was anything to show. Cannot find the reference. I also remember Microsoft saying the same when they were fiddling with Midori (now canceled windows replacement). It appears that 7 - 10 years is the typical timeline when developing an OS from scratch.

Midori had a much more uphill battle due to drivers (which would all need to be totally rewritten in M#) and the total lack of sandboxing in Windows apps.

Phones are a static collection of hardware, so vendors will have an easier time with providing all drivers needed. And Android apps are much easier to sandbox, given they run on Dalvik and have capabilities and don't just run as root with access to /.

I'm still so sad Midori was never open-sourced..

Bing was running on Midori before the project was cancelled due to factor Sinofsky from Office leadership, well known for its COM and C++ love, and how Longhorn was killed as well after Office team jumped into the mix.

To be fair; given the evidence of that timeline are Fuchsia and Midori, I think the more accurate statement is that the timeline for developing an OS from scratch is not less than 7 years, with an infinite upper bound.

Fuchsia shipped on Google Home Hubs, upgrading them in the field (replacing everything from firmware up). We did so without breaking them all.

Why? Did Fuschia enable any new features on Google Home Hubs which would have been more difficult/impossible with a Linux kernel? It's an honest question, my bias is that Fuschia is a solution in search of a problem but I am happy to be convinced otherwise.

Fuchsia has a significantly better security stance, in both implementation and in architecture.

Hubs are fully vendor locked in, and Google controls every single but running there. Security argument, while valid like everywhere, seems BS. I’d be surprised if security profile for a hub was main concern for the product.

Anecdotally my hub got much less responsive around the Fuchsia update. I’m not blaming Fuchsia directly for it, but one could wonder what would happen, if resources dedicated to that, were spent on usability improvements.

Let’s be honest, Fuchsia was rolled to a Hub, because it’s a product no one really cares about and team needed some guinea pig/stuff in production to justify its existence. I say that confidently, as someone working at big tech, and having myself experience being in rollouts like that.

The Nest team definitely won't agree with your value judgement of these devices, nor do many of the users who paid for them. To this end, even if there are different motivations for different teams and other tactical considerations, it is absolutely necessary for the owning team of a product to have a real motivation for such a significant operation as replacing the entire firmware and runtime in the field. That set of motivations are rooted in the properties of the OS: https://fuchsia.dev/fuchsia-src/concepts

What was the security issue with whatever the original OS was? It's not like I'm installing any apps on these things.

Maybe because Google has the focus of a crack addled flea?

The same Google that is actively developing three unrelated consumer operating systems that a few years ago introduced three incompatible messaging apps at one event.

I've heard that at google (urban legend?) the best way to advance is to start a promising project that catches the eyes of managers as a potential money maker, get a very basic prototype going, and had it off to a team and then move on to your next big "money maker" idea. The people who "maintain" never really get much attention from management

I think that’s the way at most companies unfortunately. Maintenance is a cost centre while a new product is a potential revenue stream (until it too becomes a cost centre).

Heeey, AFAIK all three messaging apps were on one of them. The other two OSes didn't peddle any chat apps.

I was wrong, they actually had four at the same time.

https://www.cnet.com/tech/services-and-software/allo-duo-goo...

Google has made way too many chart apps, but this article is disingenuous; Duo is a video calling app, not a messaging app.

I don’t think that’s the argument that you think it is…

The other apps also do video.

Fuchsia debuted seven years ago and was in development even before that. And it hasn't been cancelled more than 80% of the team remains.

Crack addled flea? That's how you'd describe seven years?

Fuchsia never really cared for a particular GUI. They made a couple of shells which could only really be called demos because they ended up removed later.

https://fuchsia.dev/fuchsia-src/development/hardware/intel_n...

Hmm, Fushcia builds a whole new kernel in a different paradigm and then a userspace upon that. Android was built upon an existing and mature kernel. It doesn’t surprise me that Fushcia takes longer.

It doesn't really. Fuschia is a fairly standard microkernel design written in C++. Their big idea was a sea of internet components but it doesn't seem to have worked out? Where's the eye-catching Xerox PARC style demos of what they spent so many years figuring out?

Andreas Kling develops Serenity OS, a new from scratch open source OS and kernel in C++, what does Android have to do with it?

Sorry. I assumed that the person you named was responsible for Android, given how that would have been a good comparing target for Fushcia.

For this to mean anything, you will have to make the claim that Fuchsia is more similar to Serenity OS than Android or any other full scale OS project. And if it truly is closer to the pet project, then that's not a good sign for a company like Google, who is paying X number of engineers to work on this

one good way to slow things down is to add more people

Especially Perf(now GRAD)-evaluated people who care about promotions and performance reviews. Which is of course completely normal, not blaming the individual engineer here. Engineering directions just start taking weird turns that don't make that much sense globally, but locally they're the best course of action.

Not sure why you're being downvoted for this. It's true. From what I could see, the Fuchsia team had basically unlimited headcount and blew up into an absolutely massive thing in a short period of time. They still missed deadline after deadline.

Overall my experience at Google, esp in consumer hardware and things touching it, can be summarized as: Has a single person here read (or heard about) The Mythical Man Month?

There is no part of this statement that is true. Fuchsia never had "unlimited headcount" (early on there was a buffer more than needed but our VP told use we couldn't hire too quickly as to not to overwhelm things and then later that buffer was taken away in reorgs when it was actually needed later which meant we had to do more with less). Fuchsia also never missed deadlines we made outside of the team. Any delays to market were not the fault of Fuchsia team itself but delays our internal partners needed.

It sounds like Google is rushing their resources into AI. They might see this project as one that they can stop and start cold when needed.

ChatGPT has a real chance of disrupting Google Search. Heck I'd say OpenAI's generative art models could even disrupt google images search.

Google has its own ChatGPT-like/LLM thing that is at least competitive. There's almost for sure a bunch of very legitimate reasons they haven't tried to mate search with that.

#1 from my perspective is that those systems... lie.

I suspect that the main problem with these systems is that Google doesn't know how it will affect Google revenue, i.e. if there's a possibility of destroying a lot of revenue without replacing it with larger revenue. I.e. it's what's known as an innovator's dilemma.

I don't think that's the main problem. They wouldn't invested money into Google Assistant, which directly hurts Google revenue. They also already give direct answers to questions on Google without ads.

Realistically, the main problem is scaling and costs/benefits. You have to remember Google is accessed by billions of people a day. Now imagine they serve a product like ChatGPT. You would be pissing off a lot of users with a product that is laggy, often crashes, requires a hourly rate limit just to protect your business from going bankrupt on computing costs. It's not really beneficial to Google to rush a model into their golden product that isn't ready for the scale it operates at.

But it is beneficial for a smaller competitor to slowly chip away. Meanwhile googles investors might rather the legacy company return value to shareholders with greater dividends/share repurchases. While the new competitor is given blank checks with investors not wanting to monetize too much just yet.

I would guess the bigger problem is that someone is going to get GoogleGPT to write Nazi propaganda and they don't want the reputational harm.

For a LLMs as search replacements there's a real big issue putting the 'Google' brand on something that could essentially replace wikipedia and be inaccurate yet trusted in unexpected ways. ( https://i.redd.it/40cnc2bluh6a1.png )

ChatGPT opened the box but there's a real reason to be hesitant

This holds for random blog posts too. Arguably, ChatGPT lies less in some cases because it aggregates multiple sources. (However, it doesn’t counter common misconceptions and blogs typically don’t either.)

How?

Most people that use Google images aren't looking for generative art. They are looking for images of real things that exist.

Dall-e is fun to play around with for a bit that about it lol.

No. But their systems can point me to all the images it has seen that has a pink Cadillac with a man in a black top hat. And if no image found is suitable you can then select any of the unsuitable images and have it altered to be suitable.

You underestimate the ability for these things to index data.

But most people don't seek a random image of "pink Cadillac with a man in a black top hat". They either search a specific picture they can describe that way (famous photo, specific guy to reference...) or don't know what it looks like and need a refence (and the uncanny, not-quite-right thing AIs tend to generate on unspecific prompts is totally useless). In both cases, AI-generated look alikes are the wrong answer to the question.

It might work for some use cases, but it would never replace the search function.

My impression is that a lot of people use Google Images for things they can drag into their slideshows and posters. In that case, it's probably fine to use generated art?

Are there ads on Google Images? Then Google doesn't care.

Yes, because would be very happy about people getting their images from Bing AI Images instead of them. :p

ChatGPT has almost no chance of disrupting Google's search in the short term, it's not even a search engine?

where I can find a video to learn about making hinanese chicken rice:

YouTube: There are many cooking channels that feature videos on how to make Hainanese chicken rice. Some popular ones include "Tasty," "Gordon Ramsay," and "Maangchi."

Yeah, I think people are just obsessed with talking about ChatGPT lately?

The appeal of chatgpt at this point is twofold: SEO spammers haven't yet discovered how to manipulate it to their benefit, and Google can't inject ads into the returned results. In other words if it does return a useful result to you, it's where Google was when they were competing with Alta Vista and the like in their heyday.

I agree with you that there's a ton of hype about chatgpt, but if I can get useful and unadulterated answers from it, I will definitely pivot to searching there first. This is above and beyond the ability of chatgpt to seemingly generate tailored results to queries beyond simply returning a list of web pages ranked by relevance.

I thought so too, but then I tried it myself

There are currently unsolved issues such as hallucination, query cost, scalability, fast model update with newly indexed data and toxicity that would need to be addressed for ChatGPT (or LaMDA - Google equivalent of OpenAI ChatGPT) to replace search functionality.

ChatGPT, like crypto-based finance and self-driving cars, is overhyped by people who only understand the technology superficially or seek to directly benefit from the hype (not to be confused with benefitting from the success of the technology itself).

But I don't fault Google. For a business that makes in the ballpark of $50B profit per annum, even something that has a 0.1% chance of disrupting you is worth throwing a few billion at.

Time will only tell but search is definitely Google's bread and butter core and they will do anything to prevent any real competitor that might capture significant market share from them.

I so, I guess they’ll find out the hard way what happens when you chronically underinvest in something.

Given that Google's whole IoT world is seemingly moving to Fuchsia this is a bit of a surprise. is I think the only one with user facing UI, and an extremely limited one; really more an appliance. With a couple fixed hooks/extension points 3rd parties can tap into (versus being a general OS).

I still have an ebay search I havent cancelled to show me results for one of the old 2-in-1 machines that had been a dev target, the Acer Switch Alpha 12 (https://fuchsia.dev/fuchsia-src/development/hardware). But that was sooo long ago now. And I dont really know what user-facing intents have really emerged or grown since. Seems hard to kickstart a new consumer os, turning a viable platform into something emd userable. Yeah, having your own niche appliance os is fine, but less interesting.

It'll be interestint to see what happens with 3rd party Chromecast/Google Cast products over time. Will Fuchsia ever have inroads on 3rd party products? Do many of them already run & will they continue to run CastOS? Are there other off the shelf Chromecast implememtations? I was surprised when Android Things got dropped, when the ecosystem fractured, given how much infrastructure/packaging/services they had been able to use. Interesting times.

> Nest Hub type product is I think the only one with user facing UI, and an extremely limited one; really more an appliance

And that took them probably 5x longer than they told mgmt it would and missed multiple deadlines.

(I worked on the product for its original launch, and before it was all rewritten for Fuchsia)

The whole world of shipping multiple OSes and home software stacks at Google was a minefield of politics, turf wars, and "shipping the org chart" that was just painful to watch and be a close spectator to. Much dysfunction.

BTW, I don't think any third party devices run Chromecast Linux; that was always only first party devices when I was there. The third party stuff was running Android generally.

We shipped the first round of the Home Hub stuff running on a Chromecast stack with a UI written in HTML/TypeScript, and it worked pretty good. Then a few months later they decided to rewrite the whole thing from scratch in Dart/Flutter and replace the whole OS with Fuchsia for some reason. So a whole elaborate transitional shim layer was built to get Dart/Flutter UI to host on the Chromecast distrib, while the Fuchsia people did their own thing that would also run the same UI. Every little had to be re-built from scratch, on an already shipping and successful product. It was insanity.

You had that due to poor C-Suite leadership. A good C-Suite would have noticed the in-fighting, and re-organized the group into a single threaded leader who is a benevolent dictator. Google's disfunction is due to the lack of clear organizational decision making leading towards a 'democracy' with no movement.

If the group is well organized, you seldom need to have cross organization collaboration to achieve your goals. Then you cannot claim to have leadership or complexity because stuff accomplishable by a single team is considered simple to do. Bad for perf.

At the C level is hardly the only place where there is/was bad rot.

And in this case, it wasn't just one group. It was multiple product areas. Each trying to make the other one execute their vision for them, it seemed. It wasn't "democracy", it was more... feudalism.

Anyways, the parts of Google that were just shipping boring old unsexy $$ making core/search/ads stuff were managed much better.

To be fair: From the little I heard at least the transition was super seamless, which is impressive for such a big change.

I always thought it was supposed to be a demo that Fuchsia is viable and can reasonably easily be migrated to.

From the consumer's POV it was fairly seemless. Behind the scenes it was a slog. And they kept missing deadlines. For what?

It seemed more sensible to me that they should have just done it with a new device, instead of taking an existing device that had already shipped successfully into millions of homes.

Financially, made no sense, if one actually cared about labour costs. Duplicated engineering effort all over.

Well the real question is what new capabilities did anyone justify that project with? Fuschia seems to lack a killer edge, some radical new take that can justify such an investment. If you read their website it's all about components and stuff but consumer IoT is the opposite of an extensible component platform.

Linux on embedded devices is not sunshine and rainbows, especially when it comes to GUI. It is one of those messes that developers instinctively want to throw out. Companies like Google can afford lateral moves where the immediate business case is not there, but opportunities may pop up down the road. Of course, missing deadlines can easily get such moves killed.

Interesting, what's the biggest pain point in your view?

syscalls and drivers tend to be a big one. Android has a HAL to make it easier, and Fuchsia offers something richer. I know Amazon took a similar tack at one point of trying to build their own thing from FreeRTOS.

Well, true. The main material advantage I've been able to gather is that it has "simpler licensing" (you can keep more code proprietary). But I don't know if manufacturer's really care about keeping their drivers proprietary all that much.

> Seems hard to kickstart a new consumer os, turning a viable platform into something emd userable.

Actually ... the Android runtime is pretty portable and works VERY well on Chrome OS (you can even connect to VPNs and use custom keyboards through it). That's kind of 80% of what you need. I mean of course it's still hard, but it seems to me like they have all the parts they'd need. Which makes it all the sillier that they still haven't managed to fix Android updates.

You do understand that the Android runtime on Chrome OS didn't just pop out of the ground fully-formed, right? Years of effort went into it.

I do, but it proves to me that it's possible. Google can do it and so (kind of) can Microsoft and the Linux community. Of course it's still hard, but a lot of what you need is already there; most importantly Google's moat, the proprietary Play Services.

Indeed!!

Making Android vaguely somewhat compatible with the rest of the Linux universe has been an incredible long & difficult struggle. There have been so many heroic works, to tackle the vastly radically different conpletely NIH-ed Android universe & make it deal-with-able.

Given that Google already killed two IoT OS projects, Brillo and Android Things, they can as well kill Fuchsia, it wouldn't be their first.

Were Google's IoT products affected by layoffs as well?

I know Nest lost headcount.

I don't think there was a single PA at Google that didn't lose people. From the stuff I've seen on LinkedIn via ex-colleagues, etc. I have a hard time believing that they d didn't just run a random # generator over the set of US employees, just drawing people out of a hat until they hit 12k.

Certain teams were impacted more than others. The actual distribution for like comp/seniority/whatever seemed random but some were entirely unscathed and some saw massive cuts.

I hope you all are taking noted of employers for whom you won't work for. This isn't fat trimming apparently just an attempt at trying to get near-zero interest rates.

Are you saying that all the tech companies are conspiring to lay off workers to cause the employment rate to drop from its near record levels so that the Fed will panic and reduce interest rates?

They're going to have to lay off a lot more people to move the needle aren't they?

Not GP but yes I'm saying exactly that. If you concede the following axioms:

1. All tech companies' management and investors are primarily incentivized to increase the stock price total return.

2. High interest rates are kryptonite for growth tech company stock prices and will always be the biggest dampener on stock price recovery and growth.

It makes sense that company management and investors are incentivized to get rid of high interest rates. You can't manage your way out of a problem that wasn't caused by poor management. These companies were darlings of the market one week and shunned as uninvestable the next. The only thing that changed was the macro environment.

You don't need actual synchronization to observe a synchronous response. If disjoint parties have the same incentives and are equally rational, they will move in a direction that appears synchronized. Of course, these companies aren't even actually disjoint - the same management consultants may well be advising multiple tech companies and coordinating a response.

Big tech growth companies are an extremely rich and powerful segment of the economy. I wouldn't put it above them or beneath them (depending on how you view it morally) to not take givens as givens and manipulate their environment itself.

i'm not saying i buy this conspiracy theory, but laying off ~100k high-paid tech workers in the noisiest way possible will definitely move the needle. maybe not directly, but there's a lot of uncertainty and knock-on effect caused by all these big layoffs getting a lot of press coverage

Hmm, I wonder, would that be treason? It would basically be a organized attempt at crashing the US economy. Though I suppose banks have gotten away with it before.

If the government's official policy is to crash the economy then helping them isn't treason.

But here it would crashing it in the opposite direction, trying to influence the fed through threatening the job market

That's a bit silly.

It's more likely it's an attempt to break our (software engineering industry workers) bargaining power, compensation rates, and "entitled-ness." Expect this across the whole economy this year, but it shouldn't surprise that corporate sector would go after the most $$ workers first.

Oh, and most importantly, to get share prices up... in anticipation of immanent Q4 earnings report. Sundar Pichai's compensation is being tied to this: https://www.reuters.com/technology/alphabet-links-more-ceo-p...

They cut 6% of their staff and killed dead-end products. That's a normal and healthy decision. The number of highly compensated tech personnel laid-off barely moves the needle when it comes to inflation and employment metrics.

It's not about moving metrics. It's about generating headlines and swaying public opinion. In fact over a 12 month period most of the companies who did layoffs probably see a net increase in headcount.

It will be interesting to look at commit history in a while to see how this week compares to preceding and succeeding ones.

It's strange that the Fuchsia homepage says that it is written in 'C', when wikipedia says it's written in 'C, C++, Dart, Go, Rust, Python'. I've counted the number of lines in the fuchsia codebase myself using `cloc .` and the results are below. It is not written in solely C or even mostly C. In fact there are more lines of code written in Rust. The fuchsia team should rectify this incorrect information on their homepage. Why is the implementation language even listed as the first thing on the homepage? It's really odd.

$ cloc . pts/third_party/d3_v3/d3.js #x2F;diagnostics/archivist/src/inspect/mod.rs #x2F;lib/diagnostics/selectors/src/parser.rs #x2F;proc/tests/android/gvisor/expectations/default.json5 #x2F;settings/service/src/light/light_controller.rs d_party/golibs/vendor/golang.org/x/net/idna/tables10.0.0.go d_party/golibs/vendor/golang.org/x/net/idna/tables11.0.0.go d_party/golibs/vendor/golang.org/x/net/idna/tables12.0.0.go d_party/golibs/vendor/golang.org/x/net/idna/tables13.0.0.go d_party/golibs/vendor/golang.org/x/net/idna/tables9.0.0.go d_party/gvisor_syscall_tests/expects/netstack3/loopback_expectations.cc d_party/gvisor_syscall_tests/expects/netstack3/tcp.json5 d_party/rust_crates/vendor/tuf-0.3.0-beta11/src/metadata.rs s/auto_owners/src/main.rs

github.com/AlDanial/cloc v 1.90 T=142.47 s (328.0 files/s, 107722.2 lines/s) --------------------------------------------- blank comment code --------------------------------------------- 50 0 4674452 404572 625522 3339090 402599 236729 1937881 91268 152628 671742 8236 174436 243376 619602 108561 72 318674 35435 18812 293467 47081 64048 271343 0 0 140374 10788 11546 61906 11559 13427 47039 5210 4997 38375 960 33038 36364 3762 5518 21375 976 1555 19767 3340 10315 19305 3536 5394 16504 2882 4408 14259 66 362 11872 2551 5322 7493 614 39 7344 303 2199 6100 468 172 5258 1214 1746 3007 325 299 2367 538 152 1034 192 418 856 123 8 838 70 0 537 100 34 371 13 31 250 72 124 232 44 123 227 81 99 217 42 30 211 50 68 209 26 16 190 16 0 175 0 0 156 27 18 98 18 43 92 14 3 80 7 0 46 6 0 38 11 3 31 3 0 20 2 0 16 1 0 6 --------------------------------------------- 1314012 1442694 12590890 ---------------------------------------------

This is the Fuchisa homepage, https://fuchsia.dev/, not what is on that screenshot.

And who is surprised?

Time to change industries, it looks like it will be downhill from here (at least for a while). My gut tells me we have hit a saturation point in IT, a lot of very busy people getting very little done. What happened at twitter was eye opening. I agree with earlier posters about the impact of inflation and interest rates, seems we have hit a tipping point.

There is certainly not a surplus of capable software developers/engineers in USA. Maybe some corners of SV.

Reply on news.ycombinator.com