[ironman1478]

Having worked at meta, something I noticed is that the orgs that were well run were ones that were bought. WhatsApp, reality, insta, etc. I worked in an org that was not associated with those products and was purely homegrown and it was awful. Things got done but horribly inefficiently due to over hiring and extreme requirement and schedule shifts.

I believe that the cultures that were developed outside of Meta are used to launder the image that meta as a whole has a good engineering culture.

[svachalek]

I had a similar experience at Google, they were so convinced they were the only good engineering company in the world and had to protect themselves from all the wrong-thinkers outside and yet the only progress they made was via acquisitions.

[surajrmal]

Good engineering isn't always about building new things, but making existing ones continue to work well. Funding new ideas is generally a hard problem for large organizations and that's not entirely an engineering culture problem.

There are a lot of terrible practices out there in the world that you should stay wary of. Too many false positive alerts, flakey tests, not enough tests, not listening to users, taking a solution because it's easy and popular but not necessarily a good fit for your specific requirements, etc. Many of these practices are popular unfortunately. That's not to say others don't have great ideas, just don't copy them blindly.

[abustamam]

> making existing ones continue to work well

Say what you will about MS, but the fact that Windows was backward compatible for so long (it may still be, I haven't kept up) is damn good engineering.

[apejcic]

Having led engineering in SF big tech and now building my own startup, I’ve heard this exact argument countless times. I believe it is precisely why large tech companies are losing their ability to innovate.

Engineering shouldn't be an academic pursuit of "good engineering" for its own sake. It is fundamentally a business function. The objective is to achieve real world goals and build things people actually want, even if that means the solution is scrappy and unpolished today, so long as it solves an important problem cheaply and effectively.

In large organizations, accountability for actual value delivery is so diluted across the org chart that it practically disappears.

Leaders gravitate toward building bloated "internal platforms" that offer little external utility. Engineers over-engineer simple problems because their performance reviews reward technical complexity, not business impact. It becomes a systemic shield, allowing the old guard to justify their headcount without delivering new value.

Software engineering is uniquely plagued by this. Civil engineers understand their objective: build a bridge that safely carries traffic within a specific budget. They don't invent a novel, hyper-complex suspension architecture when a standard short-span bridge will do the job.

[inigyou]

I would decouple "build things people want" from "business function". I would put in the former bucket things like: making the computer work smoothly and run apps. I would put in the latter bucket things like: show advertising to the user when the log in. Engineering should always be about the first, or it's bad engineering. The second is far more controversial - business leaders would like you to consider it good engineering to do that, but you get to interpose your own ethics, but you'll probably be fired for having ethics, but that doesn't make it bad engineering.

[surajrmal]

Considering the large increases in revenues bug tech extracts year over year, the maintenance and small improvements are not necessarily completely disregarding the business side of things. You cannot keep sustained performance without proper practices. Too much short term thinking like you see in startups leads to loss of confidence that new changes won't break existing customers, or hellish on call that no one wants to be part of. There is a lifecycle to product development and different phases require different tradeoffs. You sound like you align with the builder phase and that's great.

I will also note that most civil engineering is about maintaining existing structures and roadways, not building new ones.

[brazukadev]

> Considering the large increases in revenues bug tech extracts year over year

I don't know if bug tech was a typo or intentional but from now on I'm using it in place of big tech.

[Yoric]

> Engineers over-engineer simple problems because their performance reviews reward technical complexity, not business impact.

Having tech led teams at both high-profile and low-profile tech companies, that is _not_ my experience.

Most engineers worth their salt value quality. Product Management generally doesn't see quality as a feature, therefore doesn't take it into account. For short-lived code, quality is indeed over-rated. For long-lived code, quality is the thing that determines whether you can keep improving or tuning other features or whether you'll miss deadlines. Sadly, retrofitting quality in an existing codebase is awfully expensive.

There's also one important, but often undervalued, aspect. Overwhelmingly, today's tech world has been built by neurodivergent engineers. In my experience, neurodivergent engineers tend to value getting to the end of things, rather than letting them drop mid-way. This can absolutely be seen as over-engineering, but it's often a cognitive scaffolding that will ensure that the work can be resumed (by them or anyone else) even after context-switching to something entirely different.

Whether it actually _is_ over-engineering often depends on how well the engineer is aware of the actual needs, rather than being spoon-fed instructions without visibility. Or on the maturity and skill of the engineer, depending on the case.

[jwoq9118]

Relevant

https://terriblesoftware.org/2026/03/03/nobody-gets-promoted...

[galkk]

Dremel/big query Spanner Borg Chrome

None of those were acquisitions

[lokar]

And they solved real, hard problems the company was facing for which there were no good external solutions.

[ath92]

Waymo too

[throwaway2037]

V8 that backs Chrome and NodeJS

[JacobAsmuth]

You'd say that Google Maps hasn't improved much in the last 20 years?

[da_chicken]

My best friend bought a house. She noticed within the first month that the name of the road she lived on was misspelled in Google Maps. Specifically, it's a slightly unusual spelling of the road name. Like "Chickorie" vs "Chickory". It's particularly annoying because there's a road the next town over with the traditional spelling. It doesn't even share address numbers, but Google Maps still frequently misdirects people.

It's correct everywhere else. The road sign, the municipal tax parcel GIS, the post office, Apple Maps, MapQuest, OpenStreetMaps. All of them had the correct name, except Google Maps. So she reported it through Google Maps. And reported it. And reported it. Every few months she reports it again. She's asked friends to report it as well.

It's still wrong.

She bought her house in 2015.

[aeturnum]

My "oh google maps" story was that I was hanging out at a new cider bar in my neighborhood and asked the owner why I couldn't find it on google. They said they got a message that they'd been banned for listing boosting - which they said they didn't do any had no idea. So I reached out to an acquaintance who knew a lot of maps people. Some investigation later it was entirely unclear why the bar was banned - it was some conflict with overlapping systems deep in the map bowels - and the bar is visible again. It's just good luck that the owner happened to talk to me and I happened to know someone who looked into it!

[rcbdev]

This kind of support should be mandated for any digital utilities company such as Google.

The difference between small business success and insolvency was based on the shier luck of being graced with the presence of someone in contact with the priesthood of Google, where no real contact from the plebian citizenry is allowed.

Exactly this kind of thing is why the EU feels the need to regulate the shit out of U.S. big tech.

[inigyou]

I think in the EU this is mandated. Every algorithmic profiling decision must come with an explanation and human appeal, under GDPR?

[vovavili]

Legislating costly support for a free and accessible service is how that service stops being free and accessible.

[areoform]

These things can accumulate and ruin lives. I'm surprised that there haven't been more class action lawsuits against "errors" like these. Because it might seem like a benign accident -- but how many people have lost important parts of their lives – banking, photos of their children and more – because "computer says no?"

Eventually, these systems will be mostly artificially generated, and perhaps the machines will have fewer error rates than the humans. Perhaps not. But how many humans will understand the machines well enough to ask these questions in the first place?

Machines were supposed to free us from bureaucracy. Not freeze it everywhere with few avenues for escape.

I have had an encounter with something like this via Wise / Transferwise. It has been half a decade and nada. And I estimate that it has cost me north of $20k+ over that time.

Google, Wise and heck Maps were started with the ambition of adding something to the world — e.g. Google's original "organize all knowledge" mission - but over time cruft accrues and these companies rapidly accumulate negative side effects / drift away from their core mission.

When was the last time Google / Alphabet / whatever did something that involved improving access to the world's knowledge? They've degraded their search to the point of uselessness and beyond. Slowly alienated their best researchers and engineers. And done their best to turn away from the entity that made Google Books — "we'll scan all the books for the good of all humankind."

[inigyou]

Google just lost a lawsuit by an unnamed big media company in Germany because its AI kept telling people the media company was a scam. This is the only form of feedback that companies like Google actually listen to.

[andsoitis]

> I have had an encounter with something like this via Wise / Transferwise. It has been half a decade and nada. And I estimate that it has cost me north of $20k+ over that time.

What happened? I also use Wise.

[eitally]

My anecdote is worth just as much as yours. I moved cross-country in early 2016 (for work at Google) and noticed when we closed on our house that the next street over was mis-pronounced by Google Maps. I reported it and it was fixed within a week.

[gnerd00]

you moved there --to work at google; you reported it --while you worked at google. What is different hmmmm

[eitally]

Nothing -- I reported it via the consumer Google Maps function. In fact, I've reported several Maps errors over the years and all have been addressed fairly promptly. This includes before I worked at Google, during my time there, and after leaving.

[brazzy]

If they reported it via the normal Google Maps reporting function, it doesn't actually make a difference, does it?

[antonvs]

The difference is the location of the house.

[annzabelle]

Google Maps put a town label near my parents' house that is a name that was on some old maps in the 1910s, but fell completely out of use before the current suburb was built in the 70s and 80s. I had never heard it once in my life until I was in my early 20s and met a transplant who had picked up the name from Google Maps. I attempted to report it as inaccurate (USGS has had it as being out of use for decades) multiple times over the years, but no reply.

[inigyou]

Google puts the "Warsaw-Berlin Urstromtal" (glacial valley?) right in the middle of the river near the Friedrichstraße train station.

You can be browsing the area and suddenly, yep, there's a glacial valley there in the middle of the river. According to Google.

This is a large-scale geographical feature that crosses most of Germany and Poland. But actually it's in the river next to Friedrichstraße station.

[dmurray]

Most of the neighbourhood names in my large European city were like this on Google Maps. Either old names or hyperspecific landmark/street names that I had never heard being used as synecdoche for the area.

This was definitely true 4-5 years ago. I looked now and it's mostly better at most levels of zoom.

[consp]

Google maps uses suggestive names (e.g. a park walkway given a name by the municipality) as official road names, and when it does that it marked it as a cycling path. Both are invalid and cycling on them is illegal. This has been going on for years.

[iamacyborg]

Google Maps have an address API some companies use when filling out addresses for shipments.

Despite my building existing for over 10 years, putting in “flat number, building name” into that system will be rewritten to “house number, building name road” which is an address on a he other side of town (and London ain’t exactly small).

I’ve had multiple orders go missing as a result.

[inigyou]

Falsehoods programmers believe about addresses...

The only universal address input is a freeform text box. Optionally with a country, which should also be a freeform text box because countries get created and destroyed and you won't remember to update it. City or district is probably okay - as a freeform text box - addresses without one can write (no district). But street addresses are different everywhere and must be freeform unless you're the government of that region.

[ralferoo]

This is why we have postcodes.

More concretely, a postcode is usually specific to between 7 and 30 dwellings, in the usual case it's about 15 houses.

Contrary to common belief, they're not always on the same street, but in the vast majority of cases (like maybe 99%) house number + postcode is sufficient. Definitely flat number + building name/number + postcode should get your stuff there.

[iamacyborg]

Yeah, the google maps address thing just overwrites my post code, and some senders in the US don’t appear to pay attention to that.

[21asdffdsa12]

The trick is use the right keywords. "Dear Sir and Madam of google, im writing you to inform you, about a possible liability your software opens itself up to, by referencing emergency vehicles to the wrong road. Example:

Regards"

[Ferret7446]

Usually street data comes from an official source. You probably should have reported the issue to the local government rather than Google. (Just because the name is correct on some government material doesn't mean it's correct everywhere.)

[jazzyjackson]

It’s doing a great job making sure I know where Burger King is when zooming into a city

[nightshift1]

Interesting submission on that topic : How Google Maps allocates survival across London's restaurants https://news.ycombinator.com/item?id=46203343

[tasty_freeze]

No kidding. When I zoom in on a few square miles around my house and ask to see restaurants, it shows only a few of them even though there is plenty of space -- it isn't like it has to prune some to make it readable. I zoom in and some appear that weren't there before ... why? There was room to display it before I zoomed in. Gaaah!

[morganf]

Oh wow I always complain about this, I didn't realize other people were also so frustrated by this. Tons and tons of cafes by my house, but I zoom in with about a six block radius, where there are at least 20 cafes or cafe-restaurants there, and literally 1 or 2 come up.

[gwking]

The grocer on my block is so tiny that I hit the max zoom before it shows up. If you pan around fiddle I can occasionally get it to appear, but it effectively does not exist according to google maps.

[andreareina]

Not to mention that zooming in will make some listings disappear

[telchior]

Yeah I utterly despise using Google Maps now because of this. In my neighborhood, there are restaurants that will only appear at a very specific zoom level. Zoom in OR out and they'll disappear.

I see things when I'm out and about all the time that just never appear on Google Maps -- cafes for instance. You can search cafe, coffee, restaurant, etc and they won't appear. Search the exact name and it comes up with 300 reviews. Sometimes, they'll come up in general search in week 1, disappear in week 2, and reappear in week 3.

Google is not good at consumer software and not a responsible company in many ways; Maps really deserves to be replaced with something, but unfortunately there's no competitor to speak of. Apple Maps is nice but listings don't come up because they're literally not in the database.

[bell-cot]

I'd assume it's pruning to make only the best-paying (to Google) restaurants visible.

[pjc50]

This is to force you to use search rather than your eyes.

[rickypp]

I started to notice "Turn left at the light, after the Chipotle™" last year... ads in my turn by turn navigation is next level innovation

[johannes1234321]

I believe that the initial idea there was to refer to notable landmarks and these company signs are readable from a distance.

But yes, they could serve well as ad space.

[decimalenough]

This also works much better in the vast slabs of the world where roads don't have names, or if they do, they're signposted badly if at all. (India, Japan, most of Africa, etc.)

[bobro]

So you think they weren’t getting paid for this? You think they just freely mention corporate food companies purely for users benefit?

[asdff]

It is still drawing bike lane networks in cemeteries and random bits of grass and pavement. So no. The way I see it is is actually worse than it was in 2006. When I opened it today it asked me if I would like to see Dua Lipa's recommended restaurants.

[21asdffdsa12]

Sir, this is Dua Lipas favorites Wendys - do not cause a promotion commotion, not again Sir!

[blt]

Not relative to its available resources. For example, there's still no option to penalize intersections. I'm still getting absurd side-street routes and unprotected lefts/crossings to save 1 minute on a 20-minute trip.

[throwaway27448]

It's rather baffling how little they've improved it in certainly the last 15 years. Waze is certainly the better consumer product, even if it relies on similar (the same?) backend software. Apple Maps has caught up and passed it in terms of usable software, even if the actual map quality is worse in some places.

For instance: there's no way to get it to stop killing directions on reaching the destination. Apple Maps puts you into "parking" mode where you can still see the route to the destination—extremely useful for cities where you might have to drive around a bit to find a space.

[boelboel]

It's been up and down for about 8-12 years. Problems get fixed and new ones get added without a particular direction, at least that's what it feels like as a regular user. Some come back years after they were first fixed.

[driverdan]

Maps peaked after Google acquired Waze in 2013 and integrated the data. Since then they've added worthless annoying social features and more advertising.

[joquarky]

I feel like every other time I open maps I get some "look at what's new" pop-up, which derails my thought process and muscle memory.

[allthetime]

Go to a park near your house with walking paths in it. Compare Google Maps with OSM.

That said; of course Google Maps has improved and is likely better than alternatives in lots of ways, but it’s actually not great or anywhere near the detail and granularity of OSM when it comes to the actual map part.

[intexpress]

Apple Maps has been better than Google Maps for at least 10 years and nowadays the way that Google Maps improves is often just by copying Apple Maps

[californical]

I’ve agreed for a long time… Unfortunately though, Apple Maps just added Ads, so expect them to start having the same issues as Google Maps does (like showing big ads for every location whenever you’re moving around at the map).

After Apple Maps being one of my favorite reasons for having an iPhone for the last 7ish years, I’m back to OpenStreetMaps mostly, or still Google to look up business hours. Sad… but having downloaded maps will probably be a good thing long-term

[walt_grata]

Id say it improved until about 2016, since then the additions have generally made the software worse.

More overlays and popups More needlessly verbose navigation instructions Less predictable routes.

Thats just the start

[forgetfreeman]

"needlessly verbose navigation instructions"

You ever try to navigate the highway systems in either DC or Houston with verbal instructions enabled? It sounds like someone is having a stroke.

[walt_grata]

Yep to the point where its easier without the gps.

[daedrdev]

Google has said to turn right at an intersection where you are turning left for years because it doesn't know about a T shaped intersection that tens of thousands use a day.

[awinter-py]

odds of a given street being labeled on gmaps for android still feels like 50/50 unless you get the zoom exactly right

[3eb7988a1663]

This is my biggest annoyance. Such a crap shoot if I can actually see the street names.

[ElectricalTears]

Maps thinks it's fine to do a u-turn on the Sydney harbour bridge.

[XorNot]

Maps straight up can't navigate sanely in Australia I've found.

I've switched to OsmAnd for driving because it'll give a sensible main road following route.

Google Maps reliably sends me down residential roads and adds a whole bunch of turns and then winds up with idiot ideas like "turn right into this two way 4 lane road we tried to slightly avoid".

[saltcured]

I wonder if it ever does that for the SF Bay Bridge.

"Perform Immelmann turn in 200 feet"

[adithyassekhar]

Is this like a platform 9 3/4 thing?

[meroes]

My house and street and surrounding streets still aren’t on Google Maps after a month post closing (new development). It’s on Apple Maps and Uber; Amazon and FedEx deliver.

Google Maps feels like I’m constantly in an A B test for how bad it can be and I’m always getting the bad side. It repeatedly doesn’t update in CarPlay, showing the same miles to destination as I progress.

It doesn’t label streets well either. I try to manually find routes in SF and I have to switch to Apple Maps because not enough streets are labeled no matter how much or little I zoom.

[prasadjoglekar]

It's a mixed bag at best. The product has had numerous features pushed, but in the US, the navigation feature is objectively worse than 5+ years ago. Too much spoonfeeding of directions to ensure the app is kept on. But the app won't actually tell you the freeway number and direction - which the driver can see on the dozen signs on the road.

Or, the incessant "police activity" shit from Waze. That creates all kinds of rear ending hazards as morons try to slow down.

[Noumenon72]

I didn't realize till you said this how it's their fault I have to turn on my screen at every important turn to tell which road they mean, because they won't just say "turn onto the interchange, EXIT 7A".

[Fordec]

I used to be able to view maps in 3D. Not anything fancy, just a tilt shift. This year I noticed the feature appears to be gone. Also, if I want to see my home city with an up to date overhead view from this side of 2023 to see what has been built lately, I have to go look up Sentinel imagery.

I would claim it has indeed become worse.

[Citizen_Lame]

I would say Google Maps regressed, over bloated, showing ads or at least unrelated results. Mysterious hang-ups, where app refuses to work every once in a while.

[Animats]

Google bought Google Earth from Keyhole.

I had a Keyhole account long before Google was in the business, and it worked about like Google Earth does now.

[defrost]

Keyhole dates back to 2001 and draws on several precursors, eg: ERMapper (Perth, Western Australia) which offered similar functionality in the 1990s - in addition to computational process pipelines for, say, warping and geomosaicing inputs, multi spectral filtering (adding and subtracting weighted bands to sharpen certain features, etc).

ER Mapper was the first software I dog fooded that transferred images without the dial up scan line by scan line top down process that existed to that point in time.

The same discrete wavelet transform (DWT) and variation techniques occurred to several entities at much the same time .. some US patent pettiness killed a lot of development for a couple of years from 1999 forward:

* https://en.wikipedia.org/wiki/LizardTech,_Inc._v._Earth_Reso....

[ipaddr]

It does less than before and it blocks more things and has more spam. Now you have to pay for map embedding on a website. Let's go back 20 years

[bigfatkitten]

Over the last 8-10 years or so, I’d say it has gotten considerably worse.

[stackghost]

I'd say it's much worse than it used to be.

[gmerc]

It’s regressed a lot.

[forgetfreeman]

I'd say it's devolved aggressively at least on mobile. Between the firehose of modal bullshit I didn't ask for, sub-optimal pathfinding, and annoying prompts from waze legacy features there's a ton of unwanted friction added to the process.

[dlev_pika]

Google: HOW ABOUT SOME AI ON IT?!? You know you want it.

Lmao

[throwaw12]

Google Cloud has a solid product, not sure about internal engineering culture.

[casualscience]

brain was grown in house and likely deserves as much or more credit than open ai for the current llm boom

[echelon]

They literally would have kept the entire thing to themselves and never productized it. Think about how much investment capital and human capital OpenAI poured onto this by having their big "hello world" with ChatGPT.

[vineyardmike]

They literally had a productionized transformer model (translate) which was the inspiration for all of this. The famous paper was written about a productionized ML Model.

They also had a productionized LLM in search (known as 'mums') before the whole "AI Chatbot" craze.

They also had a chat-tuned LLM chatbot (LaMDA) in testing internally. It was shown off over a year before ChatGPT was created (2021)... and later released as an app (AI test kitchen) before ChatGPT was announced.

ChatGPT may have been the big industry moment, but Google was releasing LLMs to production before OpenAI.

[thr0w]

> LaMDA

Wow thanks for reminding me of this. Remember that leaked story of an engineer claiming it was sentient, even before ChatGPT came out?

[somedude895]

Meta and Google both are companies that got lucky and managed to carve out a quasi-Monopoly in their space. Sure, they did a lot of things right at first and provided a popular product, but for the past years/decades they've been using their money to buy other companies, adding a veneer of "innovation" to their image, constantly coming up with "cool stuff" that they're unable to market successfully, but ultimately they're just milking their insanely profitable core product they came up with 20 or so years ago, like any legacy corp. I have no respect for either of those companies having to work with them in my day job, knowing how lazy and incompetent they are with the services that actually make them money.

[vovavili]

That's the unfortunate outcome of running a business at scale.

[boelboel]

Google has bad governance, same with many other big businesses in the tech related sector partially because of the extremely generous position founders were in during the 00s and interest rates. In industries with lower margins like chemical/material industry you don't get away with this because managers in these industries are way better at telling value from non-value. There's been activist investors ( e.g. Christopher Hohn) who tried getting Google to run a tighter ship but with the dual stock setup there's not much which can be done.

[majani]

The 2000s were a startup winter after the dotcom bubble burst. In those days you were extremely lucky if you could raise over $100k in funding and in revenue $100m was the ultimate fantasy. ZIRP was the 2010s

[boelboel]

You're right, what did happen around that time is that founders would be kept in charge of their companies instead of being replaced by CEOs (e.g. Founders Fund). The combination of the founders staying in power facilitated with dual class shares/ZIRP in 2010s (Google had their 2004 IPO) would be more correct.

[Schlagbohrer]

So in other words, "Thanks Obama"

[busterarm]

I had a friend who worked for Instagram post-acquisition left and came back to a team in Facebook.

They had always sung the praises of Instagram's culture but said they didn't recognize the company that they came back to. Literally night and day between the best and worst place they'd worked.

[PaulHoule]

Instagram hasn't updated its web application in years -- leaving out the filters which I don't use, unfunded Mastodon has an easier, faster and more reliable interface to upload photos.

[petesergeant]

That's because they don't want you to use the web application

[rozap]

My experience is a bit different. They constantly update it and make it more unusable. At this point it's totally broken. Occasionally I open it, it loads for 30 seconds, I have a notification icon for a post I saw a month ago, then my feed shows the same picture of a guy I follow driving his rally car, which was posted several months ago, then I log off. It's truly bizarre what they've managed to do.

[PaulHoule]

That's back end. With the money they make they could have rewritten the front end so it doesn't have to reload everything whenever you upload an image.

[toast0]

Having had my employer bought by FB, there were some exceptional teams in FB that were nice to work with, but those weren't really running like regular FB teams, so there you go. I really hoped the FB orgs would change, at least a bit, to reflect the purchased org, but I guess that was my repressed optimism showing up. :P

[smrtinsert]

If its any consolation, it doesn't work. Just from reading about the company and watching its business operating principles and innovation track record, its very clear what is going on inside - at least from someone with experience in the overall tech industry.

[newobj]

that may have been true in product orgs but the infra and dev infra orgs were pretty strong imho.

[aylmao]

+1. IMO the company with the strongest web infra/tooling teams for example.

Also very strong teams working on db/data, source control, performance, reliability, pl/compilers, etc.

[ironman1478]

I worked in infra as a firmware engineer and it was something else. Very amateurish

[classified]

A people hire A people, B people hire C people. Zuck is a shit CEO and he has shittier lieutenants which makes sense because having anyone competent under him would make him look bad, comparatively.

[throwsame12304]

One aspect of this sad situation that has not been reported is the degree to which Meta's H-1B hiring drives the politics of hiring and firing in the engineering org. Meta received positive attention for a lot of the US-citizen hires it made soon after Trump took office. The vast majority of those people were red-shirted and purged in the recent rounds of layoffs. The H1B loyalties within Meta, the reluctance to disrupt someone's life who is in the US on a visa, mean that US citizens get the boot first, even if a significant portoin of the retained international engineers speak English poorly and mostly manipulate JSON files. As has been reported elsewhere, those "meets expectations" engineers are now basically doing RLHF to improve Meta's internal models. Thousands of them. So I guess it's a toss up which group is worse off, those who were laid off or those who remain employed to do brain-numbing work. How telling that they are doing data-labeling because Alexandr Wang knows he can't trust the company he founded to deliver quality. And finally, I'll just note that the engineering and product orgs of the most popular social media platform in the US is staffed by citizens of America's greatest adversary. Great job, Zuck.

[zeroonetwothree]

That wasn't my experience at all, I found the orgs I was in (infra) to have great culture.

[torben-friis]

Probably well run compared to the rest of meta, not to the versions of themselves pre acquisition. Whatsapp is worsening by the day.

[junior44660]

They introduced desktop version, dark theme and multiple accounts post the acquisition by meta. Hate the ad and social stuff they're shoving down, though.

[vjk800]

In other words, Meta (excepting maybe the bought departments) has become just like any other traditional mega-corporation. Ugly, inefficient and disorganized, but terribly good at doing the one thing that makes it money.

[cletus]

I have to disagree with this as someone who worked at Meta (and Google).

WhatsApp was a textbook example of how not to do an acquisition. The story I heard was when it was acquired, a spreadsheet went around and everyone basically decided what level they were in the (then) FB job ladder and all the engineers said they were E7s (Senior Staff SWE). The way PSC worked, WhatsApp at the time was only ever calibrated against themselves (from what I heard). It had become a fiefdom and, as someone who was on a team that tried to get them to do anything, the experience was awful.

IG was handled better but it was also an almost nonexistent team when acquired, which might well explain it. They stuck with their Django/Python codebase and (IMHO) that was a mistake. The amount of duplication that we had to do for IG specifically was embarrasing. The framework and tooling FB had on the product side was light years ahead of what IG had. IG used to have a very good product focus but I think that's long dead now. It was good because IG had a clear vision for their app and ultimately (IMHO) management had a different view to "grow". They briefly tried to launch another app (IGTV) that flopped, hard. There were a bunch of UI/UX changes that clearly showed the focus had become simply following celebrities instead of sharing updates (eg where the post/compose buttons moved to).

I mention Google because I saw the same things happen at Google.

Youtube was (and my guess is, still is) its own entity. Culturally, Youtubers don't see themselves as Googlers. They didn't (AFAIK) use Google3 or any of the other stuff most of the rest of Google did. But Youtube itself was perceived very positively, technically, particularly in relation just general encoding/decoding infrastructure as well as Bandaid (where racks are shipped to ISPs to cache videos).

Android was another acquisition that prided itself in not being Google. This was very much fostered by Andy Rubin while he was still there. Obviously Google needed to write Android apps but I got the sense that it was always Google engineers who solved all the problems whereas Android just didn't care. They cared only about shipping Android. Fuchsia was an Android offshoot.

Docs and Maps were both acquisitions but they went fully Google3 and were different orgs but weren't seen as separate. The engineering director of Docs (Fuzzy) had, from what I can recall, a very positive reputation beyond Docs (now Drive).

Doubleclick was also an acqusition but went fully Google and you'll find a lot of people who don't even know it was an acquisition.

I don't know what org you worked in but they all vary. My own experience was that Infra orgs in comparison to Google were primitive and barely above just running random Docker-like (Tupperware) instances with a godawful variant of C++, probably started by someone who had done C++ at Google and had decided they really wanted mutable function parameters and exceptions for no particular reason.

The thing I really respected about FB product orgs generally was that really did ship things quickly. I used to joke that the smallest unit of time at Google was a quarter. God help you if you eneded another team (under a different VP) to do something. You'd have to spend a quarter arguing with them to get them to add it to their OKRs for the following quarter.

At FB the timeline for launching a new thing to a limited audience was measured in weeks. The biggest barrier usually was the weekly build cycle for the blue app. The release cycle for Web was S-tier and (IMHO) the people who worked on the infra for Web were generally god tier. This was another reason why IG doggedly sticking with Python just created problems.

There are many thigns you can criticize Meta for (eg the stupid crypto, the billions wasted on VR) but the Web Foundation and Ent teams were god tier and I'll die on that hill.

Anyway, even back then the ML teams and infra, not to put too fine a point on it, sucked (IMHO). Newsfeed was OK but the recommendations for a lot of things like videos just sucked. All of this was mainly because it all relied on daily offline jobs. And then Tiktok came along and showed everybody (including Youtube) just what a bad job they were doing at recommendations. And don't get me started on the IG Reels dumpster fire.

Oh ok, I thought of another one: Messenger. I knew some smart people on Messenger but overall the product and the infra were, again, a dumpster fire.

[MarkSweep]

That's funny about your last line about Messenger. When I was at Facebook in 2020 and 2021 I thought the "Lightspeed" version of Messenger [1] was something of an engineering marvel. Specically the way it was centered around SQLite as the datastore, how the sever side could integrate with it, and how the UI was power by this reactive-style framework written in C. With lots of testing to keep binary size under control and things working reliably. And the DX was not bad for all these constraints. I think CG/SQL [2] is one of the few open source artifacts from this effort.

There were other reasonably nicely engineered things there, like some of the code-as-configuration things in PHP like the ORM and the structured logging. Where you just write and commit code and the databases and data warehouses just get set up automatically.

The choice to have a monolithic web stack at Facebook had some benefits. The latency from landing a diff to being in prod was a couple of hours, not matter what part of the site you touched. And given an arbitrary URL, there was a tool that used static analysis to tell you which file handled that request. Compared to Google where every damn thing is a micro-service with its own variable release schedule and different ways of doing request routing. Trying to figure out where a request goes is combination of dynamic tracing of the RPC trace behind it plus a bunch of grepping through the related code bases.

All that said, I was far away from doing real engineering at Facebook. I accidentally joined a growth team. Which meant our teams remit was churning out all kinds of experiments to see what we could to do to make our team's number go up. I did not find this type of work enjoyable and quit. So I have a narrow perspective.

Also to give these multinational corporations some benefit of the doubt, some of the way they operate is path-dependent. They have hundreds of millions of lines code. So they can get to a point where it makes sense to hire a bunch of compiler engineers to fork PHP and make the Hack language. Whereas a startup would almost certainly never create a new programming language to write their product. To take solutions of this context and apply them to the outside world might not make sense.

[1]: https://engineering.fb.com/2020/03/02/data-infrastructure/me...

[2]: https://cgsql.dev/

[ksec]

Thank you for the insight this actually answer a lot of questions or suspicious I had with Google. Surprised about Android and now Fuchsia makes sense.

I wonder if anyone has worked in Google / Facebook and Shopify and I wonder how they compare.

[petesergeant]

> IG used to have a very good product focus but I think that's long dead now

As a user, Instagram is about the only part of the Meta empire that doesn't suck

[neonstatic]

> I thought of another one: Messenger. I knew some smart people on Messenger but overall the product and the infra were, again, a dumpster fire.

That makes sense from a consumer point of view. I don't use FB much. In fact, I recently logged on a few times after a few years of a break. It's terrible. You can't even get conversation history anymore. You can talk to someone via the app on your phone, open Facebook in the browser on your computer and not see the messages there. Makes me wonder why is Meta still around?