Google over-hired talent to do ‘fake work’, claims Keith Rabois | Hacker News Reader

Google over-hired talent to do ‘fake work’, claims Keith Rabois

finance.yahoo.com

57 pointssigna1164 comments

I was with him until "[Musk's] axing of half the Twitter team since his takeover in October should provide inspiration to other tech bosses". While yes it is likely true that Google and other tech companies have excess staff, there is a right way to handle that situation so it doesn't adversely affect the business, and what Elon did is not it.

Twitter's problem was cash flow and avoiding getting margin called amid falling ad revenue, high expenses and interest payment obligations, all related to Elon's buyout. They needed to stop the bleeding, and fast. Large tech corps are meanwhile sitting on hundreds of billions of dollars in their bank account and posting profits of tens of billions every quarter. It would be idiotic to further disrupt their operations for no reason.

Should Google do more rounds of layoffs, kill employee morale, increase overwork and burnout, shut innovative businesses, send talent to competition and significantly weaken the company's long-term prospects, all for a few hundred million in payroll savings that will go towards stock buybacks? The sad part is that the activist investor class will reply with a resounding – yes.

I’m fine with the tech layoffs and “Musk being right” movement. Mostly because it has generated a large shitlist of companies I will never work for, will never invest in and will never integrate with or recommend.

That's a weird form of investment strategy, what is your yearly returns looking like?

Pretty good so far. I put it all into property and ETFs and beat any market investment I would have made.

Not investing in a toxic and volatile shit show is a good investment strategy.

My father threw a lot of money into Tesla and is breaking even. His entire life is consumed with "oh shit what did Musk say now"

In this case, contrary to your assertion you are probably pretty heavily invested in tech giants that did layoffs, assuming some of those ETFs are index-based.

That's fair. The point is that they are diversified thus limiting the risk of an individual company or group doing something bloody stupid. 90% is in property though.

I don't understand the need to show your ego and chess thumping but I guess "this is Wendy's" so well done sir.

There's a huge difference between "in a company of 150K+ engineers, a non-negligible number of them are probably not very productive" vs "over-hired talent to do fake work".

At that scale, there will always be inefficiencies and imperfections. I also don't think a big mass firing was a smart move long term, but it does make some investors happy for some reason, and gives a short term stock boost.

Quick correction: Google does not have 150K+ engineers.

When you’re hiring a lot of people, a significant number of people will be underperformers - people who are either not motivated or capable of delivering enough value to justify their salary.

In a bull market, large companies don’t want the negative publicity that comes from layoffs. So instead of firing them, they get shifted into underperforming teams.

The axeman cometh with the bear market. That unwillingness to tell people you’re cutting costs suddenly becomes exactly what they want to hear, so you take the opportunity to get rid of all those underperforming teams.

It’s garbage collection at an organisation level. This is capitalism.

> So instead of firing them, they get shifted into underperforming teams.

I work at Google. I've seen this one time. A manager didn't want to do the hard work of firing somebody on their team and then misrepresented their productivity to another manager when that person tried to switch teams. This manager received a bad performance review and left the company (I don't know if it was their choice).

It is easy to move teams within Google (or, it used to be at least). But it is very hard to move teams within Google if you have bad performance reviews. You can't really shift underperformers to underperforming teams at scale.

> It would be idiotic to further disrupt their operations for no reason.

Tech has the worst P/E ratio, even big tech. It is fine if the value is inflated for disruption and future growth, else the valuation should be less than half of what it currently is if we say $1 of tech profit is equal to $1 of any other profit.

And...? Thanks for this non-sequitur.

Please explain how you come to this conlucsion of the worst P/E?

Google brings 1.30$ after tax for each 1$ invested into operations and it trades at 19.5 P/E. This is very conservative.

While I wouldn't say fake work, "busy work with dubious value" holds quite true. Eg spending 5 months a year in the perf/promo dog and pony shows, cloning existing documents and deprecating older ones in the name of inflating artifact count, meetings to show case leadership and influence all point to this.

Wow, that Marc Andreessen quote is incredibly obnoxious, especially given how much money his firm has thrown into an entirely useless industry.

Useless maybe, but still new and constructive. The laptop class he derides is mostly nihilistic and destructive. At the absolute worst he facilitated the transfer of funds from wealthy investors to crypto programmers?

He is the laptop class, just richer and more useless.

You think that guy builds anything with his hands?

He's even more removed from anything useful being done, he's not even the one doing stuff on a screen, he uses his screen to tell other people to do stuff on their screens.

Real rich having that complaint come from him of all people

Yeah that was embarrassingly out of touch seeming.

Yes such useful and constructive new ways organize Ponzi schemes. Andreessen is a member of the laptop class himself, just one of the most grating ones. Seems to think telling other people to build makes him a builder. YIMBY in his tweets, NIMBY in his streets.

> an entirely useless industry

synbio? green tech? crypto? adam neumann? you're gonna have to be more specific

I love (and agree) that Adam Neumann is an industry.

More about Andreesen. Yes, he is investing in the WeWork idiot's latest project.

https://www.youtube.com/watch?v=7GWIR3h2twI

The problem with over-hiring is not that there's a lot of people doing nothing. The problem is that there's a lot of people who need to do something to write in their annual reviews. Since there's only so many useful things to do, people start doing non-useful things e.g. over-engineering simple things. These non-useful things then slow down the development of useful things by

  1. Creating needless technical complexity.
  2. Creating needless compatibility issues.
  3. Distracting the focus.

So over-hiring isn't just a waste of money (which G has plenty of), it is also a waste of time (which G can't afford with competitors, like ChatGPT, breathing down it's neck).

Hiring 2X more people is not going to result in 2X more value creation, but laying off half the staff is not going to 2X the productivity either.

It's not only in companies, the same logic works for any bureaucracies. Examples such as overregulation in EU or large universities bodes well for this argument.

The downside, Rabois said, was that the new hires just had to "be entitled, sit at their desks, and do nothing."

"There's nothing for these people to do —it's all fake work. Now that's being exposed, what do these people actually do, they go to meetings."

I think I see the problem, workers are taking over tasks traditionally reserved for middle and upper management: entitlement, desk-sitting, inane meetings, LARPing as people who have real work to do.

The difference being that for people with actual talent, the state of affairs is intolerable. So you have to wonder what kind of efforts management were exerting to prevent these capable hands from being set to necessary work (of which there is plenty).

Then again, if you've ever worked in the tech industry, you needn't spend too long wondering since your own experiences will furnish you with more examples than you would want to enumerate, lol.

I would be very surprised if anyone at these companies go to their desk to just do nothing. Less productive than should be, sure, but doing nothing would cause general revolts from people that do something (of course there are people that do something, innovative or not new features come and general uptime is held). It's basically impossible to both have an entitled tech elite but also members that don't work among them unless 100% are not working.

Yeah my experience is people who have nothing contribute tend to make frenzied efforts to LARP as people who do something. Often creating needless busy work for others, or creating constant messes for others to deal with. Usually those people are at least in middle management, if not in upper management, so when people are trying to fix or ameliorate the damage they've done, they can't even stand up and say that that's what they're doing for fear of offending these peoples fragile egos. Alternatively, if the culprits are not in management, then management has hired them, but simply may not willing to admit they made a hiring mistake so will deal with any information from subordinates as simple personality issues for the employees to solve themselves.

As a result, the people that work have to take on themselves an additional unpaid responsibility the actual restorative work or, at the very least, the emotional labour of pretending the mess being cleaned up wasn't made by someone, and just fell out of the sky.

This leads to a syndrome called "burnout" and "staff turnover" far more often than it leads to "open revolt". You can't have an open revolt of one (that's just called "a mad shouty fucker who can't hack it"). And you can't come together with others to revolt if you're unsure to what extent people have internalized the state of affairs as being their own fault, or at least, being something that is not okay to be discussed (either for fear of retaliation, or they've internalized the notion that it's an unsolvable personality conflict and the done thing is to be polite).

Washed up failed entrepreneur who hired boyfriend into a quid-pro-quo job and then accused by him of sexual harassment before being forced to resign says: "other companies are making bad hiring choices." in bid for relevance. -- Yahoo (all the news that's shit to print)

Anyone who presides what Musk had done at Twitter is obviously either a Musk-can-do-no-wrong acolyte, or an ignoramus who's out of touch. I'd bet the former, if for no other reason than pity.

I dislike Elon Musk, but apart from a few stability glitches the product seems fine and product delivery seems to be moving faster than it has in years.

I know everyone is waiting for it to crash, and you may be right, but so far, stripping (hundreds of?) millions of cost out of the business seems to be have been a good strategy from a technical perspective at least.

Only need about 20% of the engineers are dedicated to keeping services running in any tech company.

The other 80% are focused on new product features, growing revenue, growing engagement, growing new user signups, l blocking bad actors and legal compliance.

Having worked in a "growing engagement" role at a big company, I'd say it's more accurate to describe it as "running no-op experiments over and over until the A/B testing tool tells you one had a 'statistically significant' result". It seemed to me like the exact kind of make-work Musk fans criticize.

Axing all of those roles could make it easier to ship valuable user-visible features, since the remaining engineers wouldn't have to work around feature flags for fifty ongoing experiments.

Let's not project your role at Company X onto Twitter. Especially, given that neither of us have worked at Twitter X.0

Definitely doesn’t seem like that other 80% did a lot of that at Twitter.

Definitely doesn't seem like you have any data points at all to make a judgement, apart from the revenue collapse of 40%.

Honest question, as a non twitter user, what does "product delivery seems to be moving faster than it has in years." mean? Has there been new features?

> product delivery seems to be moving faster than it has in years.

They launched Twitter Blue which was a disaster, then redid the timeline, which at best is a mild improvement, and they rushed out things that were already in development.

I'm not in the Twitter Death Cult camp, but I don't think we can judge new development based on the last four months.

Twitter services that are publicly exposed have seen a global latency average has racked on 150ms with regular spikes of several seconds or more. [1]

Core services for Twitter have been throwing 500s a dozen times every few hours. [2]

Twitter performance for users in Australia has fallen apart twice due to what was found out to be someone losing access to a root certificate and mitigation delayed because the team that could fix it is no longer employed. [3]

This is far more than a few stability glitches.

[1] https://deadbird.singlepane.io/d/hI9vrUO4k/home?orgId=4&refr...

[2] https://deadbird.singlepane.io/d/hI9vrUO4k/home?orgId=4&refr...

[3] https://deadbird.singlepane.io/d/hI9vrUO4k/home?orgId=4&refr...

Hmmm private equity licking their lips. Tech is still over valued in P/E terms so they haven’t been biting but expect a lot of companies purchased and taken private. If you are in the right position you can make a lot of money on the upside as an IC if they cut you into the turnaround. Otherwise go find a new job.

Why can’t you be a bit of both? Twitter was awfully run by all means. They were unable to turn large profits even during the pandemic.

Musk’s handling of Twitter has also been pretty poor by all means.

People have been treating this like some political issue with sharp partisan divide. Which I don’t get at all - it’s a badly run business taken over by a new boss who is out of his depth. Nothing more, nothing less.

I detest Musk as much as the next guy (or more, if my history on HN is any proof). I think he's a despicable human being; I would not be caught dead in the general vicinity of a Tesla, etc.

That said, for the casual user, Twitter works pretty much the same as it did before he took over. It hasn't crashed durably. Makes one wonder what all those people were doing.

Obviously 100% of the engineers weren't just keeping the lights on. The long-term future of Twitter is way more bleak right now than it was a year ago, and that has, in part, to do with the fact that all of the insane engineering talent at Twitter is gone.

Twitter has some insane infrastructure, which was way ahead of it's time, and which is arguably why the company is still doing fine.

> Makes one wonder what all those people were doing.

A "good" social media giant needs huge headcounts in departments like Troll Control. Last I heard, the EU was threatening to sanction or ban Twitter - because the gutted Twitter staff wasn't keeping up with such things.

Yes, maybe, but I haven't seen all those supposedly new trolls (and to be fair, I never saw the bots Musk was constantly complaining about before purchase).

I think it's possible trolls, spam and "bots" exist in a kind of invisible state: they can be found if one looks for them (what regulators are probably doing) but they never appear on most users' timelines.

Speaking of fake work, didn't Meta recently cancel the metaverse?

Keith Rabois?

Sexual harasser Keith Rabois? http://www.huffingtonpost.com/2013/01/26/keith-rabois-square...

Founder of effectively bankrupt OpenDoor Keith Rabois? https://www.nasdaq.com/market-activity/stocks/open/advanced-...

Is Fuchsia fake work?

According to folk like this: yes. It does sadly look like fuchsia seems to have been a victim of google's culling of talented engineers :-/ (disclosure: I did work on fuchsia briefly and the engineers working on it were all incredibly talented and very hard working)

Anything that isn't immediately profitable is fake work, anything that fails to work out means that it was also fake work.

It's more that there doesn't seem to be any kind of synergy between Fuschia and Google's business of selling ads, hence the fake work label. If Google tried to launch a FM radio ad business and it flopped, it would be at least related to their core business.

Now we're just dunking on people who think the purpose of the economy is (ought to be) to connect idle hands with needed work?

Instead, we're lamenting that a company should "at least be trying" to pump more sewage into the river we all drink from.

What a sad state of affairs :)

The goal of fuchsia is/was to create a better platform for future products - a lot of engineering goes into securing the linux base of android which is not free. There was also talk about being "better at modern hardware", "lighter weight", etc ideas but I generally feel that claims like that from any software/platform are questionable: they're not really measurable, and what is "good" could easily be different in a generation or two of hardware/software changes.

The more robust security basis for the platform is a much stronger claim, and remains true as hardware and software change around it. E.g. we can improve the security of a platform by adding CHERI say, and that improves the security of a linux+android based system, but fuchsia would also get those wins on top of the existing more secure basis - the gap may be reduced but fuchsia stays ahead.

So the goal is to create a platform that is more profitable in the long term, so not "fake work".

It would appear that Google decided that there has not been sufficient evidence of success to warrant any further investment, but that does not mean it is "fake work".

Such a definition of "fake work" would mean that any project that you should never do or attempt anything if you do not have guaranteed success.

Note: Google's business model is ads. Android exists for the purpose of selling ads. The long term goal of fuchsia under this model would be a cost reduction in the maintenance of some android style product (whether it would still be branded android shrug). That's the business argument for fuchsia, and it would appear that google has decided that it doesn't think the end result is likely to warrant the cost of getting there. I think that's a shame, and I feel bad for my former coworkers, but I can't say "companies should be required to continue developing projects forever, even if it appears they won't succeed in their goals" because that would also be dumb :-/

I hope so because it’s shameful that they’re wasting all that work on vaporware. You ask people here and you get glares and condescending reminders that “Operating systems take a lot of time to develop, okay?”. Android took five years or less to develop. We’re going on seven years for Fuchsia at the very least, likely more. There’s just no excuse to have wasted all that engineering talent on nothing. They might as well be Aperture Science.

When you're playing with yourself it can take years to make a baby.

In general, OS release 1.0 takes about 4 years. That's feature complete and ready for users, not necessarily production ready and stable.

> Android took five years or less to develop

And it runs on a kernel that started development on 1991. Also, Fuchsia is already running in some devices, - https://arstechnica.com/gadgets/2021/05/google-launches-its-...

...Yes, they literally rewrote an OS kernel for no reason

No reason other than making a kernel with a sane security model?

Like we have a bunch of those under each tree in a park...

Fuchsia has nearly been 10 years in the making but has yet to make any meaningful impact. They probably could have spent a quarter or two with Yocto to create a stripped down linux for Nest devices and called it a day.

Instead they had 400(!) full time staff pulling in $300k+ developing something which nobody wanted and has not had any meaningful impact.

Maybe in 10 more years they will manage to replace Android…

And they literally rewrote a Browser for no reason, And they literally rewrote a bunch of infrastructure for no reason. And they literally reqrote an operating System (Android) for no reason. Oh wait...

Chrome never was a rewrite

Fuchsia has always had an element of "talent retention program" about it. (see https://news.ycombinator.com/item?id=17569631)