Amazon Takes Fresh Stab at $16B Housekeeping Industry | Hacker News Reader

Amazon Takes Fresh Stab at $16B Housekeeping Industry

bloomberg.com

54 pointslxm80 comments

Amazon and Google will end up being the middle men for everything.

You simply can't compete with a company that has all the data, all the money, the best engineers, the patents, etc... Once they like your business model, you are out of business. This is happening at an increasingly faster pase. There will be an Amazon AirBnB, and Amazon Uber, Amazon Food, Amazon Freight, etc... the future of free markets is not free. They will own every slot in the monopoly game and we will receive a basic income to keep the game going.

> Amazon and Google

A couple of weeks ago, your list would have included Facebook. The future is a lot less deterministic than you think.

Yes, but does this make OPs forecast any less possible?

The OPs forecast is predicated on an understanding of companies like Google and Amazon being so big and powerful that they are immune to the ebb and flow that define normal market conditions and allow new entrants to prosper and old companies to whither when they no longer create enough value for consumers.

My argument is that when Facebook, a company until a few weeks ago commonly understood to be every bit as immortal as Google and Amazon, can be so deeply upset by a scandal that isn't even predominantly their own making (I don't mean to imply that FB isn't culpable, but CA did the bad stuff and explicitly and broadly broke terms in doing so), this immortality is a lot shallower than it seems. Facebook isn't dead, and won't be for a long time, but the idea that Facebook moving into Airbnb's space will immediately monopolise it for the detriment of consumers seems rather more of a silly fantasy than a "forecast".

No it doesn't.

It merely points out that the possibility is very low in the absolute sense.

The future is a lot less deterministic than you think.

Unless our current understanding of physics is completely wrong, the future is entirely deterministic.

I think you mean to use the word predictable, which insinuates that we don't have enough information to accurately model the future.

I believe Mr. Heisenberg would like to have a word with you.

Mr. Heisenberg would like to have a word with you about classical versus quantum phenomena.

We don't use quantum physics to measure significant amounts of time, we use classical physics. In addition to the uncertainty we encounter at that scale, the arrow of time is often irrelevant anyway.

You seem certain of that, but I don't think it's been quite that clear for some time. I'm not a physicist, but AFAIK the Copenhagen interpretation essentially requires indeterminacy. Suffice it to say that I don't think it's as clear-cut as you do :)

Quantum mechanics has no bearing on the macro scale that classical physics inhabit in our current models. We currently have not synthesized the two models into a singular theory.

Classical physics is deterministic. I thought this was common knowledge...

The whole point of quantum mechanics is that classical physics doesn't fully describe the world. Obviously phenomena that require quantum mechanics to describe them are rarely observable at the macro level, but it's clearly not the case that they are never observable because otherwise we wouldn't have been able to make the observations that led to the idea of quantum mechanics.

It seems readily apparent that the future is not predictable solely using classical physics, because the development of the theory of quantum mechanics is something that happened and it could not have been predicted solely based on classical physics.

The whole point of quantum mechanics is that classical physics doesn't fully describe the world.

Quantum mechanics doesn't have a point, it's a theory based on observation of quantum phenomena. It is currently irreconcilable with classical physics, which we still use to model the universe on useful scales.

Quantum mechanics therefore are irrelevant here. The current standard model, which is the best we've got right now, is deterministic on classical scales.

Which brings me back to my original comment:

Unless our current understanding of physics is completely wrong

It's at least a little wrong, but to say that our theory of classical physics isn't deterministic is just wrong. Until we have a better theory, that's just what we have to go with.

I'm not saying that classical physics isn't deterministic. I'm saying that classical physics is provably not a fully accurate model even at macro scales. For most problems at the macro scale, our best model is deterministic, but it is also easy to construct scenarios where macro scale events are not deterministic unless quantum phenomena are also deterministic, so clearly our best model is one that acknowledges that macro scale events might not be fully deterministic.

Show me such a case in the macro scale where the arrow of time reverses or disappears, instead of just postulating.

I don't understand what you are asking. Clearly quantum phenomena are observable at the macro scale. Every quantum mechanics experiment is an example of this, e.g. the double slit experiment.

How is the double slit experiment an example of non-determinism?

Let's say that I set up the double slit and then decide that I will make a macro level decision based on the observation of which slit a particle passes through. The observation of which slit the particle passes through is non-deterministic according to most interpretations of quantum mechanics, so my macro-level actions are now also non-deterministic.

Alternatively, take the example of Clinton Davisson and Lester Germer in 1923 and then attempt to predict what papers they will publish in the future using classical physics. Clearly it can't be done because it is dependent on quantum effects.

> The observation of which slit the particle passes through is non-deterministic according to most interpretations of quantum mechanics, so my macro-level actions are now also non-deterministic

That's not true. The double slit experiment shows that as the arrow of time moves, wave patterns predictably emerge.

Again, until a better unified theory of everything emerges, we currently have no synthesis between quantum and classical physics and thus measured quantum phenomena are currently irrelevant to physics measured in the classical scale.

> Alternatively, take the example of Clinton Davisson and Lester Germer in 1923 and then attempt to predict what papers they will publish in the future using classical physics

This is simply a matter of not having enough information. The amount of information required to "predict the future" in the sense you're talking about is ginormous.

Your concept of determinism is flawed.

Further reading: https://www.lesswrong.com/posts/JrhoMTgMrMRJJiS48/decoherenc... https://www.lesswrong.com/posts/eWuuznxeebcjWpdnH/the-so-cal...

> Again, until a better unified theory of everything emerges, we currently have no synthesis between quantum and classical physics and thus measured quantum phenomena are currently irrelevant to physics measured in the classical scale.

That means that we don't know how quantum physics and classical physics interact. It doesn't mean that we know they don't interact. Clearly they do in some way or else it would be impossible for us to make the observations that led the development of quantum physics.

> This is simply a matter of not having enough information. The amount of information required to "predict the future" in the sense you're talking about is ginormous.

That information doesn't exist. Clearly it is not possible to predict the existence of quantum mechanics using only classical physics. If it were, quantum mechanics would be part of classical physics. Even knowing the entire state of the entire universe, you could not predict that they would publish a paper about quantum mechanics based solely on classical physics.

Did you read my links?

I never made a claim that classical physics can predict quantum mechanics. In fact it's the opposite, I said the theories have not been synthesized.

We are no longer arguing about your original claim of non-determinism in the classical scale, and you don't seem to have a comprehensive enough understanding of the concepts involved here and you're not interested in changing your mind, so I'm done arguing about this. I really suggest you read up on decoherence.

The subject was whether Facebook's stock price going down could be predicted based solely on classical physics. I am arguing that it cannot.

I am arguing that classical physics can not in general predict human behavior. Publishing a paper on quantum mechanics is a human behavior. Classical physics cannot predict whether someone will publish a paper on quantum physics, so clearly it cannot predict all human behavior.

This in itself is not enough to prove that classical physics cannot predict Facebook's stock price, but since Facebook's stock price is dependent on a vast number of human behaviors, it should cast serious doubt on the idea that the stock price could be predicted using classical physics.

Is there some part of this that you disagree with?

I am arguing that classical physics can not in general predict human behavior.

That has nothing to do with determinism, and everything to do with incomplete information.

Your postulate was such:

It seems readily apparent that the future is not predictable solely using classical physics

That means the future is not deterministic, but our current standard model says otherwise on scales that matter for this kind of situation.

The inability to predict human behavior and other external events is entirely because of lack of complete knowledge about the system.

That doesn't mean it's not deterministic, nor unpredictable given the right knowledge.

You fundamentally misunderstand these concepts you are arguing, and again I recommend further reading before you continue to engage in such debates.

> The inability to predict human behavior and other external events is entirely because of lack of complete knowledge about the system.

That is clearly not the case. You cannot predict the contents of someone's research paper about quantum physics no matter how much foreknowledge you have using only classical physics. That is a macro scale event that cannot be modeled using only classical physics. That is proof that classical physics can not model macro scale events with perfect accuracy.

You're using circular logic. Classical physics can't predict the future because classical physics can't predict the future. Try again.

It is not circular logic.

My premises are:

1. Classical physics cannot predict the results of at least one physics experiment. This should be uncontroversial because otherwise the field of quantum mechanics wouldn't exist.

2. The results of said experiment can be observed at a macro scale. This should also be uncontroversial because the publication of experimental results is observable at the macro scale.

It follows directly from those two premises that there is at least one macro scale phenomenon that classical physics can not predict.

Amazon and Google are engaging in concrete industries. Google is still the largest advertising company in the world, facebook is not. Facebook has been and will be second fiddle to Google in the ad industry.

"His" list wouldn't have included Facebook because they haven't really dominated anything, let alone stretched their arms out into other fields. Facebook can only hold the dimmest of candles up to Google and Amazon. There are giants, then there are titans.

New ideas thrive best without constraints. Large companies may have more resources but they must have institutional constraints to manage a large workforce.

Startups will always have a leg up on new ideas because of their flexibility. There’s no bullshit holding back a small team from executing. That’s a big reason why startups are so effective at pulling the rug out from underneath larger companies who seem to be invincible.

Consider what startups are/have actually been building... collectively.

There are a lot of individual startups, but most of the money driving them comes from the market cap of "tech giants" directly or indirectly. First, VCs invest based on the probability of a startup becoming a tech giant. Second, most exits are buy-outs by existing giants using their market cap as currency. Ultimately, either new giants are created or existing giants are made bigger. That's where all the startup money comes from, and that's where it ends up in success scenarios. Startups are a part of that, not the antidote to it.

What are these tech giants, and why are they worth so much?

As Thiel put it, monopoly. MS learned this game early. They understood, platforms and moats, network effects and ringfenced markets. Google is a search monopoly. Amazon has its fiefs. Uber raised all that money because it looked like they could become a monopoly. Now "their" market looks a little more competition-friendly. There are actual market dynamics pressuring prices down to marginal costs... Hey! how are we supposed to justify $100bn market cap with competition?!!

Unicorns are a dime a dozen (well $1bn, technically). Unless the market crashes, we'll probably have 3-5 dragons(TM) within 3-5 years worth >$1 trillion each. They will all be relatively unexposed to market forces.

> the best engineers

You're living in a bubble. Lots of companies have better engineers on an average.

I am not sure about that, but what I meant is that they can afford to hire the best in their field.

The railroad barons are going to be the middleman for everything. They have all the best engineers and control all the suppliers.

They were for a while there, and it took legislation to fix it, it didn't just magically get better.

I am always surprised by the turn of events when legislation is not involved.

Several years ago, documentaries were coming out attacking Walmart over everything. Now it's Amazon. I don't hear as much about Walmart from the rabble-rousing press anywhere now. Moreover, Amazon ended up killing many of the mom and pops anyway. We could not get a Walmart near our city because of the political opposition, and in the end, the political opposition only delivered us a different overlord.

I'm not sure exactly what you're getting at. Both Amazon and Walmart have problematic side effects on society. The general trend over this period has been more consolidation of corporate market power and rising income inequality.

The idea that these problems don't just fix themselves is certainly supported, rather than undercut, by these observed realities.

I think the point is that a narrative of "Amazon is destroying awesome local companies... like Sears!" is both misleading and not useful in any particular way.

A more general narrative of the dangers of our current "winner take all" market environment might be useful.

I think about this quite a bit. Why does there only have to be one boogeyman?

I also remember when Wal-Mart was the company destroying boogeyman. It was easy to see why. Their sales in electronics, some furniture, housegoods, and some clothing were causing massive erosion to some key competitors. It wasn't really mom and pop shops either, it was places like Sears and other big box stores. And, of course, other shops were eating other businesses for some of them as well (for example, Home Depot and Lowes having a devastating impact on appliance sales at Sears).

But somehow, the media somehow decided that Amazon should be the new boogeyman. Thinking about that list above, are those even things that I would buy at Amazon? Generally, they are not.

Amazon has had a huge impact, but not necessarily in the obvious way that gets bandied about in our politics and our press.

Nah, both companies have a major weaknesses that no amount of money seems to be able to solve. Namely, user experience.

Example 1, Dropbox: Google & Amazon seem like companies ideally positioned to destroy Dropbox. They excel at making network infrastructure at scale, they both have competing products and massive existing customer basis. Yet Dropbox has managed to compete and win on user experience.

Example 2, Netflix: Both google & amazon have tons of money and massive amounts of users to cross promote a premium video service to yet neither YouTube nor Amazon seem to have figured out how to produce consistently good original content to the same degree that Netflix has. Netflix should be dead by now yet it’s doing very well. YouTube/Amazon also have much poorer interfaces than Netflix.

I could go on. My point is that plenty of companies successfully compete with both Google & Amazon.

Not the most airtight examples. Dropbox was first to market with an innovative service, they've been losing market share ever since Google Drive which now has 25%+ of the market. Netflix was also an innovative service that was first to market. So now there's a new monopoly type of a company who dominates a new market > cable tv. Smaller company can't oust bigger ones unless it is actually possible for them to #1. Out innovate and implement a better service and #2. Hope the monopolies don't catch on to your idea and overwhelm it with their resources (i.e what Blockbuster could have done by adapting an online model). Those 2 critical circumstances don't have all that often.

In the case of Netflix, there are also some distinctions between their strategy and both Google's and Amazon's strategies that can't easily be erased.

Netflix has a 100% subscription model that includes original content plus enough licensed content (mostly TV) to fill things out.

Google is basically an engineering organization (ADDED: that makes money through advertising) that probably doesn't have a lot of interest in becoming a studio or in licensing a lot of content.

Amazon wants to offer subscription video as part of a broader bundle at a price they want to be attractive even for people who don't use the subscription video a lot. And they also want to sell video a la carte whether online or as disks.

I think the playbook at these two companies is very different. Amazon will eat your business completely, but Google will mostly just take a cut off the top. I say this since I don't think Google is willing to hire large amounts of people who aren't software engineers.

> The online retailer is swapping the low cost of contract workers for the greater control of employing its own people. Doing so puts it on the hook for things like minimum wage, workers compensation and overtime pay. But it also lets Amazon determine how the workers are trained, which cleaning products they use and how they organize their schedules.

As an Amazon employee, I very much hope this trend continues. Amazon logistics has gone the industry standard route of contractors hiring contractors and no one seems to be happy with it at all. When Amazon is the employer, Amazon can control quality better. Customer obsession means more than just lowest possible price.

(My own personal opinion and of course I don't speak for the company)

I think the fact that cleaning services are a lot less standardized is what drives the need to hire people. There's very little difference in logistics: its both easy to do an average job (just take the route google maps suggests) and its easy to monitor if contractors are doing a reasonable job.

The impression I get for cleaning services from people who have hired them is that the person doing it will generally do the smallest/easiest amount of work possible in the allocated time, which leads to poor customer satisfaction/retention, so exercising more control as an employer is necessary.

Does anyone remember HomeJoy? See [1] and [2] below.

Amazon isn't necessarily using this business model so that it may differentiate itself from hands-off matching services. In this case, regulation and legal battles that have destroyed entrepreneurial enterprises like HomeJoy are guiding Amazon. Improved customer experience is a side effect.

[1] https://www.forbes.com/sites/ellenhuet/2015/07/23/what-reall...

[2] https://www.recode.net/2015/7/17/11614814/cleaning-services-...

Aaron Cheung, co-founder of HomeJoy, launched Homeaglow which seem like the same business model as HomeJoy. Xiao Wei Chen, COO at HomeJoy, is also a co-founder of Homeaglow.

I'm not saying it'll succeed this time, but two execs at HomeJoy have decided to dedicate another 3 years of their lives to the idea with Homeaglow.

I know that the articles mention the lawsuits over classification of independent-contractor vs employee, but nestled in there is a huge customer acquisition cost against a low lifetime value. Offering an initial $19 cleaning means very high acquisition costs and you're getting customers who probably won't return with $100 a month later. The Forbes article notes that Handy has a much better customer retention rate and recurring revenue is so important.

It also sounds like HomeJoy might have been over-staffed. Re/code notes that 20 employees of HomeJoy would be joining Googles product and engineering teams. That seems like a lot of salary. I don't know what HomeJoy's site was like, but Homeaglow's seems simplistic - some PHP just meant to handle the booking/matching/reviews. Even small things like the forms are the kinda simplistic looking ones where the zip-code box takes up the full width of the form on a new line (the way it would if you grabbed Bootstrap and didn't want to customize anything) rather than having a form with a more natural City, State, Zip layout. But it gets the job done and looks like it can be maintained with minimal effort.

If you have 20 employees of Google-caliber, you have to imagine a lot of salary cost. From LinkedIn, it looks like Homeaglow is a 2-person shop.

It could just be that while it's a large industry, they had trouble retaining customers and had a high run rate at HomeJoy. With a lower run rate and some alterations, Homeaglow might be a happy business. They don't seem to discount first cleanings hugely. It looks like they offer a free half hour on a 3 hour cleaning for first-time users, but that means that someone is willing to spend a hundred compared to someone that just sees a deal at $19. They also only allow recurring plans that you can cancel, but it gets you into the mindset that this should be a weekly/bi-weekly/monthly recurring expense.

Data point: I just moved and Amazon sent me a $20 off cleaning offer, which I used to clean out my old rental this past weekend. They did a nice job (I think, haven't gotten my deposit back yet). I was talking to the cleaner and she said she used to work through HomeAdvisor exclusively but has put HA on hold because Amazon takes less commission and keeps her just as busy.

The price was OK with $20 off, I know you can do better if you search out the right service, but I imagine that you can already ask Alexa to schedule a house cleaning, so the integration & convenience would have to be pretty compelling for me to make it a regular thing.

The general problem with "Uber for X" is that mostly consumers don't want one-time anonymous services. It works for cabs and probably for one-time cleaning, but if I'm going to have someone clean my apartment every week I want the same person there every time, and I think that's pretty typical.

> A weekly cleaning of a 1,500-square-foot home runs about $156.

It doesn’t say what’s included but off hand that seems 2-4x expensive for a weekly house cleaning of that size.

That (relatively) high pricing may be specific to Seattle though as they’ve got a high ($15/hr) min wage laws there that would apply. Combined with non-contractor employees, FICA, and benefits, they’re easily looking at $20+/hr in labor costs.

I have housecleaning every few weeks. It's about $100 and takes 2-3 hours for about the same square footage.

Came here to say this - seems like what's included in a nice big markup for Amazon ;)

> "Any products that require additional services beyond opening a box, Amazon doesn’t do well," Kalyanam said.

Well, there’s AWS, which smashes that mold.

How's AWS human customer service?

Quite good for us. We have a team from AWS on-site every other Thursday (and subject matter experts from AWS that we can call in as/if needed).

AWS started out as a REST-services-only take-it-or-leave-it business. They've expanded quite substantially their human-provided sales, service, success operation in recent years.

Even on my personal account (spending <$50/month), I get quick and sensible email replies to my account-related inquiries.

Pretty good if you're a big fish. I've been mostly satisfied with their human CS.

Amazon seems to do a mostly reasonable job of customer service, especially given their scale. However, I'd argue that's something quite different from managing an army of low-wage contract workers. Yes, they have their warehouses and there's even Mechanical Turk (though you don't really hear much about that these days) but those seem different and much more constrained that people providing personal services of various sorts. That's generally proven to be a hard type of thing to deliver at scale.

I'm in the weird position where, we are family planning, but (to our knowledge) unsuccessful, but started getting a series of emails from Amazon congratulating us and offering up suggestions like vitamins, baby registry and, of course, housecleaning services w/ discounts available.

(The first email was, to put it lightly, premature. Now we got one that says "week one", and it could be right, or not. We don't know yet.)

These emails puzzle me, because I pored over my shopping history, and I can't find any indicator that would tip off their algorithm.

But I do get the idea of trying to sell things to expectant mothers. Such as convenience services.

Maybe they are purchasing data from other providers about your search history, site history, etc.

We may not be there yet, but there's certainly going to be a day where online advertisers know a woman is pregnant before she does based on the changes in her behavior.

I'm not sure if that counts as a dystopia or not. Is there a good word for "Weirdtopia"?

This article talks about how Target predicts when a woman is pregnant, and the steps they go to to make it less weird:

https://www.nytimes.com/2012/02/19/magazine/shopping-habits....

One of my favorite incidents and exactly what I was thinking about.

We are not far from that point, with Target who figured out a girl was pregnant before her dad knew:

https://www.forbes.com/sites/kashmirhill/2012/02/16/how-targ...

Helpful?

I mean honestly, the whole point of computing is to do things we as humans can't do with data to make our lives easier. Now that it's actually doing those things, people think it's creepy and scary.

If a computing system knows you are pregnant before you do then maybe it can help you take steps (like taking pre-natal vitamins, abstaining from heavy drinking/smoking, starting a savings account etc...) weeks before you would have normally. How is that dystopian?

If they just sent you a notification that you’re maybe pregnant I would agree. Instead it starts with advertising.

I can't help but think about how it would give you more time to consider carrying to term vs abortion, since there is a hard time limit and earlier is better in terms of stress on the body and ethics.

This would make it a target for anti-choice groups as well as internal company ethics, however, so it's a complicated issue.

While "notifying" that you "are" pregnant, according to data and calculations seems helpful, the fact that a lot of apparent pregnancies end in miscarriage[0], or just the emotional roller-coaster that is family planning (particularly when it requires additional medical assistance) takes a toll on people, means that these well-meaning emails can really be a negative trigger.

[0] https://www.webmd.com/baby/guide/pregnancy-miscarriage

"According to the March of Dimes, as many as 50% of all pregnancies end in miscarriage -- most often before a woman misses a menstrual period or even knows she is pregnant. About 15-25% of recognized pregnancies will end in a miscarriage."

Amazon probably sees your browsing history from embedded affiliate ads on the websites you’re visiting.

I can only assume. It's annoying that I can't (trivially) block that level of information gathering. I use AdBlock and I disable 3rd party cookies. I personally don't think I've looked up anything that would "tip my hand", but I'm sure my wife has used my computer/browser enough that things she researches have leaked into the corporate overlord databases.

https://en.wikipedia.org/wiki/Brave_(web_browser)

Is this a trivial change in user behavior from using Chrome?

(My assumption is that, at the very least, I'll need to switch from using my password manager extension to using a separate app, and having to copy/paste. I'll also probably have to export/import bookmarks. I also use Chrome profiles heavily. Are all of these things trivial to migrate?)

Brave has built-in support for multiple third-party password managers.

Chrome profiles - I dunno.

This would be an interesting case for a GDPR request.

mere curiosity is enough for such a request it needs "legal effects concerning you or similarly significantly affects you"

see 22[1] https://gdpr-info.eu/art-22-gdpr/

I think it is a common social practice to disclose a pregnancy at the time of your choosing rather than amazon's.

A premature disclosure can significantly affect you.

Have you been shagging in a room with an Alexa?

That's actually a good idea. I mean, it's a terrible idea, but certainly one way they could do this. But, alas, no. We don't own an Alexa speaker, though we do own a Fire HD. That, however, isn't allowed in the shagging room.

ETA: Too late to edit, but the point I'm trying to make is that Amazon is able to gather data on potential customers, and it's possible, even likely, that when you get pregnant, you start to think "paying for housecleaning" is worth it, and the combination of knowing if this is true, and being able to predict (or know) when someone is pregnant through their data mining and algorithms gives them a position to advertise directly to newly expectant mothers with these housecleaning services. This is, obviously, a big advantage over their competitors.

"How Companies Learn Your Secrets" (2012)

https://www.nytimes.com/2012/02/19/magazine/shopping-habits....

Cached Version: https://archive.is/SdQas

They have the keys to your house through Ring and others. So now we see their conquest of our homes coming to a completion.

as a customer, I think it's great. Few purchases make me as happy as the amazon house cleaning.

Reply on news.ycombinator.com