The article says Nest may be up for sale, which would make sense. I have to think that the buyer would probably buy Nest for the underlying technology and not the brand at this point, however.
I have an original Nest thermostat but would never buy another Nest product due to the horror stories that continually pop up online. I don't follow Nest closely or have any particular interest in them, but I remember the video they reference in the article from the Google employee who couldn't stop the smoke detectors from false alarming, and I also saw the news when they decided to "brick" a $300 product they acquired.
Based on those two things alone, I don't feel like they take customer satisfaction seriously, nor do I have confidence in their ability to build a good product--even though my thermostat has never had any issues. I doubt I'm alone in this.
Oh, me neither. But today there are plenty of companies in the smart thermostat game. A quick Amazon search shows 8 at a glance, only one of which is a Nest.
As someone who lives in the Northeast the horror stories make the thermostat a non-starter. Even a tiny drop in reliability that could cause an outage in the winter isn't acceptable in exchange for a cool looking device that I can, on rare occasions, use to remotely change a setting that I forgot before leaving on a trip.
The current generation of Nest Protect actually seems pretty good--it's not nearly as bad with full volume false alarms as my previous one was when I open the oven door. But I basically don't care about the IoT aspect of it.
IoT is a perfect use for notifying you if the furnace fails and you are not home since IoT temperature sensors operate independently of the thermostat. I've put a Nest Thermostat in multiple houses now and been very happy with them - the remote control functionality, decent web UI for additional control/scheduling, and the app are simple enough for family members who are not in IT to use.
I agree that one can imagine an IoT system designed by paranoids for northern houses that does things like provide undertemperature warnings, shuts off plumbing systems, etc. I'd considered buying such a system. The Nest though strikes me as a flashy device that potentially decreases overall system reliability.
ADDED: In general, I'm sure a Nest works fine. I just have a use case where the winter reliability of my heating system is really important. I can't control every aspect of that but I don't want to hurt any aspect of it for trivial gains.
Selling Nest makes tons of sense. It's probably still valuable to someone who thought of buying when Google did, and Nest adds very little to Google's IoT/home-automation platform, which is more likely to be Android-based. Moreover it is unclear if, all else being equal, thermostats would be very high on Google's IoT priorities.
IoT is a fad. I have yet to see very many convincing arguments on why my house needs to be online. The existing products (locks, sensors, smoke detectors) are barely different from existing home security products that have been on the market for years. The functionality of Echo/OnHub is largely covered by smartphones. They promise integration with other home IoT devices that do not exist yet.
People are very hesitant to buy new appliances/home products. The last major change to the American home was the television, and that was ~50 years ago. Desktops had their day, but have been largely replaced by Laptops/Smartphones for many consumers. Adoption of new home appliances and technologies is glacially slow.
Also newer electronics have shown they need replaced/upgraded every 2-5 years. You are going to be hard pressed to sell the idea that the consumer needs to replace all their light switches and smoke detectors even every 10 years. I barely replace light bulbs at that rate now.
Cars about the only place where the IoT call home paradigm works. They are complicated machines that need semi regular maintenance. Placing sensors all over it makes sense. Trading updates for telemetry is attractive because a cars computer needs updates like any other computer. But I don't need a fridge that senses when I am low on milk, when I can just as easily look inside it and tell for myself.
EDIT: I think I should clarify that home IoT is a fad. Having machines feed back telemetry is useful, especially commercial machines with a service contract. The data is useful when its actionable. Data for the sake of data in the home is a waste.
It's not exactly the same, but in the UK, Economy 7 is effectively a low tech version of this. At night the price of electricity is cheaper, so electric storage heaters and boilers are switched on then to take advantage of the lower costs and reduce demand at peak times. Rather than using timers, it is controlled by a radio signal broadcast throughout the country.
IoT is a fad in the current way we're going about it.
What's not a fad is getting devices to communicate their health and status to an upstream repository for things like predictive analytics.
This is already happening in the B2B sector. It will reach the consumer sector next, but like you said it will take 2-3 consumer product cycles to make it real.
It won't be about remote control or any of the barely useful use-cases you're seeing with fridges and door locks and water bottles. It will be knowing your home's A/C has two weeks of operation before it seizes up. It will be knowing your home's plumbing has frozen and burst while you were away.
The "TV in the refrigerator" IoT is definitely a fad, but you're right, there's a fundamental need to connect devices together and to a central managing/monitoring service.
There is also a fundamental need to have those devices be secure. Consumers take this for granted because they have been secure for so long. They don't know what's coming. And when they find out then there whole thing will probably blow up. It just takes a few companies to tarnish it for everybody else.
The problem is the lack of corporate liability. If they used some alternative to the internet we'd be reading about how it makes the SWIFT network look secure add long as the alternatives are “spend money and time doing it responsibly” or “ship it now with a strong EULA”.
That stuff kills me when it's marketed like that. It's not IoT! Having a tablet or a TV or Twitter in your computer isn't IoT, it's just a shitty tablet, and it still required human intervention. That's literally the Internet we have right now. IoT would be your fridge talking to the store and telling them you need more milk, and then you get a shipment of milk. But that's infinitely harder than taping a tablet to a fridge, so Samsung sells us a tablet taped to a fridge and it's called IoT.
Connected devices are not IoT. Marketing needs to learn that.
Both of those use cases (AC/plumbing) are good, but cannot be stand alone products. These sensors will be integrated into your next AC unit or PEX manifold. They also need to be robust enough to last the full duration of the unit.
The major issue is the platform. Commercial products already have dedicated terminals, so the platform already exists. But I don't want to have to install the AC app on every new phone just to monitor it. Or renetwork my house when I buy a new router (imagine having your grandparent struggle through that).
I think what you'll see first is the manufacturers embed the transport layer directly into the hardware and use cheap cellular connections to talk directly to the factory. The entire service will be either opaque to the customer, available for a "subscription", or part of a warranty.
The last-mine problem of standardizing the link from device to internet is going to be a hard one for a long time to come, IMO.
I've bought a lot of "IoT" devices (going back to the days of X10) and yeah lot of them are utterly useless, but so were a lot of dotcom startups in 1999.
The thing I've realized, which may be obvious, is that first and foremost the device needs to work well in its "unconnected" form. Examples:
* Hue/LIFX/etc suck because most of the time I just want to flip a light switch, not pull out my iPhone to control it. I'm a fan of Z-Wave (or whatever) switch replacements that function just fine as regular switches, or I can turn off every light in my house through a wall controller, Alexa, or my iPhone. But if that secondary system fails I can still control my lights.
* Withings scales are great, because it works well as a simple scale, but also records my weight so I can go back and look at trends, etc.
* I like Dropcams because they're super easy, though I wish it weren't locked to a proprietary service
That may work, as long as you can prevent people from using the wall or lamp switch, and if it works even if your internet connection/router is down. My preference is for more hard-wired/bulletproof solutions.
I agree about everything being connected to the internet.
However, I think no truly groundbreaking device has come out, yet. I bet money that one day, a truly useful device will eventually come out. Something, years later, we'll wonder how we ever lived without it.
There are a few devices where there are good arguments for connectivity. An HVAC thermostat, for one, benefits as it's the most expensive consumer of electricity, so a smarter device (or access to external resources) makes sense. Additionally, the ability to control it remotely has a great benefit for "oh hey, I forgot to turn off the A/C before we left on our month-long trip." It could pay for itself right there.
I really think IoT ought to be much more focused on a few high-value devices. Mostly, I think this is important because they have high value for end users but, as such, are also potentially targets for attacks, and by focusing on fewer devices they can hopefully be made much more secure.
All valid points, but I think the products now are laying groundwork for the future. I really believe that at some point, the home will have more safety features and convenience because of IoT. I'm actually surprised that Nest failed to deliver more products. There will be some company that offers a wide range of products and has a huge marketing budget for explaining the use to people.
My favorite example is garage doors. People won't pay $50 to retrofit a half assed automation onto their existing opener. When many openers come with wifi without much of a price premium, lots of people will start using the feature.
Just out of curiosity why would you want to open a garage door via the internet? If it somehow enables you to close the door without seeing what it might be closing on that sounds like a spectacularly bad idea.
Obstruction sensors and automatic reverse have been required for decades (in other words, the openers are already meant to be safe to operate without observation). Anyway, just knowing it is open can be useful.
The most obvious is: "Siri/Alexa open the garage door." (assuming some established authentication mechanism) rather than dealing with the unique garage door remote.
I can imagine others. "This is Alexa, it's 11pm and your garage door is still open. Should I close it?" and things in that vein. But those seem pretty marginal. I have very rarely left my garage door open by accident.
I'm not sure it's a fad, but it is definitely not ready for prime time.
I have a "smart washer" largely because the smarts came with the washer I wanted. The "smart" functionality is completely unusable. First, it can only be connected to the network via WPS. (And the reviewers for the smart phone app don't seem to be able to connect it even with WPS. I haven't tried.) Second, the actual functionality of the connected app is a joke. So why do they have smart functionality at all? My guess is marketing. Is anyone going to return their otherwise awesome washer because the smart functionality is crap? Probably not, since there is no washer or dryer with true smart functionality on the market. But having it on there as a feature might sell a couple units.
On the other hand, while a refrigerator that detects when I'm out of milk is useless now, I can imagine a future in which it then orders the milk for me, so I never have to worry about it. I honestly can't stand doing inventory and ordering/shopping every week. I would love for that process to be automated.
What if you have milk but are out of orange juice or cheese or steak? Unless the 'milkman' is turning up with all these things on the of chance you're low on one this system is much more efficient.
That's a fad. I wouldn't be suprised if people just wanted to take a look inside their fridge while they're shopping, so a webcam at each level would be enough.
> I honestly can't stand doing inventory and ordering/shopping every week. I would love for that process to be automated.
A beautiful fact about refrigerators is that they make for perfect whiteboard surfaces and they are magnetic. I take a picture of my fridge before I go shopping, on which I have written everything that needs to be purchased. A magnetic whiteboard marker with builtin eraser stays on the fridge. There is no way a "smart fridge" is going to catch up to that level of ease + utility any time soon.
I have an actual magnetic whiteboard, which I carry with me to the store and erase as I put things into my cart. The magnetic marker+eraser goes in my pocket, and I put it back on the fridge when I'm putting away my groceries.
What if there were and "Evernote-OCR + Amazon Fresh" app that would let you take a picture of the whiteboard and then have the stuff show up in an hour?
To me, this seems far more inconvenient than just taking out my phone and just putting in to my grocery list what I'm out of in something like Our Groceries. It syncs to my phone and my wife's phone and keeps track of items we've previously purchased. No need for a picture of the fridge, one person can just pick up everything on the list without even asking.
I am perfectly capable of maintaining an inventory. In fact I am quite efficient at it. The reason is that I despise having to do it; which is also the reason I would love an automated process which doesn't require me to snap photos or go to stores. Your solution wouldn't work well for me, though. My refrigerator is covered in stained wood. :)
I will avoid buying 'smart' appliances for probably forever. Besides having mostly useless features, I'd rather not have another electronic probe in my life relaying all of the details of my life to everyone.
The existing products (locks, sensors, smoke detectors) are barely different from existing home security products that have been on the market for years
The existing products have miniscule market penetration, for decades, precisely because their networked integration is abysmal. IoT is yet another attempt to impose some order on the market, which is still focused on incompatibility as a way to retain customers rather than grow the market.
There is ENORMOUS pent up demand for networked everything, if only a cheap universal standard arose and was accepted.
I think that demand is from the companies and not the consumers. Stand in front of a Lowes in the Midwest and ask consumers who walk out if they would like whatever item they bought to be hooked up to the internet. I bet you would get more No's than Yes's.
Some IoT is a fad. But there are very useful things that you can do if you have many connected data devices and sensors at home.
I have ~15 temperature sensors and door sensors that alert me when doors get opened. I was able to use that data to definitively measure the positive impact insulating my house had. I can tell when people enter or exit my house, and I also have ~10 cameras for security. If I get a notification that someone is around my house while I'm gone, I can call my local police department a lot faster than any alarm company.
My Nest also tells me how long my furnace is on every day. Of course their data is useless so I constantly request and store this data in violation of their terms of services, because fuck Nest and their useless TOS. But with that I can make smarter decisions on how to save money in heating and I've halved my bill by manually adjusting my heating (not relying on their stupid algos).
Do I need a fridge that tells me I'm out of eggs? No. But having metrics and data lets me as a homeowner make smarter decisions that save money.
So you got a smart thermostat, wrote some code to hit an API, and do some fancy math to half your heating bill.
Why could you not just turn your thermostat down 4 degrees for the same thing?
I get that there might be some marginal gains you can get with tech ("A heat wave is coming tomorrow, let's shut off the furnace 1 hour early"), but it seems really small and crude compared to existing offline tech can do (look for leaky holes in your insulation, add 6" more blown insulation in your attic, etc"). Or for example replacing a 80% efficient with a 99% efficient furnace has a pretty huge impact. Does nest magic timing have more?
The nest cost $250, insulation cost 10x that, and new furnace is 40x.
I agree that insulation and upgrading the home would produce the nest results but my point is that the data is useful, not useless.
I could shut my heating completely and save 100% but that misses the point. What I did was turn on the heating in the morning to a temperature that was comfortable and then shut it off so that by the time everyone left for work/school the temperature would still be acceptable but on the lower bounds. That saved about 45 mins of heating a day.
Overnight as well I could see how my furnace behaved and tweaked based on that, etc.
The key isn't just the nest but the other temp sensors so I have a complete idea of what my houses characteristics are.
Ya so you don't just need a nest.. you need a nest + other sensors + a bunch of other code. If your bill rate is $200, it is pretty quick before the amount of effort you spent engineering a nest solution would have been beat by just having some guys come blow $1000 in insulation in your roof.
> . I can tell when people enter or exit my house, and I also have ~10 cameras for security. If I get a notification that someone is around my house while I'm gone, I can call my local police department a lot faster than any alarm company.
What kind of cyberpunk dystopia are you living in that you need this?
It's fun as hell knowing what's going on in the house at all times and feeling that I'm protecting myself and my family. Whether or not that's true is to be determined.
What about the increased security liability and added attack surface you've exposed for the benefit of the smarter criminal or random online script kiddie just wanting lulz?
I'd seriously be interested in how you addressed that.
What security liability? If you're talking about a random attack, then they would need to somehow piece together my various email addresses and break different passwords. The cameras (Dropcam, Smartcam, Ring) don't keep ports open on my firewall.
If you're talking about a targeted attack, there's likely nothing I can do except sit in my house with a loaded gun.
I use wirelesstags.net which is good enough for me. Pretty effective for my uses, batteries last about 1 year for updates every 10 mins, and you have access to all your data for all time.
Just imagine you could connect to all your devices (washing machine, heating system, warm water system, garage doors, TV and hifi) from the same app om your phone. And just imagine you could just go to your parents place and do they same thing there. Just imagine you could just buy some random LED bulbs from alibaba and they would just work in your home without thinking about the right standards.
But all of this is just not possible because there is no common standard. I'd really like to buy some remote controllable LED bulbs and perhaps a remote for my heating system. But there are dozens of companies which offer the samw things but they don't work with each other... This is why home automation is not taking off.
My parents are planning to build a new house and it will be 100% traditional without any fancy home automation stuff because there is no universal system.
> Just imagine you could connect to all your devices (washing machine, heating system, warm water system, garage doors, TV and hifi) from the same app om your phone
But if this would work flawlessly then life could be just a bit more comfortable.
For example I am laying currently on the bed and the light in the kitchen is still on. Yes, I could just get out of bed and walk over there, but if there would be an easy gui for this is my mobile it would be good because at the moment I am very lazy.
A quite real problem I have in winter is the temperature in my sleeping room. Sometimes I wake up and it is either too warm or too cold and I have to get up and change the settings. Usually I then have the problem that I can not fall asleep again. If I could just press a button, I could turn around and continue sleeping.
So yes, I don't really _need_ this, but my lazy self would love the comfort.
In all fairness, I'm sure there was a day when many people said: "A remote for the TV--how lazy can you be?" (Admittedly there were only about four channels or so then.)
>If I could just press a button
The more likely model is to go "Alexa--turn off the downstairs lights."
Joking aside, there's probably some confluence between consumer IoT devices and natural language recognition that brings some things into the lazy but genuinely useful range. Unfortunately, most of my annoying home tasks fall into messy physical world interfaces (do the laundry and put it away) that don't really lend themselves to digital automation.
Eh, people also looked at a bunch of terrible inventions and correctly said, "who would ever want that?". Sometimes an invention is mistimed (eg, if the remote control were invented when we had 4 channels and not 40 or 400) or lacking just the right amount of polish (does anyone remember browsing the web on a pre-iPhone smart phone?), but sometimes an invention is just solving a problem almost no one has in a way essentially no one wants.
Fragmentation of various sorts is a problem though some of it is probably inevitable to a certain degree as there's probably a need for, at minimum, a variety of protocols and physical interconnects.
That said, even if you hand wave away this problem you still need to have solutions that mainstream consumers find genuinely useful/valuable. Yes, lack of interoperability adds additional friction. But I see a lot in the consumer IoT space that I'd still have zero interest even if it were completely standardized and interoperable.
I don't have anything useful to say about IoT, but I can think of three major changes newer than television: microwave ovens, personal computers, Internet. Mobile phones are also a major change that you might also put in this category.
Like AI, IoT is a catch-all term for the things that are not yet significant enough to have their own more specific category. Useful applications arise from the primordial soup and acquire their own name.
In 2015 Nest posted job listings for its "Nest Audio Team." The team would be responsible for "developing an audio roadmap for Nest products." Industry observers suspected the audio team was building a smart Bluetooth speaker, but Google beat Nest to the punch with Google Home, an Amazon Echo-style Bluetooth speaker and voice assistant appliance. According to a report from The Information, when Nest found out about Google Home, it asked to work on the project with Google. Nest's request was turned down. We can only guess why—maybe Nest's reputation inside Google had something to do with it?
Often times Google has multiple projects doing effectively the same "thing" going at the same time. They don't work together because the idea is to have the team that can get to the "finish line" first to win, not to co-operate. This, as it was explained to me by an engineering director there, was to encourage a 'natural selection mechanism that selected for the best teams and the best products.'
While I understood the idea, that you would pit your own resources against each other like that was kind of foreign to me. My own personal learning from that, having watched them for 15 years, four of which from the inside, is that as a way of managing a company it produces few viable products and no viable businesses.
The flipside of the argument is a Project Managers hiring 9 women to have a baby in 1 month. Collaboration between teams does incur a cost. Working closely with a single team is easier than coordinating across multiple teams in multiple business units. The additional headcount may not offset the cost.
"Limited divisibility of tasks. Adding more people to a highly divisible task such as reaping a field by hand decreases the overall task duration (up to the point where additional workers get in each others way). Some tasks are less divisible; Brooks points out that while it takes one woman nine months to make one baby, "nine women can't make a baby in one month"."
Average pregnancy may be 9 months long, but that's not even the right metric: the cycle time from one successful birth to the next is longer than that.
Out of every nine women, less than 9 are able to have children once age and various conditions, medical history, choices etc come into play. I think I was over thinking the scenario...
My wife recently made two babies in 33 weeks. That's less than half the average time per baby! Granted, they spent a combined total of 62 days in the NICU, but even accounting for that, we're still ahead, I think.
Congrats! A premature baby in NICU might appear even cuter to many on the internet. The game and app reviewing public is less charitable to premature software, however.
Well full-term is often given as actually being 37-42 weeks, with the due date estimate being 40 weeks (not 9 months), making 42 weeks a normal(ish) pregnancy duration.
You can only double count the babies' NICU time if you're looking at the average baby per week. We averaged 0.048 babies per week, as opposed to the baseline 0.025 babies per week.
That example from the mythical man month, however to really put it in context, it would be like hiring 9 pregnant women and giving a bonus to the first one to deliver a baby and aborting the other 8. And yes, that is an overreach because aborting a later term baby is a way more serious thing than canceling a project you've worked on for nearly a year. However, projects that were "late to deliver" were de-funded and it was really hard on the project leaders. When it happened I did not see it instilling in those engineers a desire to work harder and faster next time, rather it just made them cynical and bitter.
I think importantly, not all approaches requires equal time to achieve. Some approaches might have been more viable in the market if the PMs had simply been given more time to achieve the product rather than just tossing out the first thing to cross the finish line.
It really explains the sloppy state lots of Google products ship in.
The notion of internal competition is really tricky, since you're also competing externally. If you believe in the classic explanations of the theory of the firm, "companies" exist so they can be effectively communist on the inside in an attempt to lower transaction costs. If you have teams competing against each other directly for resources, time-to-market, etc. you don't really have a reason to be a single company anymore.
OTOH, a little bit of internal competition to penalize complacency, if there's not too much wasting of resources, is probably a good thing. It's been fairly common IME for different teams to be working on very related concepts (eg, HPC platforms for different tasks), with a sort of implicit understanding that the team that does well will be able to do the logical extension of the project, and the team that doesn't will work on something else.
although I think fiatmoney is conflating a centrally planned command economy with communism.
Seems like a distinction without a difference, considering that every time a nation tries to implement the latter they end up with the former as an "intermediate stage" of the revolution that somehow never goes away.
In a corporate body, there is no pretense that central planning is a transitional stage. There are a few exceptions like Valve and Zappos, but even fewer genuine success stories where a non-hierarchical structure has outperformed traditional ones in the long run.
Well the fascist countries for instance traditionally had a centrally planned command economy too, and 'fascist' (although that word has perhaps even more baggage than communism) is probably a closer fit to how companies act than 'communist'. Also while "crazy ass dictatorship" lacks any sort of unified political/economic manifesto, they always seem to end up looking like centrally planned economies too, though fewer companies of any real size end up as "crazy ass dictatorships" thanks to the magic of bureaucracy.
Yes, in the US, "fascist" is a general-purpose insult directed at whatever politicians or parties the speaker doesn't like. One of those terms that brings more heat than light.
"Communist" is a loaded word, but it gets at the idea.
One of the thorniest problems in business economics is this: we know free markets produce extraordinary results, but why is it that even in the absence of regulation, so little of the economy is structured as a free market? Very few businesses of any size operate internally as a market, instead they usually operate under hierarchical command and control. Markets generally exist only at the interfaces between organizations.
Why don't teams bid competitively within a company?
How do individuals decide on the teams to form?
Why are managers allowed to rise to the point of failure?
Why do we pay people salaries instead of hourly wages?
Why do we hire hourly wage labor as if though its was permanent?
The number of questions raised by this topic is endless.
The most general explanation is this: yes, companies adopt inefficient structures, but they more than make up for this through significantly reduced transaction costs.
"Why do we hire hourly wage labor as if though its was permanent?"
At the current cost of hourly labor if you made it even more like temporary labor than it already is the quality of worker you could obtain at that price would be terrible.
Some regulation is necessary to even define a free market.
If you believe current US (at least for most of the population) law theory, all employment relationships are completely at-will. They may be terminated at any time for any reason. I would argue that is all that is required for a free market. This would seem to contradict the claim that most companies to not internally operate as a market.
I know there exist serious breaches of this freedom (ie. slavery), but it is not supposed to be the norm anymore.
I have always thought that it was because the level of planning required to run a small to medium sized company is just able to be handled by the human mind. Economy level planning is just too complex for humans and we have to fall back on the next most efficient structure - free markets.
In highly competitive environments, humans tend to adopt a "single executive" type organizational structure. In short, there needs to be a single decision maker who can set strategy, make decisions, approve direction and so on at at least the rate of the competition. Committee systems work fine when things are more steady state and there's less existential threat.
Nations almost all adopt an executive system because nations compete. The U.N. doesn't because Earth doesn't have to compete with anything else. Imagine a competitive interplanetary situation and you can bet the Secretary General of the U.N. would quickly gain significant decision making powers.
Collectivist market philosophies tend to be built on the assumption that there will be no competition internally and thus can be run by various committees. However, even countries trying to follow various types of collectivist ideologies ended up with single executive-types at the top, again because countries compete.
On the small scale, many indigenous tribal systems rely on a council during times of peace, and elect an executive (sometimes called a "war chief") during times of war.
The phrase "business is war" is particularly appropriate since most businesses operate like they're at war.
> On the small scale, many indigenous tribal systems rely on a council during times of peace, and elect an executive (sometimes called a "war chief") during times of war.
That's how the Roman Republic operated, too: the Senate would elect a dictator for six months to handle certainly irregular situations. Unlike any other Roman official, he had no colleague who could countermand his orders, and he was not legally liable for any act he committed in office.
"Why do we hire hourly wage labor as if though its was permanent?"
Could you give an example of this? Are you talking about using contractors to get through a project, and then keeping them on indefinitely after the project is over?
Read Ronald Coase's Nature of the Firm: http://www3.nccu.edu.tw/~jsfeng/CPEC11.pdf It isn't really that firms are communist internally, but that firms exist to reduce transactions costs. Such as contracting for every thing every day (i.e. I don't need a new contract every single day at my job).
I don't think it's objectivist at all -- it's somewhere between "let's not put all our eggs in one basket" and "it's unlikely we got it 100% the first time so don't be afraid to start from scratch."
> If you believe in the classic explanations of the theory of the firm, "companies" exist so they can be effectively communist on the inside in an attempt to lower transaction costs. If you have teams competing against each other directly for resources, time-to-market, etc. you don't really have a reason to be a single company anymore.
Sure you do. It allows someone at the top to flip the switch between market and central command. Ideal firm size and vertical consolidation varies between industries, and plausibly varies in different environments, different times, etc.
Ha! The team that "loses" gets to go try something new, and the team that "wins" gets to take their new product into maintenance/bug fix/enhancement mode?
Did they actually refer explicitly to "natural selection"? Because while natural selection has indeed produced some impressive results, it did so over billions of years, and not terribly efficiently, so I wouldn't generally consider it a model to be emulated, unless Google is really optimistic about how long the company is going to be around.
Yes, they used that exact term, invoking Darwin at the same time. And I agree with you about the time frame.
I've been reading "The Vital Question"[1] of late and it has been the first book in a while where I've had several "oh that makes so much sense!" moments. And one of the tenets is that Darwin was correct in the small, and wrong in the large understandings of evolution.
Something that jumped out to me while reading it, is that information extraction drives our technology systems like energy drives biological systems. Read the book and then sit back and analyze Twitter as a multi-cellular organism with information as energy and individuals as cells. A sort of Datasaurus. Fun stuff.
[1] "The Vital Question: Energy, Evolution, and the Origins of Complex Life" -- http://www.amazon.com/Vital-Question-Evolution-Origins-Compl... based on the recommendation tweeted by Gates and comments here, I read it through cover to cover while camping, and now going through a second reading to pick up what I overlooked the first time.
The German competing fiefdoms method caused real problems when the allies got a good enough design and cranked it out like the T34, sten gun and arguably the Sherman tank
Emphasis on was. It seems like these days Microsoft is moving towards more internal collaboration, with their "One Microsoft" initiative. They've found that greater focus on common, standard tools used across the entire company is better than having a bunch of teams, each rolling their own version of a given tool.
More to the point, isn't this exactly what Steve Jobs advised Larry Page not to do? Steve Jobs said, famously, to put "more wood behind fewer arrows". And for a while, it seems like the advice was sticking - Google shuttered a bunch of products (for better or for worse) and seemed to be buckling down to focus on the search and YouTube division. But now it seems like Google is going right back to how it used to be - many teams, working on many products, with no guarantee of ongoing development or support. It seems to me that Google is turning into the next Cisco.
Toyota has formalized internal competitions. The Prius drivetrain was the best out of a field of 10. Designs for new iPhones apparently work in a similar fashion. I think such formal structures produce very good results and are necessary to prevent destructive internal competition.
For me, there is an interesting difference. On the one hand you can say "everyone compete on the next iPhone antenna design". That is one component out of an over all strategy. But proposing incremental improvements at Google when I was there was death, you were accused of "not thinking big enough" and everything had to be a "moonshot".
So you couldn't propose just a new antenna, you had to propose using an entirely different spectrum with entirely different rules and an entirely different platform architecture where voice service was one of many options rather than the main option. You don't get to propose re-inventing the phone, you have to re-invent telecommunications. One of those that got out of the door when I was there was the Google "wave" project. Its vision was so big it missed the fact that people just wanted Slack[1].
Bottom line is that competition within a structure is good, its something that lots of organizations have applied successfully. Google's style, when I was there, was never structured enough to head toward anything. It always went all directions at once, leaving the organization with a net zero momentum in any direction.
[1] Yes they are nearly a decade different in time frame, but as a 'dogfood' wave user and then later outside of Google it was pretty clear the "win" was the IRC replacement, none of the vision stuff was getting much traction but it was the way to go if you're group wanted to figure out which cafe to go to lunch at that day.
But proposing incremental improvements at Google when I was there was death, you were accused of "not thinking big enough" and everything had to be a "moonshot".
Sounds like this was used as a bludgeon for attacking competitors.
Bottom line is that competition within a structure is good, its something that lots of organizations have applied successfully. Google's style, when I was there, was never structured enough to head toward anything. It always went all directions at once, leaving the organization with a net zero momentum in any direction.
My experience with such competition, is that groups actually tend to undermine each other instead of pulling together with cohesive effort. Here, the lowered cost of transactions within a company can allow other groups within a company to act as gatekeepers -- otherwise "enemy" in-company groups can be penalized with higher (outside company) transaction costs for getting things done. This is especially true when project budgets are made with in-company costs in mind.
Do you get the impression they re-designed the organization, with moonshots at their moonshot factories, while incremental improvements at the rest of the org?
Given the discussions I've had with friends still there and friends who have left, I would say that yes, Google is constantly redesigning the organization. Oddly enough the experience was very similar to Sun Microsystems as it grew, it redesigned its own internal organization several times. It was a running joke at Sun that when a big re-org hit it was "some new chapter in the MBA handbook."
Google is famously data driven (even when it hurts) and has taken a lot of criticism (justified in my opinion) about not being able to create any new business outside of search and search advertising. I saw, and commented on, the Alphabet re-org as a way to get better visibility into the data coming from the organization. Spend vs revenue. Robots aren't carrying their weight? Toss them over. Making cell phones too hard? Sell it off.
If you work in companies like this long enough you begin to recognize that the organization can constrain what products you can even attempt to build. When the Sun "Voyager" (aka the 'portable') came out we joked that Sun could build any computer it wanted to, as long as you didn't mind it looking like a workstation. A partner Tadpole had built an actual laptop. But the Sun hardware engineering organization couldn't get there from a clean sheet of paper.
For that reason alone it is critical that Google keep re-organizing until they find a formula that works for them.
My own anecdotal experience I heard a much larger competitor in my line of work have 3 teams competing against each other for a product. All failed except one which tagged up with an external partner to build a solution for them. Apparently many large companies don't know what to do with all their employees. I spoke to some of them later after my company hired some and all some engineers hope for is to actually complete a project.
Something similar happened to one of my previous employers. We were less than 100 people and cranking out many successful projects. Once we got acquired and nearly quadrupled in size, none of the projects succeeded. Many of the early contributors left one by one due to frustrations and pressure including me. The new management style had changed everything.
This is something I'm coming to grips with. Massive inefficiency is not just the domain of government agencies. There are plenty of firms in the private sector that have teams sitting idle, and the kicker is they are hiring!
It is not uncommon for a software developer to be asked to keep a seat warm in an open office seating environment. It feels very unnatural to be expected to sit in front of a computer for 8 hours and be given nothing to do. Some people convince themselves that they are riding the gravy train, others go crazy and leave to find actual work.
While I understood the idea, that you would pit your own resources against each other like that was kind of foreign to me. My own personal learning from that, having watched them for 15 years, four of which from the inside, is that as a way of managing a company it produces few viable products and no viable businesses.
I can't find any good citations online, but Electronic Arts was reputed to do that as well. Supposedly they'd put several developers on the same title, and whoever finished first got to ship.
Well if the reports are true, that is also what happened with Longhorn and later on with WinRT vs Desktop, meaning WinDev vs DevTools divisions at Microsoft.
I thought that really came through in I/O this year. The Allo team does not seem to be the same team as the one making Hangouts. Which is fine... except that has a Hangouts user I really wish they would dedicate themselves to a messaging app and actually make it good.
The more confusing one was web vs. native - there were a dozen sessions on progressive webapps, yet they barely got a mention in the keynote, while Instant Apps got a full showcase. Internal competition is all very well, but it's frustrating as a developer to not be able to get a clear answer as to which platform we ought to be developing for.
My understanding is that MSFT operates in similar fashion. The difference is that none of the implementations get green lighted until some C level exec asks years later, "Hey why don't we do X?" Some other guy in the room will reply, "Sure, we have 4 versions of X. Which one do you want."
That's what I thought while reading the post. Microsoft was infamous for internal divisions basically acting as separate companies, with divergent roadmaps, duplicated features and so on.
In this case, however, given the prominence that was originally placed on the acquisition (at a time when Google had no direct presence in home automation), I suspect that the original vision was to actually merge the two worlds. The Google rank&file probably learnt very quickly how toxic Fadell was, and then proceeded to keep Nest at arms' length and do their own thing, basically writing off the acquisition and hoping it will eventually go away.
Personally I think it's great. Coordinating 57k people is very very hard. If you had to ask 57k people "Hey, I'm going to work on this, anyone else already doing something similar" the entire company would grind to a halt.
Sure there are mistakes but I don't think it's top down "team x" compete with "team y". It's more like both teams don't know about each other until the projects are far along. At which point might as well see which one does better... Or at least that seems to be the attitude at google.
While that can be a problem, it's tough to address at Google's scale. Consider the possible solutions:
(1) Top management tells everybody what to do. Result: No initiative from low-level employees.
(2) In case of conflicting projects, top management decides which team should continue. Result: Rampant internal politics.
(3) In case of conflicting projects, let all of them continue, and see which project wins the market. Result: Wasting resources; company looks disorganized to outsiders.
Some other companies (Apple?) avoid the problem by sticking to (1). However, many Google employees joined in the first place thanks to the tale of "A company where you can work on the problem you want to": it will be very tough for them to ditch the model wholesale, and it may harm the company in the long run.
I think Google does a bit of all three. While Google could certainly do better in some cases, it's not like Google's willfully ignoring an obvious solution.
The problem with 3) is not just wasted resources or company looking bad to outsiders. It also leads to rampant internal politics as people simply do everything in their power to win, or worse, to survive. Just to name a few common tricks: developing features secretly, not sharing resources or technologies, poaching coworkers from other teams, lobbying to get more resources, developing moating features instead of focusing on what users really need, and trading long-term objectives for short-term gains just to get to the so-called "finish line".
This type of practice is common in government work. Hire a dozen teams to make a thing by some deadline, and award the best result with a contract.
It seems wasteful, but as we can see here, clearly the Nest work style couldn't compete with Google, and so no need to waste further R&D. But until you didn't the contest, how could you know?
I think the fundamental problem with this type of approach is that the ugly, sausage making, competition ends up getting aired too much with the public which creates market confusion and drives potential customers to places with a better, simpler story.
My understand is that great product companies, like Apple, also have a kind of Darwinian development practice, but only ever reveal the "winner" once it's clear that it's the winner.
As an outsider, Apple looks like it has strategy and vision, while Google looks completely baffling, even releasing and pulling products from the market within weeks or releasing clearly incomplete, barely beta level products with no real vision and just expecting the public to sort it out.
> While I understood the idea, that you would pit your own resources against each other like that was kind of foreign to me. My own personal learning from that, having watched them for 15 years, four of which from the inside, is that as a way of managing a company it produces few viable products and no viable businesses.
Interesting you say this as the first example that I can think of is Apple themselves when Steve Jobs was forced out of the Apple Lisa team and he joined the Macintosh team which then became in direct competition with the Apple Lisa team. I don't need to expunge on that but i found it to be a very famous case study on how 2 internal teams in direct competition affected the company.
I'm not sure I believe that's an intentional strategy or just a side-effect of size and the philosophy of letting teams start projects on their own.
So you have two teams somewhere in a company start a project at the same time.
Your challenge is, what does the corporate structure need to look like such that these two teams learn about each other's project early enough to collaborate?
If you rely on project information being passed up and down the corporate ladder, until one person learns that two subtrees of his reports are working on the same project, that's never going to work because bandwidth is reduced as you go up the tree, so less and less information is passed up, and early/experimental projects will frequently not make the cut and get filtered out. So your LCB will never hear about both projects.
If you just rely on direct communication between teams, you've got an N-squared problem as you grow, so that will never work.
If you try to give each team a carved out domain, new projects will frequently exist in the "chocolate & peanut butter" intersections, and both the "chocolate" and "peanut butter" teams will start the same project.
It's a difficult problem, and I'm not sure there's an answer between "companies should be small".
Right before they went under Enron was doing the same thing: they felt having three teams compete for fastest to market was a competitive advantage and worth the wasted resources sunk into the two losing teams.
"The present system includes competition between people, teams, departments, divisions, students, schools and universities. Although economists have taught that competition will solve our problems, we now know that competition is destructive."
As I recall Apple did a similar thing with the original iPhone OS, one team took the iPod software as a starting point and another took OS X as a starting point. I think I have that right.
If in the end it makes the best product it's probably worth the waste of having competing solutions. But that is kind of a big if. (I would say the iPhone passed that test, personally.)
The debacle of the smoke alarm made me immediately regret buying one of my friends for their newborn son's room. But wow, the following sounds like an utter fail...if quadrupling the headcount doesn't bring new products, then you at least expect the old product to have tried overexpanding into new markets...they weren't even able to do that:
Even by Fadell's own standards, the Google acquisition doesn't seem to have led to huge wins for Nest. In an interview with Bloomberg shortly after the sale, Fadell was asked what the move would mean for Nest. "For us it was about getting out to the world faster" Fadell said. "[Our products are imported] in 96 countries where we don't even sell today. We can see these devices being connected but we can't get there fast enough. For us, going with Google, we can get access to resources to allow us to move these products around the world much more quickly." Before the Google acquisition, Nest was for sale in three countries. In two-and-a-half years under Google and Alphabet, Nest expanded to just four more countries.
I love my Nest thermostat- it's still better than any other smart thermostat out there in build quality and reliability (tho the Ecobee series is slightly more feature rich) - but I can't say I'm surprised.
Nest is, in many ways, the most disappointing company. I'm not sure how much of that is just down to them lacking vision or leadership though - they have the skills to do great things.
I have a Nest thermostat as well, and I like it. I've wondered before if they just had that one good idea and that was it. When they announced the smoke detector my jaw dropped. It just seemed like such an awful idea.
I had thought maybe they would do a smart lawn watering system next. I know it's a smaller market, but it seemed better aligned with the thermostat as something that could be marketed as a money-saver. The smoke detector doesn't save you any money, it just seems like a solution in search of a problem. The lawn system could basically sell itself: local rebates on the up-front cost, the promise of saving you water/$$, and the name that Nest had already established from their thermostat.
There are some other companies making products like this now, but I don't know if they're good.
Drives me slightly crazy seeing lawn sprinklers going while it's raining and humid..
The underground sprinkler system my parents installed 30 years ago has a sensor to prevent that, it's certainly not unusual. It would be interesting to see rain sensors for regular sprinters although I'd be shocked if they didn't already exist.
The build quality is definitely good. However I am very unpleased with the temperature sensor, it seems to only be accurate to more than 1 degree, with the software further adding a tolerance in when to turn on/off the AC/Heating I find myself constantly manually setting the temperature as I'm either freezing or getting too hot. A solution would be remote temperature sensors which people have been asking for since the beginning, but that never happened. I believe there are now thirdparty solutions that hook into Nest perhaps via IFTTT, but that turns the Nest in just a thermostate with remote on/off switch, which is basically the only feature that mostly works for me. (yeah even that isn't always robust)
At this point, I doubt it will matter much if they do. Many other companies are working on it now. Eventually, one will get it right, and the auto companies will copy/follow.
Or it will be the auto companies that turn out to get there in the best incremental practical way.
I know this is probably a minority opinion here, but one plausible scenario is that 5 or 10 years from now, Google will come to the realization that full no steering wheel autonomy is still "just over the horizon" and, in the face of declining ad revenues (or whatever), the whole effort gets ramped down.
No, Google's core business, search, is doing great. Starting a conglomerate of unrelated businesses is not working out.
Google's basic problem: search ads are incredibly profitable when you're #1. But they've saturated that market. Nothing else they do with their excess cash can match those returns. If they do something less profitable, the stock price goes down.
Hence Alphabet's search for a "moonshot". But what management meant by "moonshot" was another high volume product with a huge return on investment. That's probably not going to happen in the home appliance space, which is price-sensitive and competitive.
Not sure how much has changed in a year, but Youtube runs at breakeven/potentially a loss[0]:
"The online-video unit posted revenue of about $4 billion in 2014, up from $3 billion a year earlier, according to two people familiar with its financials, as advertiser-friendly moves enticed some big brands to spend more. But while YouTube accounted for about 6% of Google’s overall sales last year, it didn’t contribute to earnings. After paying for content, and the equipment to deliver speedy videos, YouTube’s bottom line is “roughly break-even,” according to a person with knowledge of the figure."
In his recent Recode interview, Elon Musk said he doesn't consider Google to be a competitor to Tesla. They're not selling cars today. They're not even on the path to making and selling cars. Their most likely option will be to license their tech to car companies. They're working with Fiat now.
Just because they did Ads Search, GMail and Android right, doesn't mean they know how run or integrate a company into their ecosystem. It take a different culture / management / skills to do that.
In fact one can argue there is an inverse correlation there. The more efficient they are at developing new products with moonshot ideas, the less efficient they'd be at managing integrating a newly acquired company.
Doesn't Nest have like 1200 employees? What are they doing day-to-day? With that many people it seems like you'd almost release a product accidentally.
Ah, yes! And because of the current team and structure of management, everyone involved, consciously or unconsciusly knows that it will be nothing but a flop.
They make a move, they lose. They don't do anything they lose as well, it will just take longer. So they chose the second option.
That's a nice 3B$ flop right there, 3.5B$ if you take the dropcam into account.
Why couldn't they just focus on small inexpensive products that sell without thinking? Take chrome cast for example, great product, cheap price, don't really have to debate if I want to buy it.
Hey Google, here is an idea, why don't you make an IoT button that can talk/control other devices like bttn or flic but costs just 5$ (I'm looking at you Amazon dash).
They will sell like hotcakes!
After that you can introduce some other products, maybe under the home brand, like a device that control your AC (it just needs to work, doesn't have to be apple like the Nest) or one that monitors your water heater.
What about a small device that checks for leaks in your pipes.
There is so much stuff todo here.
I don't think they get much more in ad revenue because of chrome cast, if I want to watch something I'd watch it on my mobile or PC, casting to the TV is just more comfortable, it's not an enabler to something that wasn't watchable.
Moreover chrome cast is used outside of google products like Plex, Netflix, HBO etc.
Having a low bar for entry into the home is huge for a company which is all about data and insights.
Well, it's not like there's been any positive Nest press in the last year or so to obsess over.
Based on the stories, what's been happening at Nest is an extreme case of many of the same organizational antipatterns that people in the tech industry see in their own companies on a smaller scale. The things that have happened at Nest (if the stories are to be believed) are like cancer to a tech company. If left unchecked those antipatterns can grow and consume a company, and can ultimately kill it. No amount of reform will save a company if they can no longer retain talent or recruit replacements of the same or better caliber.
Given that so many HN readers are in the tech industry themselves, they probably worry about these kinds of cancers taking root at their own companies (I know I do). This stuff exists everywhere, it's just that for stable companies it's the exception and not the rule. I'm not surprised that it's been a hot topic.
> Well, it's not like there's been any positive Nest press in the last year or so to obsess over.
Their performance versus the rest of the Google ventures? The product line that was previously leaked? And, in the eyes of Hacker News perhaps celebrating Tony stepping aside.
A lot of these stories, if they are in fact true, likely happened prior to Google acquisition and those people are now just feeling comfortable to speak out without the fear of it being spun as retaliation.
I have a lot of faith that the Google HR department wouldn't tolerate any of the behavior, which is why it smells like a marketing campaign.
all of this comment and not a single one seems constructive...
Nest the company problem as consumers (myself being one) seems to be Software...
I use the nest software and simply put it's not up to scratch
The Hardware is beautiful if only the software was...
For example the nest protect has temperature sensors yet the Nest thermostat can not use these...
Nor can the software operate without access to the "cloud" even in a reduced ability.
Google OnHub suffers from the same Software problems, Great hardware and nice low level drivers however it cant even use IPv6 which for a router is pretty basic
my advice for the incoming CEO, interoperability
Simple
make use of thread & brillo (software) to unify things...
(produce an alarm but have the option of ADT as the responder)
Unify the products so they can work within a LAN and expand outside of the USA.
181 comments
[ 2.8 ms ] story [ 140 ms ] threadI have an original Nest thermostat but would never buy another Nest product due to the horror stories that continually pop up online. I don't follow Nest closely or have any particular interest in them, but I remember the video they reference in the article from the Google employee who couldn't stop the smoke detectors from false alarming, and I also saw the news when they decided to "brick" a $300 product they acquired.
Based on those two things alone, I don't feel like they take customer satisfaction seriously, nor do I have confidence in their ability to build a good product--even though my thermostat has never had any issues. I doubt I'm alone in this.
The current generation of Nest Protect actually seems pretty good--it's not nearly as bad with full volume false alarms as my previous one was when I open the oven door. But I basically don't care about the IoT aspect of it.
ADDED: In general, I'm sure a Nest works fine. I just have a use case where the winter reliability of my heating system is really important. I can't control every aspect of that but I don't want to hurt any aspect of it for trivial gains.
What do they have that is hard to do from scratch?
People are very hesitant to buy new appliances/home products. The last major change to the American home was the television, and that was ~50 years ago. Desktops had their day, but have been largely replaced by Laptops/Smartphones for many consumers. Adoption of new home appliances and technologies is glacially slow.
Also newer electronics have shown they need replaced/upgraded every 2-5 years. You are going to be hard pressed to sell the idea that the consumer needs to replace all their light switches and smoke detectors even every 10 years. I barely replace light bulbs at that rate now.
Cars about the only place where the IoT call home paradigm works. They are complicated machines that need semi regular maintenance. Placing sensors all over it makes sense. Trading updates for telemetry is attractive because a cars computer needs updates like any other computer. But I don't need a fridge that senses when I am low on milk, when I can just as easily look inside it and tell for myself.
EDIT: I think I should clarify that home IoT is a fad. Having machines feed back telemetry is useful, especially commercial machines with a service contract. The data is useful when its actionable. Data for the sake of data in the home is a waste.
https://en.m.wikipedia.org/wiki/Economy_7
What's not a fad is getting devices to communicate their health and status to an upstream repository for things like predictive analytics.
This is already happening in the B2B sector. It will reach the consumer sector next, but like you said it will take 2-3 consumer product cycles to make it real.
It won't be about remote control or any of the barely useful use-cases you're seeing with fridges and door locks and water bottles. It will be knowing your home's A/C has two weeks of operation before it seizes up. It will be knowing your home's plumbing has frozen and burst while you were away.
The problem with IoT is the I
Connected devices are not IoT. Marketing needs to learn that.
The major issue is the platform. Commercial products already have dedicated terminals, so the platform already exists. But I don't want to have to install the AC app on every new phone just to monitor it. Or renetwork my house when I buy a new router (imagine having your grandparent struggle through that).
The last-mine problem of standardizing the link from device to internet is going to be a hard one for a long time to come, IMO.
The thing I've realized, which may be obvious, is that first and foremost the device needs to work well in its "unconnected" form. Examples:
* Hue/LIFX/etc suck because most of the time I just want to flip a light switch, not pull out my iPhone to control it. I'm a fan of Z-Wave (or whatever) switch replacements that function just fine as regular switches, or I can turn off every light in my house through a wall controller, Alexa, or my iPhone. But if that secondary system fails I can still control my lights.
* Withings scales are great, because it works well as a simple scale, but also records my weight so I can go back and look at trends, etc.
* I like Dropcams because they're super easy, though I wish it weren't locked to a proprietary service
Pulling a new wire is better but also much more expensive. Adding a new switch for me is~200 and I can't move it around like the Hue.
However, I think no truly groundbreaking device has come out, yet. I bet money that one day, a truly useful device will eventually come out. Something, years later, we'll wonder how we ever lived without it.
Like the iPhone did for smartphones.
I really think IoT ought to be much more focused on a few high-value devices. Mostly, I think this is important because they have high value for end users but, as such, are also potentially targets for attacks, and by focusing on fewer devices they can hopefully be made much more secure.
http://www.dezeen.com/2016/04/27/smart-technology-driverless...
My favorite example is garage doors. People won't pay $50 to retrofit a half assed automation onto their existing opener. When many openers come with wifi without much of a price premium, lots of people will start using the feature.
The most obvious is: "Siri/Alexa open the garage door." (assuming some established authentication mechanism) rather than dealing with the unique garage door remote.
I can imagine others. "This is Alexa, it's 11pm and your garage door is still open. Should I close it?" and things in that vein. But those seem pretty marginal. I have very rarely left my garage door open by accident.
I have a "smart washer" largely because the smarts came with the washer I wanted. The "smart" functionality is completely unusable. First, it can only be connected to the network via WPS. (And the reviewers for the smart phone app don't seem to be able to connect it even with WPS. I haven't tried.) Second, the actual functionality of the connected app is a joke. So why do they have smart functionality at all? My guess is marketing. Is anyone going to return their otherwise awesome washer because the smart functionality is crap? Probably not, since there is no washer or dryer with true smart functionality on the market. But having it on there as a feature might sell a couple units.
On the other hand, while a refrigerator that detects when I'm out of milk is useless now, I can imagine a future in which it then orders the milk for me, so I never have to worry about it. I honestly can't stand doing inventory and ordering/shopping every week. I would love for that process to be automated.
Regular milk deliveries might be a little bit more inconvenient than automatically ordered milk, but it's vastly simpler.
What if you have milk but are out of orange juice or cheese or steak? Unless the 'milkman' is turning up with all these things on the of chance you're low on one this system is much more efficient.
That's a fad. I wouldn't be suprised if people just wanted to take a look inside their fridge while they're shopping, so a webcam at each level would be enough.
A beautiful fact about refrigerators is that they make for perfect whiteboard surfaces and they are magnetic. I take a picture of my fridge before I go shopping, on which I have written everything that needs to be purchased. A magnetic whiteboard marker with builtin eraser stays on the fridge. There is no way a "smart fridge" is going to catch up to that level of ease + utility any time soon.
Digital devices take off when they have a killer application. So far I've seen nothing along these lines.
The existing products have miniscule market penetration, for decades, precisely because their networked integration is abysmal. IoT is yet another attempt to impose some order on the market, which is still focused on incompatibility as a way to retain customers rather than grow the market.
There is ENORMOUS pent up demand for networked everything, if only a cheap universal standard arose and was accepted.
I have ~15 temperature sensors and door sensors that alert me when doors get opened. I was able to use that data to definitively measure the positive impact insulating my house had. I can tell when people enter or exit my house, and I also have ~10 cameras for security. If I get a notification that someone is around my house while I'm gone, I can call my local police department a lot faster than any alarm company.
My Nest also tells me how long my furnace is on every day. Of course their data is useless so I constantly request and store this data in violation of their terms of services, because fuck Nest and their useless TOS. But with that I can make smarter decisions on how to save money in heating and I've halved my bill by manually adjusting my heating (not relying on their stupid algos).
Do I need a fridge that tells me I'm out of eggs? No. But having metrics and data lets me as a homeowner make smarter decisions that save money.
Why could you not just turn your thermostat down 4 degrees for the same thing?
I get that there might be some marginal gains you can get with tech ("A heat wave is coming tomorrow, let's shut off the furnace 1 hour early"), but it seems really small and crude compared to existing offline tech can do (look for leaky holes in your insulation, add 6" more blown insulation in your attic, etc"). Or for example replacing a 80% efficient with a 99% efficient furnace has a pretty huge impact. Does nest magic timing have more?
I agree that insulation and upgrading the home would produce the nest results but my point is that the data is useful, not useless.
I could shut my heating completely and save 100% but that misses the point. What I did was turn on the heating in the morning to a temperature that was comfortable and then shut it off so that by the time everyone left for work/school the temperature would still be acceptable but on the lower bounds. That saved about 45 mins of heating a day.
Overnight as well I could see how my furnace behaved and tweaked based on that, etc.
The key isn't just the nest but the other temp sensors so I have a complete idea of what my houses characteristics are.
What kind of cyberpunk dystopia are you living in that you need this?
I'd seriously be interested in how you addressed that.
If you're talking about a targeted attack, there's likely nothing I can do except sit in my house with a loaded gun.
It is a fad because there is no common standard.
Just imagine you could connect to all your devices (washing machine, heating system, warm water system, garage doors, TV and hifi) from the same app om your phone. And just imagine you could just go to your parents place and do they same thing there. Just imagine you could just buy some random LED bulbs from alibaba and they would just work in your home without thinking about the right standards.
But all of this is just not possible because there is no common standard. I'd really like to buy some remote controllable LED bulbs and perhaps a remote for my heating system. But there are dozens of companies which offer the samw things but they don't work with each other... This is why home automation is not taking off.
My parents are planning to build a new house and it will be 100% traditional without any fancy home automation stuff because there is no universal system.
Why would I want to.
But if this would work flawlessly then life could be just a bit more comfortable.
For example I am laying currently on the bed and the light in the kitchen is still on. Yes, I could just get out of bed and walk over there, but if there would be an easy gui for this is my mobile it would be good because at the moment I am very lazy.
A quite real problem I have in winter is the temperature in my sleeping room. Sometimes I wake up and it is either too warm or too cold and I have to get up and change the settings. Usually I then have the problem that I can not fall asleep again. If I could just press a button, I could turn around and continue sleeping.
So yes, I don't really _need_ this, but my lazy self would love the comfort.
>If I could just press a button
The more likely model is to go "Alexa--turn off the downstairs lights."
Joking aside, there's probably some confluence between consumer IoT devices and natural language recognition that brings some things into the lazy but genuinely useful range. Unfortunately, most of my annoying home tasks fall into messy physical world interfaces (do the laundry and put it away) that don't really lend themselves to digital automation.
That said, even if you hand wave away this problem you still need to have solutions that mainstream consumers find genuinely useful/valuable. Yes, lack of interoperability adds additional friction. But I see a lot in the consumer IoT space that I'd still have zero interest even if it were completely standardized and interoperable.
In 2015 Nest posted job listings for its "Nest Audio Team." The team would be responsible for "developing an audio roadmap for Nest products." Industry observers suspected the audio team was building a smart Bluetooth speaker, but Google beat Nest to the punch with Google Home, an Amazon Echo-style Bluetooth speaker and voice assistant appliance. According to a report from The Information, when Nest found out about Google Home, it asked to work on the project with Google. Nest's request was turned down. We can only guess why—maybe Nest's reputation inside Google had something to do with it?
Often times Google has multiple projects doing effectively the same "thing" going at the same time. They don't work together because the idea is to have the team that can get to the "finish line" first to win, not to co-operate. This, as it was explained to me by an engineering director there, was to encourage a 'natural selection mechanism that selected for the best teams and the best products.'
While I understood the idea, that you would pit your own resources against each other like that was kind of foreign to me. My own personal learning from that, having watched them for 15 years, four of which from the inside, is that as a way of managing a company it produces few viable products and no viable businesses.
"Limited divisibility of tasks. Adding more people to a highly divisible task such as reaping a field by hand decreases the overall task duration (up to the point where additional workers get in each others way). Some tasks are less divisible; Brooks points out that while it takes one woman nine months to make one baby, "nine women can't make a baby in one month"."
https://en.wikipedia.org/wiki/Brooks%E2%80%99_law
42 weeks overall
4.33 weeks in a month
42 weeks -> just shy of 10 months
I'm not so sure if you are ahead on weeks. But you are ahead on kids! Congratulations and good luck
Not that I want to derail the well-wishing!
It really explains the sloppy state lots of Google products ship in.
OTOH, a little bit of internal competition to penalize complacency, if there's not too much wasting of resources, is probably a good thing. It's been fairly common IME for different teams to be working on very related concepts (eg, HPC platforms for different tasks), with a sort of implicit understanding that the team that does well will be able to do the logical extension of the project, and the team that doesn't will work on something else.
I don't buy the "first to win argument", but it would seem useful to take multiple approaches to solving the same or similar problems.
https://en.wikipedia.org/wiki/Theory_of_the_firm
although I think fiatmoney is conflating a centrally planned command economy with communism.
Seems like a distinction without a difference, considering that every time a nation tries to implement the latter they end up with the former as an "intermediate stage" of the revolution that somehow never goes away.
In a corporate body, there is no pretense that central planning is a transitional stage. There are a few exceptions like Valve and Zappos, but even fewer genuine success stories where a non-hierarchical structure has outperformed traditional ones in the long run.
http://www.inc.com/audacious-companies/leigh-buchanan/mornin...
One of the thorniest problems in business economics is this: we know free markets produce extraordinary results, but why is it that even in the absence of regulation, so little of the economy is structured as a free market? Very few businesses of any size operate internally as a market, instead they usually operate under hierarchical command and control. Markets generally exist only at the interfaces between organizations.
Why don't teams bid competitively within a company? How do individuals decide on the teams to form? Why are managers allowed to rise to the point of failure? Why do we pay people salaries instead of hourly wages? Why do we hire hourly wage labor as if though its was permanent?
The number of questions raised by this topic is endless.
The most general explanation is this: yes, companies adopt inefficient structures, but they more than make up for this through significantly reduced transaction costs.
At the current cost of hourly labor if you made it even more like temporary labor than it already is the quality of worker you could obtain at that price would be terrible.
If you believe current US (at least for most of the population) law theory, all employment relationships are completely at-will. They may be terminated at any time for any reason. I would argue that is all that is required for a free market. This would seem to contradict the claim that most companies to not internally operate as a market.
I know there exist serious breaches of this freedom (ie. slavery), but it is not supposed to be the norm anymore.
Nations almost all adopt an executive system because nations compete. The U.N. doesn't because Earth doesn't have to compete with anything else. Imagine a competitive interplanetary situation and you can bet the Secretary General of the U.N. would quickly gain significant decision making powers.
Collectivist market philosophies tend to be built on the assumption that there will be no competition internally and thus can be run by various committees. However, even countries trying to follow various types of collectivist ideologies ended up with single executive-types at the top, again because countries compete.
On the small scale, many indigenous tribal systems rely on a council during times of peace, and elect an executive (sometimes called a "war chief") during times of war.
The phrase "business is war" is particularly appropriate since most businesses operate like they're at war.
That's how the Roman Republic operated, too: the Senate would elect a dictator for six months to handle certainly irregular situations. Unlike any other Roman official, he had no colleague who could countermand his orders, and he was not legally liable for any act he committed in office.
Could you give an example of this? Are you talking about using contractors to get through a project, and then keeping them on indefinitely after the project is over?
Sure you do. It allows someone at the top to flip the switch between market and central command. Ideal firm size and vertical consolidation varies between industries, and plausibly varies in different environments, different times, etc.
Looks like the joke is on the winning team.
I've been reading "The Vital Question"[1] of late and it has been the first book in a while where I've had several "oh that makes so much sense!" moments. And one of the tenets is that Darwin was correct in the small, and wrong in the large understandings of evolution.
Something that jumped out to me while reading it, is that information extraction drives our technology systems like energy drives biological systems. Read the book and then sit back and analyze Twitter as a multi-cellular organism with information as energy and individuals as cells. A sort of Datasaurus. Fun stuff.
[1] "The Vital Question: Energy, Evolution, and the Origins of Complex Life" -- http://www.amazon.com/Vital-Question-Evolution-Origins-Compl... based on the recommendation tweeted by Gates and comments here, I read it through cover to cover while camping, and now going through a second reading to pick up what I overlooked the first time.
https://en.wikipedia.org/wiki/V_bomber
http://mobile.nytimes.com/2010/02/04/opinion/04brass.html?pa...
More to the point, isn't this exactly what Steve Jobs advised Larry Page not to do? Steve Jobs said, famously, to put "more wood behind fewer arrows". And for a while, it seems like the advice was sticking - Google shuttered a bunch of products (for better or for worse) and seemed to be buckling down to focus on the search and YouTube division. But now it seems like Google is going right back to how it used to be - many teams, working on many products, with no guarantee of ongoing development or support. It seems to me that Google is turning into the next Cisco.
So you couldn't propose just a new antenna, you had to propose using an entirely different spectrum with entirely different rules and an entirely different platform architecture where voice service was one of many options rather than the main option. You don't get to propose re-inventing the phone, you have to re-invent telecommunications. One of those that got out of the door when I was there was the Google "wave" project. Its vision was so big it missed the fact that people just wanted Slack[1].
Bottom line is that competition within a structure is good, its something that lots of organizations have applied successfully. Google's style, when I was there, was never structured enough to head toward anything. It always went all directions at once, leaving the organization with a net zero momentum in any direction.
[1] Yes they are nearly a decade different in time frame, but as a 'dogfood' wave user and then later outside of Google it was pretty clear the "win" was the IRC replacement, none of the vision stuff was getting much traction but it was the way to go if you're group wanted to figure out which cafe to go to lunch at that day.
Sounds like this was used as a bludgeon for attacking competitors.
Bottom line is that competition within a structure is good, its something that lots of organizations have applied successfully. Google's style, when I was there, was never structured enough to head toward anything. It always went all directions at once, leaving the organization with a net zero momentum in any direction.
My experience with such competition, is that groups actually tend to undermine each other instead of pulling together with cohesive effort. Here, the lowered cost of transactions within a company can allow other groups within a company to act as gatekeepers -- otherwise "enemy" in-company groups can be penalized with higher (outside company) transaction costs for getting things done. This is especially true when project budgets are made with in-company costs in mind.
Google is famously data driven (even when it hurts) and has taken a lot of criticism (justified in my opinion) about not being able to create any new business outside of search and search advertising. I saw, and commented on, the Alphabet re-org as a way to get better visibility into the data coming from the organization. Spend vs revenue. Robots aren't carrying their weight? Toss them over. Making cell phones too hard? Sell it off.
If you work in companies like this long enough you begin to recognize that the organization can constrain what products you can even attempt to build. When the Sun "Voyager" (aka the 'portable') came out we joked that Sun could build any computer it wanted to, as long as you didn't mind it looking like a workstation. A partner Tadpole had built an actual laptop. But the Sun hardware engineering organization couldn't get there from a clean sheet of paper.
For that reason alone it is critical that Google keep re-organizing until they find a formula that works for them.
Something similar happened to one of my previous employers. We were less than 100 people and cranking out many successful projects. Once we got acquired and nearly quadrupled in size, none of the projects succeeded. Many of the early contributors left one by one due to frustrations and pressure including me. The new management style had changed everything.
It is not uncommon for a software developer to be asked to keep a seat warm in an open office seating environment. It feels very unnatural to be expected to sit in front of a computer for 8 hours and be given nothing to do. Some people convince themselves that they are riding the gravy train, others go crazy and leave to find actual work.
I can't find any good citations online, but Electronic Arts was reputed to do that as well. Supposedly they'd put several developers on the same title, and whoever finished first got to ship.
The more confusing one was web vs. native - there were a dozen sessions on progressive webapps, yet they barely got a mention in the keynote, while Instant Apps got a full showcase. Internal competition is all very well, but it's frustrating as a developer to not be able to get a clear answer as to which platform we ought to be developing for.
In this case, however, given the prominence that was originally placed on the acquisition (at a time when Google had no direct presence in home automation), I suspect that the original vision was to actually merge the two worlds. The Google rank&file probably learnt very quickly how toxic Fadell was, and then proceeded to keep Nest at arms' length and do their own thing, basically writing off the acquisition and hoping it will eventually go away.
Sure there are mistakes but I don't think it's top down "team x" compete with "team y". It's more like both teams don't know about each other until the projects are far along. At which point might as well see which one does better... Or at least that seems to be the attitude at google.
(1) Top management tells everybody what to do. Result: No initiative from low-level employees.
(2) In case of conflicting projects, top management decides which team should continue. Result: Rampant internal politics.
(3) In case of conflicting projects, let all of them continue, and see which project wins the market. Result: Wasting resources; company looks disorganized to outsiders.
Some other companies (Apple?) avoid the problem by sticking to (1). However, many Google employees joined in the first place thanks to the tale of "A company where you can work on the problem you want to": it will be very tough for them to ditch the model wholesale, and it may harm the company in the long run.
I think Google does a bit of all three. While Google could certainly do better in some cases, it's not like Google's willfully ignoring an obvious solution.
It seems wasteful, but as we can see here, clearly the Nest work style couldn't compete with Google, and so no need to waste further R&D. But until you didn't the contest, how could you know?
My understand is that great product companies, like Apple, also have a kind of Darwinian development practice, but only ever reveal the "winner" once it's clear that it's the winner.
As an outsider, Apple looks like it has strategy and vision, while Google looks completely baffling, even releasing and pulling products from the market within weeks or releasing clearly incomplete, barely beta level products with no real vision and just expecting the public to sort it out.
I mean, how many chat apps is google up to now?
Interesting you say this as the first example that I can think of is Apple themselves when Steve Jobs was forced out of the Apple Lisa team and he joined the Macintosh team which then became in direct competition with the Apple Lisa team. I don't need to expunge on that but i found it to be a very famous case study on how 2 internal teams in direct competition affected the company.
So you have two teams somewhere in a company start a project at the same time.
Your challenge is, what does the corporate structure need to look like such that these two teams learn about each other's project early enough to collaborate?
If you rely on project information being passed up and down the corporate ladder, until one person learns that two subtrees of his reports are working on the same project, that's never going to work because bandwidth is reduced as you go up the tree, so less and less information is passed up, and early/experimental projects will frequently not make the cut and get filtered out. So your LCB will never hear about both projects.
If you just rely on direct communication between teams, you've got an N-squared problem as you grow, so that will never work.
If you try to give each team a carved out domain, new projects will frequently exist in the "chocolate & peanut butter" intersections, and both the "chocolate" and "peanut butter" teams will start the same project.
It's a difficult problem, and I'm not sure there's an answer between "companies should be small".
Deming said this over 40 years ago:
"The present system includes competition between people, teams, departments, divisions, students, schools and universities. Although economists have taught that competition will solve our problems, we now know that competition is destructive."
http://maaw.info/ArticleSummaries/ArtSumDeming93.htm
If in the end it makes the best product it's probably worth the waste of having competing solutions. But that is kind of a big if. (I would say the iPhone passed that test, personally.)
Even by Fadell's own standards, the Google acquisition doesn't seem to have led to huge wins for Nest. In an interview with Bloomberg shortly after the sale, Fadell was asked what the move would mean for Nest. "For us it was about getting out to the world faster" Fadell said. "[Our products are imported] in 96 countries where we don't even sell today. We can see these devices being connected but we can't get there fast enough. For us, going with Google, we can get access to resources to allow us to move these products around the world much more quickly." Before the Google acquisition, Nest was for sale in three countries. In two-and-a-half years under Google and Alphabet, Nest expanded to just four more countries.
Nest is, in many ways, the most disappointing company. I'm not sure how much of that is just down to them lacking vision or leadership though - they have the skills to do great things.
I had thought maybe they would do a smart lawn watering system next. I know it's a smaller market, but it seemed better aligned with the thermostat as something that could be marketed as a money-saver. The smoke detector doesn't save you any money, it just seems like a solution in search of a problem. The lawn system could basically sell itself: local rebates on the up-front cost, the promise of saving you water/$$, and the name that Nest had already established from their thermostat.
There are some other companies making products like this now, but I don't know if they're good.
The underground sprinkler system my parents installed 30 years ago has a sensor to prevent that, it's certainly not unusual. It would be interesting to see rain sensors for regular sprinters although I'd be shocked if they didn't already exist.
I know this is probably a minority opinion here, but one plausible scenario is that 5 or 10 years from now, Google will come to the realization that full no steering wheel autonomy is still "just over the horizon" and, in the face of declining ad revenues (or whatever), the whole effort gets ramped down.
Google's basic problem: search ads are incredibly profitable when you're #1. But they've saturated that market. Nothing else they do with their excess cash can match those returns. If they do something less profitable, the stock price goes down.
Hence Alphabet's search for a "moonshot". But what management meant by "moonshot" was another high volume product with a huge return on investment. That's probably not going to happen in the home appliance space, which is price-sensitive and competitive.
"The online-video unit posted revenue of about $4 billion in 2014, up from $3 billion a year earlier, according to two people familiar with its financials, as advertiser-friendly moves enticed some big brands to spend more. But while YouTube accounted for about 6% of Google’s overall sales last year, it didn’t contribute to earnings. After paying for content, and the equipment to deliver speedy videos, YouTube’s bottom line is “roughly break-even,” according to a person with knowledge of the figure."
[0]http://www.wsj.com/articles/viewers-dont-add-up-to-profit-fo...
In fact one can argue there is an inverse correlation there. The more efficient they are at developing new products with moonshot ideas, the less efficient they'd be at managing integrating a newly acquired company.
As an owner of many Nest products, and thus invested into the success of the Nest ecosystem, I'm also fucking furious.
What a waste.
They make a move, they lose. They don't do anything they lose as well, it will just take longer. So they chose the second option.
Nest needs to innovate and rapidly evolve products very quickly to keep ahead of the market, which it seems includes their colleagues (madness).
Given a team of that size and complexity nothing but the best management will get products out the door.
Reference: https://en.wikipedia.org/wiki/Infinite_monkey_theorem
Working on their resumes' ?
Hey Google, here is an idea, why don't you make an IoT button that can talk/control other devices like bttn or flic but costs just 5$ (I'm looking at you Amazon dash). They will sell like hotcakes!
After that you can introduce some other products, maybe under the home brand, like a device that control your AC (it just needs to work, doesn't have to be apple like the Nest) or one that monitors your water heater. What about a small device that checks for leaks in your pipes. There is so much stuff todo here.
What does the a button or a flood sensor buy them ?
Moreover chrome cast is used outside of google products like Plex, Netflix, HBO etc.
Having a low bar for entry into the home is huge for a company which is all about data and insights.
Holy.. what are they all doing?
Based on the stories, what's been happening at Nest is an extreme case of many of the same organizational antipatterns that people in the tech industry see in their own companies on a smaller scale. The things that have happened at Nest (if the stories are to be believed) are like cancer to a tech company. If left unchecked those antipatterns can grow and consume a company, and can ultimately kill it. No amount of reform will save a company if they can no longer retain talent or recruit replacements of the same or better caliber.
Given that so many HN readers are in the tech industry themselves, they probably worry about these kinds of cancers taking root at their own companies (I know I do). This stuff exists everywhere, it's just that for stable companies it's the exception and not the rule. I'm not surprised that it's been a hot topic.
Their performance versus the rest of the Google ventures? The product line that was previously leaked? And, in the eyes of Hacker News perhaps celebrating Tony stepping aside.
A lot of these stories, if they are in fact true, likely happened prior to Google acquisition and those people are now just feeling comfortable to speak out without the fear of it being spun as retaliation.
I have a lot of faith that the Google HR department wouldn't tolerate any of the behavior, which is why it smells like a marketing campaign.
Nest the company problem as consumers (myself being one) seems to be Software...
I use the nest software and simply put it's not up to scratch
The Hardware is beautiful if only the software was...
For example the nest protect has temperature sensors yet the Nest thermostat can not use these...
Nor can the software operate without access to the "cloud" even in a reduced ability.
Google OnHub suffers from the same Software problems, Great hardware and nice low level drivers however it cant even use IPv6 which for a router is pretty basic
my advice for the incoming CEO, interoperability Simple
make use of thread & brillo (software) to unify things... (produce an alarm but have the option of ADT as the responder)
Unify the products so they can work within a LAN and expand outside of the USA.
I hope that helps
John