Hacker News new | comments | ask | show | jobs | submit login
Amazon Is Hiring More Developers for Alexa Than Google Is Hiring for Everything (forbes.com)
111 points by eplanit 10 months ago | hide | past | web | favorite | 86 comments

This comes down to the fact that each Amazon team manages its own hiring (and so has its own job postings), whereas Google has one top-level posting per role.

Good observation. If true, this story is a complete waste of time. Apples/Oranges.

Isn't it a good observation _if_ it's true?

Urgh. Okay. I thought it was an insightful comment but didn't invest the time to vet it fully. Was trying to highlight the comment because (again, if true) it basically invalidates the whole story. That doesn't happen often. The notable part is that one short comment (could) make the story shit.

If google does indeed limit, for example, "SRE II/ Location" to one posting, and Amazon expands that to, say, 100 postings, then the premise is worthless. I don't know how to confirm that.

This is not exactly how it works. Managers at Amazon can request headcount all they want, but the request bubbles up and needs to be approved. Sometimes, there are top-level hiring freezes as well. The planning and hiring process is more complicated than that though.

He didn’t mean individual teams get to decide how many new people to hire. He meant managers make their own job postings for each small team.

Wanted: Software Engineer. Various levels, focuses and locations needed.

Disclosure: I work at Google.

As others said, there's no data on how many people Google or Alphabet are going to hire (postings are not 1:1 with individuals). Alphabet apparently hired 8,057 people in FY 2017 [1]. Note that while HTC had ~2000 people come over [2], that was not closed within FY 17.

[1] https://abc.xyz/investor/pdf/2017Q4_alphabet_earnings_releas...

[2] http://www.businessinsider.com/google-htc-employees-acquisit...

How do they determine the number of open requisitions? Seems like number of job postings is misleading since one posting could be used to hire multiple people.

Exactly what I was thinking.

Google, for one, likes to hire a lot of people under just two-three job descriptions. The author doesn't take any of that into account and it's not clear what the original analysts said, either — perhaps they compensated for this by calibrating numbers against historical employee headcount. Which is still broken, because, historically, Google had varying mixes of engineers vs. non-engineers, for example (but also: employees vs. contractors; lots could be said there).

Yeah, wouldn't surprise me. We have 2 public "open positions" for software engineers at my company, when the actual number is more like 80. We just don't post multiple time for the same role/level

Seems like they don't.

A quick Google search (ha) seems to indicate that Alphabet's growing by about 10,000 people per year or so. Let's be fairly conservative and say 20% of those folks are programmers. That's still much more than 1,147 people.

Besides that, the sentence "Amazon is hiring 1,147 people for its Alexa business unit alone" is very different than the headline, which is "Amazon Is Hiring More Developers." Not everyone working on Alexa is going to be a developer.

Maybe I'm misunderstanding, but as stated, this whole article fails the sniff test to me.

Check the link on the second page of the article about what kinds of jobs are open in the Alexa division. Pretty solidly development/technical/product/research jobs

20% sounds like a very low estimate for a Software Company. Especially since we are implicitly assuming 100% on Amazon's side. If you want to compare Apples to Apples, take 100% of each side.

As an example, lets look at Google Boulder office for software engineer.


As of right now there are 5 postings. Software Engineer, Engineering manager, student research, and Software engineer new-graph/PhD.

That "Software Engineer" title covers many job levels and teams. I can't even imagine how many people are hired under that same title in Mountain View or other locations.

Wouldn't a company have to post every single position because of labor laws, etc. They can't just have one generic position to cover many, or? That would open the doors for bad hiring practices and just picking anyone you want (without proper due diligence) rather then the best candidate possible.

> Wouldn't a company have to post every single position because of labor laws, etc

No, labor laws in most of the US do not require posting any positions, much less every individual position. (Civil service may require posting civil service positions individually, or at least explicitly identifying the individual positions being hired under a combined posting, and labor agreements may specify requirements for how and where union-represented positions are posted.

> They can't just have one generic position to cover many

Yes, they can.

> That would open the doors for bad hiring practices and just picking anyone you want (without proper due diligence) rather then the best candidate possible.

Bad hiring practices are not generally illegal; those that are (racial discrimination, for instance) are prohibited directly rather than by prohibiting things that might open the door for them.

What about H1B sponsorship? I thought that required some sort of advertising of the position at least.

Google is hiring internationally (e.g H1-Bs) for that you need to publish all positions afaik - because you need to give Americans the possibility to apply and have the position public for a while.

AFAIK, yes, for positions for which you would sponsor an H-1B, you need to have posted publicly and demonstrated an inability to find a qualified American candidate.

I don't think (but I could be wrong) that requires separate, per position postings or even publicizing the number of positioms: if you publicize for a general job description and can't get adequate American candidates the openings for that position, that can, again AFAIK, satisfy that part of the requirement.

No, that is pretty much how they and other large tech companies hire.

You apply as a SWE or SRE. Then you basically either pass their hiring bar or not. If you pass, they find out what you are interested in, what you would be a good fit for, and what team is interested in having you.

There are exceptions of course, for specialized positions or knowledge. Or if they have big gaps to fill they might do targeted recruiting for them.

I don't see why it would run awful of labor laws. They have a very structured hiring process, it doesn't allow much room for "picking anyone you want". If you pass they find a job for you.

No, that’s not how it works in most of the world.

Also, you are not required to show that you picked the best candidate possible, just that you didn’t discriminate against someone on the basis of protected categories.

Well with an H1B you have to show you couldn't hire an American at the prevailing wage. So it gets murky.

I think if you want to hire internationally it's quite different in the USA.

anecdote: I did a google team match two months ago as a new grad and was asked if I would be interested in "Search Infrastructure". I said yes and got set up with 5 calls with different Google Assistant teams.

I don't if Amazon/Google would ever do this or even need to, but my last place used to open reqs and do interviews based on "pipeline" openings which means jobs we might need to staff quickly if new work comes in, but we're not going to call you back if it doesn't.

Great! Maybe in the future it’ll be able to understand, at 10pm, “Alexa wake me up at 7” without asking if I mean “In the morning, or in the evening?”.

Considering the time this has been out and the apparent resources, so much of the basic functionality is locked behind formal command patterns rather than anything natural (and every “new features” email I get seems to just be variants of “tell me a joke/fortune cookie/fact”)

Picard had to remind Computer that he likes drinking his tea "hot" 15 times a day -- all signs point to voice assistants forcing us to be hyper literal because they like to mess with us.

Holy shit, that first one is so annoying. I don't understand how even the shoddiest of voice assistants on the market get it right but Alexa doesn't.

Wow, how hard is it to say 7am?

This is new tech, I think your expectations are a little too high. Remember windows 3?

The thing is, ALL other voice assistants I have tested get this right. I have submitted three reports to Alexa support about this very thing and they don't care. It's ridiculous that doing something that is so common would be so ingratiating.

If I meant 7pm I'd say 19:00 instead.

Please tell me one of those new hires will help to release an Echo/Alexa Bluetooth Low Energy developer SDK.

For years many developers of BLE hardware have been waiting for a BLE capable smarthome hub to win the market. Amazon has won but whenever I ask about the BLE SDK, I am told to "stay tuned". They have used that answer for years now.

That means my company will reluctantly release a Wifi Bridge this year. I say reluctant because we aren't motivated to collect $50 per customer on a bridge that will become instant garbage if Amazon does finally release the SDK. We are motivated by how much value we offer our customers, not by how much money we can collect from them.

Does anyone know if Amazon is actively working on a BLE SDK for Echo ?

I really hope that you work for a dallas based smart lighting company.

They must have high hopes. Alexa is not something I can ever imagine using. I like Amazon's web store, but Alexa doesn't seem to offer anything I would ever want. Do others find it useful?

My neighbor's wife uses it for things like setting a timer for cooking, setting reminders, playing music, queries about things she and her kids are curious about, and ordering things on Amazon. She seemed pretty happy with its usefulness.

>My neighbor's wife uses it for things like setting a timer for cooking, setting reminders, playing music, queries about things she and her kids are curious about, and ordering things on Amazon. She seemed pretty happy with its usefulness.

This is why I will never understand the appeal of these "smart assistants". All of the above can be done with simple voice command recognition that has existed for decades. Why do I need to send my commands through a third party selling my data and tracking me?

Sure, but Alexa does it much better. It just works. With voice platforms the biggest hurdle is not trusting it will work (like with Siri) which means you stop trying. When I drop pasta in while watching over a saute pan I don't want to go open an app I just want to say "set a timer for 8 minutes" and get on with it.

Why does any software need to send commands back through home base? Embedded systems are powerful these days and storage costs are tiny. I’m so tired of these “smart” devices that needlessly require an internet connection.

In order to watch video from my Dropcam over in the next room, that video stream gets uploaded through my internet connection to some backend, then downloaded back through that same connection to my viewer. Madness.

I agree with you on the dropcam issue. They should've done two TCP streams. My guess is it's a hardware limitation or something. Also their big thing is to sell cloud DVR.

I’m tired of “smart” employees that needlessly require Google.

Has there been a device before that worked as well as Alexa? Remember, user is not a programmer.

I have read like 1/3 of the comments but its interesting that the most interesting problem of everything we are talking about was only noticed by a non-programmer (so far). Everything that was talked about in the comments all hinge on the state of the art (SOTA) of ai in natural language processing (nlp). John Carmack recommended https://www.youtube.com/watch?v=OQQ-W_63UgQ stanford's nlp class (i didnt know who he was until i saw that post on hn). That class is good. it should take a long time to grok. you'll be solid up to LSTMs a general current standard.

I can do it without a voice assistant even quicker.

Exactly. I'd love to use one of these in the kitchen to set timers and ask ingredient amounts -- so long as it was local only.

Playing songs with the kids (this is actually a lot of fun, changing songs while wrestling with the kids, lots of laughs =) ), audio books, news, weather, timers, answers, buying things, turning things on and off, such as the lights. I can do all that without doing anything more than talking.

These are the average things my wife and I use ours for at the moment.

Edit: Hey, and just to be clear, yes, we find it useful enough we have several devices.

I dunno. I can and do do almost all of those things(minus maybe changing songs while wrestling with my kids, but I don't have kids) without a smart speaker and none of them are very difficult to do in the first place. For instance, when I want to buy something, I open an app on my phone. If I want to turn on a light, I walk to the light switch, and then I turn it on. I'm not trying to be facetious, I really don't see much of a benefit to having an alexa or similar device. I think smart speakers are going to have to do a whole lot more than I can currently imagine in order for me to adopt them.

Sure, you can do all those things, but I can still do them when you can't. Like when your hands are holding something, or dirty because you are washing dishes, or turning on the lights while your hands are full.

And the benefit is, I don't have to remember my phone. It can be sitting wherever, and I can still do these things.

> I really don't see much of a benefit to having an alexa or similar device.

So? I like having a remote car starter, too, despite being able to walk out and start the car manually.

I like GrubHub, even though I could call up and order from places on the phone.

I enjoy streaming video over DVRing shows.

I value convenience. Maybe you value other things. That's fine.

So then...

> I dunno.

What? What don't you know?

>If I want to turn on a light, I walk to the light switch, and then I turn it on

Compare this to: "What's the point of remotes? If I want to change the station, I get up and walk to the TV". Once you get used to being able to turn lights off by voice, you get lazy and happy.

And it's not just one light, BTW. Suppose when I go to sleep, I want all lights except the porch lights off. I don't have to walk around the whole house[1] switching them off, when instead I can just say "Sleep mode lights" or something equivalent.

[1] You know, the 3000 sq ft houses people outside of SV have.

I typically turn a light on once when i enter the room and when I leave the room- I do the same thing for my exterior lights on the outside of my house. Once every couple of months or so I or my wife will have to get up and go shut a light off in the living room when we go to bed that we forgot to turn off. People often change the channel multiple times in on sitting...which you really don't do with lights. I guess maybe having children complicates this.

Also, maybe I'm the outlier here your proposed situation is particularly accurate (and funny) to me- a couple of years ago a dog chewed up our TV remote and I have still yet to replace it. I get up and physically go to the TV to turn the volume up or down or to turn it on/off.

I'm really not trying to sound like a luddite, I just really don't care enough to adopt this stuff, at least not for the foreseeable future...I don't see the purpose of life as an endless march towards total comfort and ease, though. I will concede that these devices and smart home technology makes life much easier for those with disabilities, and I ultimately support it's development for this reason alone.

Well, no one is saying this is definitely a gain for every person. As an example, I refused to use a smartphone until 2012, and even after that, I refused to use cell phones for non-emergency calls until, well, even now. I still have a conventional phone at home, and unless I'm coordinating something, I don't answer my cell phone and simply call people back when I get home.

But on the flip side, it's a poor assumption that if someone is not turning off the lights manually, that the primary goal is comfort or ease. I used to think manually doing all chores was a virtue in itself - I had a dishwasher in my apartment for years before I started using it. I no longer think that way. Now for me, if I can eliminate mundane stuff like dishes, lights, etc - it is because I want to spend the time on more quality stuff.

As an example, too many older people say they wish they had just paid the money to repair stuff in the house, rather than spending the time to repair it themselves. They took pride in their work, on top of also saving money. But later in life, they wished they had used that time better.

I definitely agree with some of what you're saying- I guess my point would be that I don't think that these smart speakers are anywhere near the innovation that the smart phone was...and I think that people who aren't early adopters have a valid place in the consumer space because we are the majority and therefor we are growth. If things like Alexa want to be as ubiquitous as a smartphone then I think they're going to have to become a lot more novel.

In its current form, I find it less useful than a lot of people seem to. But reasonable extrapolations of natural language processing and other advances in conversational interfaces make it seem at least plausible that Alexa could become a fairly general-purpose virtual assistant. Lots of hard problems to solve but at least plausible in 5-10 years.

> Do others find it useful?

They have sold tens of millions of Alexa devices, is it really surprising that people find them useful?

That logic could also apply to fidget spinners.

> That logic could also apply to fidget spinners.

Or stamps. Or shoelaces. Or umbrellas. Or wallets. Or lipstick.

All useful things to millions of people.

One fundamental difference with fidget spinners: they have already peaked and faded in popularity, Alexa continues to grow year by year. That's the true difference between something tens of millions of people (soon to be hundreds of millions of people) find useful, and something people temporarily find amusing.

... You write as if Alexa wont also 'have already peaked and faded in popularity' in a couple of years. Growth doesn't mean shit. Retention does.

The example of fidget spinner is to highlight a counter example to the belief that selling well is an indicator for usefulness. It isn't.

That's a stretched supposition. How do you know something is actually useful before you purchase it? You're ignoring the sell-ability factor- At 50 dollars it's something people can give to another that offers interactivity and an experience with less guilt if you don't end up using it than a goldfish that gets flushed down the toilet.

The customer reviews are across the board overwhelmingly positive. 4.3/5 with over 100,000 reviews of the Echo Dot for example.

Also the facts that they keep bringing out new devices and are hiring extremely heavily would lead you to believe they see a very bright future in the product.

When Alexa can control microwaves, ovens, blenders, TVs, phonelines, dishwashers, bathtubs and showers, Roombas, pressure cookers, WOW holy shit, an OS for your entire house.

I envision pick and place robot arms that swing around your kitchen grabbing things and preparing meals too.

Home automation is something that can be handy. Example: When I'm watching a movie, I like all the lights nearby to be off. Being able to say "Turn off TV lights" and have specific lights turn off is nice (and expensive - all those smart switches cost money).

Being able to make a sound when an action occurs (e.g. garage door opens alerting you that someone has arrived home) is nice.

Not sure what else. But satisfaction seems to be high. I think it falls squarely into the category of "Seems not worth the expense, but once I have it, it is kinda convenient".

(I do not own one).

I tend to agree, the trust I have in Amazon around handling my information security is quite limited - I mean that goes for a lot of cloud providers, but Amazon just doesn't seem like an organization that cared much about data protection. I once submitted a couple of bugs to them and wanted a CVE, but only thing they said is. Thanks you, we fixed it but we won't acknowledge any of this externally via CVE, because well, it's the cloud, we don't have bugs or something along those lines.

> but Amazon just doesn't seem like an organization that cared much about data protection

I challenge you to provide a single shred of evidence in support of this claim.

Did you read my paragraph, I have a first hand example. Not being transparent is exactly part of this

Perhaps the bug you submitted wasn't worthy of a CVE. It's impossible to know from a single anecdote with basically no details.

What I do know is that, to my knowledge, there has never been a single verified publicized breach of any AWS systems that were due to a vulnerability in AWS itself. Care to provide a counterexample?

Anecdotally, it's certainly gaining market share. Most of my friends have an Echo/Alexa at home now (silicon valley selection bias disclaimer). Even though they do not use it much yet, I can see why Amazon would want to capitalize on this.

We started with Alexa but now replaced with the Google Home.

I am a super curious person by nature as also a couple of my kids and do get a lot of use out of the Google Homes. Often times during family discussions.

It is a $150 kitchen timer. And it is worth the money.

You could bolt a Kindle Fire HD to your wall and install nothing but a timer app and save money.

Amazon is also laying off a lot more developers than Google is.

Not laying off- just reducing headcount totals. I believe I am in the org being effected.

People keep moving to Alexa and not being back-filled.

They also lose engineers much more quickly. Back when I was there, they leaked SDEs like a sieve.

I know Google's overall headcount numbers for the next year, and while I can't disclose them, I'll just say this article is silly uniformed nonsense. In previous years, the reports they have produced have been off by a factor of 10 off what Google later reports in their SEC filings. So I'd take it all with a large grain of salt.

Any insight on why Facebook hiring is surging? (13% of workforce). Is it mostly on the AR/VR side?

How many will be there in 2 years or how many have left in the past 12 months? They have insane churn, something like avg tenure is 16 months for an engineer. Between burnout and pip I wonder if this is just hedging bets.

“Amazon is hiring 1,147 people for its Alexa business unit alone,...”

I’m sure Google is not going let Amazon run away with the voice assistant market.

There’s nothing like an arm’s race between tech titans to move the needle.

With the additional efforts of Apple and Microsoft, voice as a user interface will soon be a solved problem.

70% of people have bought alexa.


I think amazon has a pretty big lead in the voice assistance market.

I know you can use google assistant on your phone, but if you have your phone it might easier to just type it in.

Amazon has 70% of the voice assistant market. That’s not percentage of total households.

Amazon has taken the Echo global so look for hundreds of millions of users soon.


Which people? Surely not all people.

> 70% of people have bought alexa.

No, the claim in the source that that article is based on is that 70% of people who interact with a “voice enabled speaker” at least once a month do so with an Alexa.

Lets supposed the true sales were 50/50. That would indicate a majority of google people are just buying the equipment and not ever using it.

The true sales of “voice enabled speakers” could be 100% Amazon, and they could still not lead the “voice assistance” marketplace, or reach the claimed 70% of people. The big issue isn't confusing usage rates with market position (those probably are close, though in theory they could be divergent), it's confusing the voice assistance market with the voice-enabled speaker market (and then also confusing the current market for those products with the whole population.) Voice assistance isn't about selling dedicated voice devices, it's about voice-enabled services, for which dedicated devices are simply one gateway. Dedicated devices are a small share of voice assistant usage, far behind smartphones, and even behind the combined category of tablets and general purpose computers [0].

[0] https://www.voicebot.ai/2017/12/12/46-percent-americans-use-...

Makes sense as the Echo has some catching up to do. We started with the Echo but since switched to the Google Home as it is far more capable and also you can use natural language for most things versus the Echo requires much more rigid language.

I've just hired an extra 8 women, since me and the wife want a baby next month. /s

The Mythical Man Month doesn't mean "don't ever hire". It just means "hiring is not a short term silver bullet," and I have a feeling Amazon is thinking about the long term. (Bezos talks a lot about how he's currently planning 3 years out.)

Apple has 800 people working on its iPhone camera.


Do you think that natural language processing, knowledge graph, translation, speech recognition, multiple hardware devices, etc will require fewer people?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact