Hacker News new | past | comments | ask | show | jobs | submit login
How Amazon Rebuilt Itself Around Artificial Intelligence (wired.com)
87 points by jonbaer on Feb 2, 2018 | hide | past | web | favorite | 30 comments

One thing which is sorely missed in such discussions is the employee efforts. Out of all the companies, Amazon's hours seems to be the worse of the lot.

That said, I really like Bezos (or Amazon) that they took risks on thing like AI, AWS etc. I find that lot of companies lack that insight. Even when they are not doing great, I routinely find executives hem and haw about their "core business". Most of these companies end up scraping the bottom of the barrel for profitability and keeping the company afloat for couple of years more.

An insightful comment that had to be made - I wonder why people downvoted it.

Indeed, Amazon is probably the best and the worst about the modern tech: an ambitious leader, who manages to create paradigm shifts in areas where he is not a subject matter expert; and a sweatshop-like environment where the low ranking employees are treated like oompah loompahs.

As an Amazon employee, I tend to think Amazon's reputation is overstated. Some teams definitely have very high operational loads, but in general my hours have been comparable to friends in similar tech companies and certainly less than friends in industries like medicine, management consulting, etc.

It's very strange to describe Amazon (at least engineering roles) as a sweatshop considering how well compensated employees are.

Well compensated? Amazon pays significantly below other mega cap tech companies, and as a PhD student in machine learning with a lot of friends interning and applying for jobs I can confidently say Amazon is nobody's first choice due to low salaries and working conditions.

I'd rather say I am amazed they get anyone great working for them because it really is clearly and painfully obvious that almost everyone would rather work for Google, Facebook or Microsoft regarding machine learning.

Have first hand knowledge of salaries. Amazon offers 30-50% less than other top tech companies for machine learning. There is literally no silver lining or upside as far as I can tell.

In that context, it does not matter if Amazon pays more than <other job> in <other industry> or treats employees better than <other company> in <other industry>.

Every measurable I've seen has Amazon very close to total compensation for similar levels. The real differences are in perks and vesting schedule (theirs is notoriously bad - 5%, 15%, 40%, 40% I think)

It's hard to make a case that they underpay by 50%, have horrific working hours yet still have / retain an engineering workforce that's talented enough to deliver consistently.

This is selection bias. You have first hand knowledge of what Amazon offers new graduates with little or no industry experience (and yes, PhD students fall into this category). The compensation for more experienced employees is substantially higher.

"The compensation for more experienced employees is substantially higher"

Citation needed. Higher than compensation for experienced employees at Google/Facebook? I am not even disputing Amazon pays more later, I am saying Amazon never pays top tier AND treats employees badly AND has no perks or particular benefits to make up for it, so nobody who has the option (that I know of) wants to work there.

Every company pays more later, I was trying to say that the pay gap between other companies disappears for mid+ level positions.

Also, they do not treat employees badly. Are there cases where certain managers treat their employees poorly? Of course, you can find the same anecdotal examples at company of sufficiently large size. Since the infamous 2015 NYT article, employees are treated very well on average. For perks, if you're referring to free catered meals then, yes, Amazon lacks there. But how about working in downtown Seattle instead of some random suburb outside of San Jose?

Finally, you ask for a citation about my claim while offering none for your own. My source: I work here so I actually have first hand experience.

That's surprising, if it is indeed that large a gap.

Are you sure you were not comparing positions between product groups and research? They tend to have different compensation.

No I am absolutely certain, I was offered more than 2x for the same position by another tech company.

I was so surprised by Amazon's offer I conducted a small survey amongst friends and acquaintances to find which of my offers (there were more) was 'normal'.

The thing is, you can, as Amazon shows, have an enormously successful business by treating employees badly and not paying top of the market. Of course you can. Paying more does not always get you the best, and for many many tasks you don't need Jeff Deans, you just need someone to work hard and get it done.

There is tons of talented people willing to work for less than the top to get a chance to prove themselves, to learn something, to get their shot. It's just that they tend to leave Amazon (c.f. retention rates) after that, and Amazon seems to be doing ok with that. It's just a fundamentally different approach to employee management.

It certainly sounds like a sweatshop in the descriptions I've heard from friends who work there, or have worked there in the past, compared to the experiences I and others have had in other large tech companies. I can only think of three people I know who have stuck it out at Amazon for more than two years. The egregiously back-loaded vesting structure is clearly built around the expectation of a high attrition rate, so I believe that the meat-grinder effect exists by design and not merely incompetence.

When I get stressed I just open my stock vesting page and say "two more years." :^)

> It's very strange to describe Amazon (at least engineering roles) as a sweatshop considering how well compensated employees are.

Amazon employs over half a million people (https://www.statista.com/statistics/234488/number-of-amazon-...). I doubt engineers make more than 5-10%, someone has to mind the fulfillment centres, and the fully automatic robotic future is yet to arrive (https://qz.com/1039169/amazon-is-looking-to-hire-50000-peopl...).

The blue collars make up the bulk, and while the accounts of white-collar sweatshop conditions are argued about, nobody ever contradicted the accounts of blue-collar workers, and numerous complaints have been filed: https://www.cbsnews.com/news/amazon-settlement-could-pave-wa....

>It's very strange to describe Amazon (at least engineering roles) as a sweatshop considering how well compensated employees are.

With a company as large and complex as Amazon, many things can be relative, but if you work as an engineer or programmer at Amazon, chances are that conversation doesn't apply to you, or at least, that it applies less to you than to others.

It is the inverse of Dunne's description of the free press; Amazon's job, it seems then, is to afflict the afflicted and comfort the comfortable.

I've heard there are signs in the elevators imploring employees to stay past 4 o'clock and enjoy the downtown amenities because everyone leaves at 4 and the downtown congestion gets pretty bad.

I know engineers at Amazon, many of them work very comfortable hours.

"Amazon's hours" - I think this notion is misguided.

Just like every tech company; if you have good working habits, and can push back when given too much work (having a good manager helps), you can have a great career at Amazon. I work at Amazon and the best employees, the ones who get promotions and deliver great work typically work 9-5 hours.

I'm sure it varies between teams but I'm in year three of being a SDE and I would recommend Amazon highly.

I have to say I'm disappointed by the mild boosterism that runs through Steven Levy's coverage of large corporations. For example:

> Amazon’s product recommendations had been infused with AI since the company’s very early days,

Really? Is it the kind of AI that impresses people today? Or was it just collaborative filtering. This narrative that "things were good and now they're getting better" is common in corporate communications, but shouldn't be part of neutral coverage.

Or this:

> “But they have really come on aggressively. Now they are becoming a force.”

> Maybe the force.

That's absurd. When you look at where Google has taken TensorFlow, AlphaGo and Waymo, and compare it with really any other company in the world, it would be hard to anticipate that Amazon would become the force.

Amazon was forced to adopt an outside machine-learning framework, MxNet, as its own when its in-house toolkit didn't get traction. That's in stark contrast to Google, again.

I'd say Amazon's main advantage is in its ability to sell infrastructure and platforms that enable developers to build apps -- which in this case will help productize ML in various ways. But I would not call it the force in AI.

Amazon’s Alexa is a big win for the company. It will directly increase shopping on Amazon. Integration with music and Prime Video impact Apple

It might soon be the most popular, and perhaps best, consumer voice assistant:


This reads like a Siri article from 6 years ago. "All that data from people using it means it will only get better and better, and as it gets better no one will need to use their smartphone or search the web anymore." Right...

> It will directly increase shopping on Amazon.

Is there any evidence to this speculation?

Linked article has a link that just points to the sales of the device itself as proof.



Amazon is definitely working on improving this.

“Alexa, what are my deals?”


Sounds like they will be investing significantly more into Alexa:


That doesn't say there's an increase in sales, just that people buy over VUI. It's very likely that buying over VUI cannibalizes web and mobile sales.

yes there's also this surprise element attached to it. You won't know if you have ordered a right item in right qty until you receive it or you immediately open the orders tab in your phone to make sure you ordered the right item.

Or when you say, "Alexa, order me a Chromecast" and it ships you a firestick instead.

I've heard just the opposite from someone I know at Amazon. I've heard that they still gather a lot of their data by doing Sql queries.

SQL is highly underrated by engineers.

FYI, unless Ghostery enables Adobe Dynamic Tag Manager, this article will be blocked by Wired... made invisible.

I'm having no issue with everything in that page disabled.

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