[jsnell]

Why do you trust AWS when Amazon kills failed non-AWS products left and right? I even remember Bezos bragging about this being a core part of the company culture. Why do you trust Azure when Microsoft's graveyard of non-Azure graveyards has long since overflowed?

For me, it's because Amazon killing their failed phone, their failed Yahoo Answers clone, their failed search engine, or their failed paypal clone says nothing about their commitment to their phenomenally successful $45 billion / year business with high margins and 29% growth.

Google Cloud is a $14 billion / year business growing at 45% / year. It is a smaller business than AWS, sure. But not by an order of magnitude. It's basically 3.5 years behind Amazon on the growth curve. Were you afraid in 2017 that AWS would be killed due to being too small a business?

[whoknew1122]

>"Why do you trust AWS when Amazon kills failed non-AWS products left and right? I even remember Bezos bragging about this being a core part of the company culture."

I'd be really interested in this quote and its context. At AWS, things are always iterated upon and tested. But there's an incredible emphasis on two-way door decisions. Basically, don't make big decisions that can't be reversed if things go poorly. And once you've released something customer-facing (especially something as important as a new AWS service), you're locked in for the foreseeable future.

There are still CloudHSM v1 HSMs running out there in the wild. CloudHSM v2 was released in 2017. And CloudHSM v1 hasn't been available since at least 2019.

There are still EC2 instances running in EC2 classic--that is, EC2 before VPC was introduced.

[jsnell]

It's a pretty common theme in what Bezos says, no? Here's some examples: https://www.investors.com/news/management/leaders-and-succes... has some of such quotes:

> "Failure comes part and parcel with invention," Bezos wrote in his 2013 letter to shareholders. "It's not optional. We ... believe in failing early and iterating until we get it right." Three years later, he added, "Amazon is the best place in the world to fail."

Or:

> “As a company grows, everything needs to scale, including the size of your failed experiments. If the size of your failures isn’t growing, you’re not going to be inventing at a size that can actually move the needle,”

Now, I'm sure that doesn't apply to AWS, and your culture is totally different. You're an enterprise product with real contracts, real commitments, and making real money. You're not the Groupon clone, online pharmacy or whatever that grocery service *Amazon killed today* was.

But if we're accepting that Amazon's general trigger-happiness when it comes to failed consumer products doesn't extend to AWS, why does Google Cloud not get the same treatment?