[felixgallo]

You seem to have linked to Google's slightly inferior, copycat cloud offering.

I read Steve's rant as a cri de coeur not for an AWS-style infrastructure as was relevant and genre-busting at the time of the essay's inception, but for Google to stop following slowly with slightly inferior, copycat versions of other people's grand ideas and expecting anything good to happen out of it.

Granted, Google has some pretty great, if in-the-small, consumer-facing initiatives: think Street View.

But as one of the primary governments of the internet, they could be doing so much more to move the needle for computing and advance the general state of the art.

Where's googlelang?

Where's infinite instant CI for any language?

Where's infinite immutable package versioning for every language?

Where's the big thoughts?

[nostrademons]

It takes time to move the needle. Google has probably the best infrastructure in the world, but it was designed for Google's problems, tightly intertwined with the products it supports, and built for a different performance/productivity trade-off than most independent customers are willing to put up with. Even now, GCE beats EC2 on price/performance metrics, it just lacks some of the conveniences that AWS users expect.

[felixgallo]

That's kind. You could also say that google has had a mediocre, confusing, poorly supported also-ran cloud offering for years and shows no sign of doing anything but continuing the current path.

The needle is moved by having nontraditional ideas and following through with commitment. I know Google has enough great minds for (a), where, bafflingly, is (b)?