I agree that you do not have to be a big company to make a product like that, but it seems like you have to be a big company in order to host and deliver it.

> I agree that you do not have to be a big company to make a product like that, but it seems like you have to be a big company in order to host and deliver it.

The tragedy isn't in what you've said, but rather what you haven't said. The implication is already that a product of GDocs/GSheets quality should be free, as part of a large company's moat, rather than a paid for product that people will pay for.

The tragic reality is thanks to these large companies turning what would otherwise be successful standalone businesses, into free additional features.

I've used that word because Steve Jobs famously described Dropbox as a feature. Google has effectively made MS Office a feature. Apple effectively made operating systems a feature, by giving away macOS and iOS for free with their hardware sales.

Increasingly, everything becomes a feature, in search of what? For big tech, it's to sell users attention.

Meanwhile, on the other side, big media is charging us to give them our attention...

Nah. Hosting is cheaper than it's ever been at any time in history, the costs only become a concern if you have lots of users in which case you should be generating lots of revenue to pay for the increased hosting costs.

Have you ever been on-call for any project larger than a toy? If you have, you likely noticed that keeping the whole thing up sometimes takes effort, more effort than meets the eye.

Hosting as in having some code deployed to some machines is indeed cheap. Keeping a large app like g.docs up and running, especially without breaking the bank, is a bit more tricky.

> Have you ever been on-call for any project larger than a toy? If you have, you likely noticed that keeping the whole thing up sometimes takes effort, more effort than meets the eye.

Of course, that applies to literally every piece of production software ever, but keeping a webpapp running really isn't that hard, it's honestly the bare minimum of competent software development, if you have a team of SREs up at 3am triaging the site every night you're doing something wrong. Now of course, when you get to google scale, you will encounter unique problems, but if you're at google scale your business has more than enough revenue to pay for the costs.

My experience as well. Web apps - if made slightly streamlined and lightweight - with thousands of visitors a month is easy peasy on cheap webhosting. Google is another scale ofcourse. That's like comparing elephants with mosquitos.

Hosting is at an all time low, while innovation, time and putting time to a project like this is not.

There's more to hosting than just spinning up a fleet of containers you know (not even that that is necessarily trivial...).

Also, how would you be generating revenue from these users? Just do what google does and run ads on the side? Then what's the difference?

Ofcourse there's more to hosting. But that's what the hosting company does! The webhosting landscape - at least here in Europe - is perfect: worldclass technology, local service. I can be on the phone with these companies if there's a problem.

> There's more to hosting than just spinning up a fleet of containers you know (not even that that is necessarily trivial...).

Spinning up containers is like the bare minimum I'd expect from an ops engineer, I wouldn't call it "trivial" but it's the job.

> Also, how would you be generating revenue from these users?

That's a business question, not an engineering one, if the businesses doesn't have a plan for revenue then hosting costs are irrelevant.

Hosting stuff that hosts content some people find "problematic" has its own additional layers of difficulty. Amazon is completely willing to dump you if they disagree with you.

There are a few cases of this happening, but it's not common. If you intend to host "problematic" content pick a more understanding host or colocate.

No you have to be a big company to resist the urge of getting bought out.

Or you have to either have aspirations of becoming a big platform company or a plan to survive and be happy watching big companies push you to fifth place in a category you once dominated.

The question isn't whether you need to be a big company or not, it's where you're gonna get the money.

What you'd need to host/deliver something like Docs/Sheets is: a product team (2 QE, 8 SWEng, 3 SRE, 1 product owner), the product, some cloud infrastructure, and the capital to pay for it. You could go larger than that to build it, but that is plenty of people to run/maintain/support it. Assuming "large scale" is between 1M and 100M users, figure between $750K and $3M for infra. Combine that with median salaries for employees, and you're lookin' at between $1.75M and $4M (before taxes/business fees).

If you use the cheapest infrastructure and labor, you could do it for $500K. (it is mind-blowing how cheap offshore labor is. Google engineers get paid almost 8 times what some of our contractors get paid, for about the same work)

VCs throw that much cash around for a weekend trip to French vineyards. If you can actually get paying customers, even better.

Anything that becomes a danger of breaking away from the ecosystem gets bought out. Indeed that seems the end goal of most startups

> Google docs etc aren't even Google's inventions, Google just bought them.

That's the same thing as saying that macOS Monterey isn't Apple's invention, they just copied Xerox.

There's years of development on what Google bought and what Docs suite is now and any engineer that had developed a product for years shouldn't say silly stuff like your sentence.