Hacker News Reader

Comment by d3nj4l | Hacker News Reader

Server costs are only part of the problem. They had a team of devs they couldn't afford to pay, they were hosting a lot of media (especially images) and they never quite hit the traction to make the project feel worth doing for free. I fully sympathize with their decision to shut it down instead of letting it run on fumes.

View on HN

I mean, this is a blogging platform. Users, posts, comments, likes, that's about it. People were making blogging platforms since forever. Every web framework tutorial has a "hey, let's make a blog" part.

I don't see why would it require a whole team of fulltime devs.

Inflating team size, constant feature churn etc all make sense for a VC-funded company. But they were kinda making a statement against it, so there is bo reason for them to have a whole team to maintain a blogging platform.

It's hard for me to estimate the cost of media hosting, but again, Hetzner (I'm not advertising them, it's just one dedicated server provider I heard about. There are many others) offers 20 TB / month of free traffic. Is it not enough to serve images at cohost scale? I genuinely do not know, I would love to hear estimates from more knowledgeable/experienced folks.

Or was storage the issue? How much storage are we talking about?

I'm not sure if you actually read their previous financial reports, but their expenses were somewhere around $50,000-60,000, that doesn't indicate an inflated dev budget.

If you add up server costs, the cheapest possible accounting and legal services, and any other miscellaneous business costs, it doesn't really leave much of a budget for development.

No, I haven't read their financial reports because I didn't knew they exist. I will take a look, thanks.

I only knew of them because the top-level comment specifically references the financial updates and some of the numbers from them.

https://status.cohost.org/

I see digitalocean,datadog, aws, mailgun, managed databases kubernetes.

That costs much money, that are better spend in some classic server infrastructure like a dedicated server.

The other commenter, Throwaway3... is getting downvoted but I agree with the costs being excessive, tech-wise.

This didnt need kubernetes, datadog, all that stuff. This needed a couple of hundred a month in S3 or cloudflare or whatever storage/CDN bills, and a couple of 40$/month hetzner boxes.

Sounds like a back-then reasonable attempt at building something that can be easily scaled and carried certain risks. You simply cannot know after the fact what the basis of the decision and approach to risk was, what the agreeemnts and expectations were of and with the "anonymous investor", etc.

"Hindsight is always perfect".

Embrace mistakes? They will learn from it, and I'm already looking forward to their new project(s) (and new mistakes)!

Yes, of course. You are right and I don't mean to be overly dismissive of them; I applaud their philosophy and mission, and they shipped something many people loved.

Just... this comes from some personal lament. Computers are so wonderful. In the time between a photon emerging from your ceiling lamp and hitting the floor, you can crank out 20-30 CPU cycles.

Yet an entire breed of developers has been raised to see 2000$/month Postgres RDS on AWS as "the cost of doing business". I wish costs wouldn't hold back great things like this that the internet needs.

The "web scale" brain rot runs deep. Every project I've done has been PHP, MySQL and JavaScript. I've built software solutions for a number of local businesses that are still running years on virtually unchanged because I just used simple tools, simply.

No, none of my stuff has ever hit millions of users, and I suspect if it tried it would blow apart quite quickly, but none of them ever will come to 1/10th of that, and hot take: that's fine.

Ive been working in healthcare for a few years now and yeah, most of the stacks I encounter are similarly simple and get by fine while generating millions in revenue.

Reply on news.ycombinator.com