undefined | Better HN

0 pointspclmulqdq3y ago0 comments

Whatsapp, discord, telegram, and all other chat apps have a very easy time scaling. Using user count as a metric and comparing them to Twitter is pretty disingenuous. Instagram is a much better comparison, which had 1/50th the user count of Twitter at its acquisition.

Chat apps are well known to rely on infrastructure that scales up exceedingly well. That is why there are so many of them, and why they all have 100 tech employees or less.

Instagram at the time of the acquisition could have run on 20 servers plus S3. Today's Instagram, along with today's Twitter, does a lot of work that is super-linear in user count, and has something like 2000 engineers. Timeline building is reportedly O(n^2) in user count. The scale difference has a huge effect.

Notice that chat apps have no equivalent to "timeline building." The worst scale factor a chat app has is linear.

0 comments

adam_arthur3y ago

Ok, multiply instagrams employees at time of acquisition by 50x and you still get an order of magnitude smaller employee base than Twitter.

And obviously employee count shouldn’t scale linearly with user count. No idea where you get that idea?

To argue that a 1B+ user app of moderate complexity can’t be run by a few hundred employees is simply wrong.

The end game for all web software is going to be managed services with automation built in such that it requires almost no manual intervention (absent logic bugs). This gets easier and easier to accomplish as cloud capabilities grow. And cheaper.

The long term trend for tech is obviously towards fewer employees and bigger impact per employee through leveraging higher level abstractions. Look at what you can build as an individual today vs 20 years ago and it’s quite obvious. Unless you think technological progress will suddenly halt

pclmulqdqOP3y ago

You are missing that you actually pay a lot for those abstractions. Once you get to Twitter's or Instagram's scale, it starts to pay to DIY the foundations of your technology, because the people to build it are cheaper than the cost to buy it. Building also has the benefit of aligning your concerns across abstraction layers, which pays huge dividends in efficiency.

I don't think it's obvious at all that the long term trend is in favor of few employees with "high impact through abstraction" when the highest-impact people I know at big tech companies are ones who build the underlying technologies behind those abstractions. They routinely bring in 8-figure cost savings per engineer per year compared to using OTS or open-source technologies.

Headcount should scale with actual complexity of the problem being solved, which is not AT ALL linked to a user's perception of "complexity." Twitter is far more complex than you are giving it credit for, and a big part of that complexity comes from its scale and the fact that it is cheaper for them to eschew "higher level abstractions" than to use them.

j / k navigate · click thread line to collapse