Comment by tyre | Hacker News Reader

Comment by tyre | Hacker News Reader

No, but you can frame it in a way where he was correct.

Look at some of the largest tech companies and what languages they used to get big:

  - Facebook: PHP
  - Twitter: Ruby
  - Dropbox: Python
  - Box: PHP
  - Github: Ruby
  - AirBnB: Ruby

Probably not what PG would have had in mind.

If you move away from his objective claim about Lisp's "power" to a subjective metric based on "the right tool for the job", then sure. (Though reading e.g. Hackers and Painters, his argument around Lisp was specific to the power of that language.)

PHP perfectly fits the "move fast and break things" mantra of early Facebook. Ruby allowed for quick iteration at early Twitter.

Note that Facebook built a PHP-to-C++ compiler called HipHop, while Twitter switched largely to Scala. Box (to the best of my knowledge) has supplemented PHP with Java and Scala. Github maintained a custom fork of Rails for a long time to achieve the performance it needed.

So no, it doesn't appear that Lisp-flavoured companies have an advantage.

All good points.

Even though it's hard to find any flaw in "Use the right tool for the job", I still think that the success of a start up in Silicon Valley (and probably the rest of the world) has a lot less to do with the technical choices of the company than the people who work there. And luck. A lot of luck.

It's still baffling to me that pg doesn't realize how lucky he was. He was in the right place at the right time and he was probably surrounded by competent technical and non technical people. If you have such a team, the technical preferences of the engineers will probably have very little bearing on the success of the company as long as they use the tools they are proficient with and which allow them to execute quickly.

I mean, the fact that Rails and PHP were at the origin of Twitter and Facebook alone shows that the quality of programming languages is irrelevant for the early year(s) of a start up.

where are the assembler startups if more powerful languages don't give benefits?

Where did I say that more powerful languages don't give benefits?

Besides, assembly fails the "allows to execute quickly" test.

Ruby has very similar expressive power to LISP...

I'm pretty sure that what PG was trying to say was that given two languages X and Y that if X is more powerful (expressive) than Y it will be better for startups.

Not that if you don't use X there is no way your startup will be successful, and certainly not that if you don't use LISP you won't be successful.

It's true that you can frame it that way, but I think it's clear that this wasn't the spirit of what he was saying, and is certainly a far less controversial message ("use the right tool for the job" is hardly groundbreaking!)

I always thought he was speaking largely from survivor bias - "we succeeded, and we did X, therefore X must have contributed to our success."

Yes absolutely. For the record, PG was not saying to use the right tool for the job. Or do, but that tool is Lisp.

My comment was couching against someone coming in later with a "well he really meant the right tool for the job."

This is base rate neglect: the percentage of people who are using elite tier languages (LISP, etc) are tiny in comparison to the number of people slinging Ruby/Python/etc. All that needs to be shown is that, per capita, LISP users tend to do better. Unfortunately, I don't have such data at my fingertips :)

This is very useful, thanks. I'll concentrate on learning Ruby.

Reply on news.ycombinator.com