Topics include: programming, Apple, Unix, gadgets, large-scale web sites and other nerdy stuff.

Yegge on Agile Programming and Working at Google

2006-09-28 11:37:52

Steve Yegge has a great post about Agile Development and what a load of crap it is. It then kind of shifts into "what it is like to work at Google." It's kind of an odd segway, but that's what makes Yegge so much fun to read.

It almost makes me want to go an apply for a job there, but nothing is worth the commute to Cambridge. Nothing.

Update: thinking about it (the "Google Way" of motivating engineers), it seems to me that it is a perfectly plausible and interesting way of developing software. However, for anyone to actually replicate that process would be quite difficult. Essentially, you need to start off with a massive, nearly limitless cash flow to fund the 95% of projects with insane incentives that will fail. Then the 5% that succeed with 100 fold return on investment make up for it.

It sounds like they have created a startup incubator without all the hassle of pitching to VCs, and all the other headaches that come with starting a company, including the mundane stuff like finding office space, buying bandwidth and computers, getting data center facilities, etc.

The other part of it you need is the initial talent magnet. For Google, it is their smash-hit products (search, gmail, maps) and the reputation/fame of the founding engineering team. I'd say those are their only really big, industry dominating hits.

Only a few companies in the world even have the chance to pull off a process like this. Microsoft certainly could do it: they have the cash, and they have the talent, for now at least. The only thing working against them is that there are an awful lot of smart programmers who would never want to work on Windows, and you probably would need to do that if you worked for Microsoft.

It would an interesting private equity play. Suppose you were Blackstone, and had $4 billion (or $15 billion) lying around. Could you buy some successful startup, with a great team, and Google-ize it? Basically simulate the crazy cash flow that Google has to build up a meta-incubator tied by some common theme. The problem is that there would be a clock: unless you get a smash hit to replace the simulated cash cow in a few years, the money would run out and you'd have to shut it down.

Yahoo! could probably pull it off, if they wanted to. The big problem there is that there net income is only in the $100-200 million range, and their R&D expense is already around $200 million (quarterly). There's just not much room to say "free food for everybody." I think the main thing would be that the Hollywood types that run Yahoo! would need to start recognizing programming talent the way they recognize Hollywood talent. Seems unlikely.