Ask HN: How many RFSs have been released? | Hacker News Reader

Ask HN: How many RFSs have been released?

23 pointsboggles10 comments

I read the first two RFSs that pg released but haven't seen any more since then? Are there others? Is there a list somewhere?

View on HN

I find these lists of "RFSs" and the earlier list of ideas rather disappointing... I don't think that formulating a problem in such general terms is really useful and I don't think it works at all to create such a "framework" for an idea which someone just comes and fills in with the concrete details. It really reminds me of academic grant applications where there is some grand vision laid out for 5 years, in which only the details need to be worked out. However, the details are all that matter, without the "how to do it" I don't think such ideas add anything valuable.

The line "Ex-Googlers would be particularly well suited to this project" is again the kind of credential-based test that large organizations and the government would like. This all starts to sound much less than the startup-culture that PG talks about in his essays. In this context, using a bit more informal language than in grant proposals and job adverts is slowly becoming a bit gimmicky.

I agree with your point that credentialism is counter to the notion of what we've been told YC is about. Yet, pg has always said that they want to see indicators of past performance and success, and working at Google is a pretty darn good one (usually).

Edit: I am curious about pg's thoughts on the balance between relying on strong signals from applicants (e.g. being a serial entrepreneur, working at Google, getting a PhD) to avoid false negatives, versus taking a risk and accepting teams that may be more likely to fail (false positives).

Regarding the efficacy of the RFSs: they've done the market research for you. If the proposal can eliminate a few free variables in your analysis, then you've only got to focus on the ones that remain undetermined. Furthermore, we've found that having someone pose a well-asked question can be enough to get us answering that question in a new way, and driven to build what we envision.

Imagine someone interested in data mining and such but not in working at Google, for reasons that are outlined in PGs essays (inability to get things one, large corporation, conformism, politics, etc.). YC were even running these ads in the beginning: http://www.flickr.com/photos/martindavidsson/96160482/

So when you are 100% sold on the idea that a startup is the way to go instead of accepting that offer from big co., it turns out that they look for this as validation. I find this quite sad.

About the market research part, you are right, but I think in the case of really transforming and original ideas, it is exactly the kind of approach that does not work: fix some variables, focus only on the few left. If it were so, we could write programs that come up with ideas.

> pg has always said that they want to see indicators of past performance and success,

If you've been successful in the past then that's great but then you are already part of a very select group of people. Reminds me of the ads for employees: has to be between 18 and 25, have a degree and at lesat 5 years of work experience relevant to the field.

YC attracts young folks, willing to move (that means no current large responsibilities) and willing to neglect their social life for a period in order to take this chance.

People with a success in their back pockets are not going to intersect in large numbers with people willing to meet those conditions.

> and working at Google is a pretty darn good one (usually).

Why ?

That only says that you were good enough to get hired by google but it does not mean at all that you are 'start-up founder material', rather the opposite, it says that you are the kind of person that seeks large brand name companies as employers.

That's where the failed 'starter-uppers' go when they've had their shot and need to recover for a bit, not where you go before you decide that you want to run your own company.

That's the sort of thing that drives your whole being, pretty hard to mix that attitude with having a day job, that must mean that you are not very comfortable in your day job.

The list is a good one though, (I assume we're talking about this one: http://ycombinator.com/ideas.html ), even if some of them are quite ambitious for a YC funded startup, after all, the (initial) funds are quite limited, and to get to a working prototype for some of those ideas in such a short period is stretching things quite far.

Would you honestly consider working at Google to be ceteris paribus equal to working at merely a "typical" software company? There is information there. Refer to what I said about false negatives versus false positives.

To reverse that, everything else being equal do you really think it matters if somebody has worked at google on some project vs someone working at say Yahoo, Microsoft, NASA or any one of a hundred other large operations writing software ?

I don't buy that. There are plenty of places where you can get excellent experience relevant to solving some of those problems, not all of them are 'google'. And not all of them are big company efforts either. (though the samples listed here are).

Google is a 'great' company, but they're definitely not the only game in town and I find it strange that having worked at google would be a 'pre' compared to having worked at other places.

It's not like every coder/analyst at google has personal access to the whole GFS system and can work with it to their hearts content. I figure plenty of the jobs there are like every other job in industry.

If you're part of the core search team that changes things, but having 'worked at google' is way too arbitrary a selection criterion.

So, you're saying that Google, as an employer, is a member of an equivalence class of other worthwhile employers. And, knowing the name and reputation of an employer is not enough, we also need to know what the person worked on while there. I think everyone agrees with that :)

The list is located here: http://ycombinator.com/rfs.html

There was also a list of startup ideas that YC likes posted earlier which comes with less description of the final product: http://ycombinator.com/ideas.html

I just checked the site. Doesn't Amazon already do what RFS 2 is suggesting?

They are, though in a pretty limited fashion, and chang eis at giant company speed.

It's also not their core competency, which has always been fulfillment, despite what they may wish.

Reply on news.ycombinator.com