As of last batch, which had 45 startups in it, we were able to give everyone the attention they needed, as measured by the results (= most observers thought w2011 was the best batch so far).

I am obviously a fan of YC. But, I would caution putting too much weight on what investors tell you or even their investment activity. At this point you are such a big deal, there is incentive to tell you whatever you want to hear.

Interestingly, much like investment portfolios (and most other things), investor opinion quality falls under some distribution. The opinions of a very small number of very no-nonsense, intelligent ones is probably the best indicator of batch quality. The rest matter very little.

On a related note, I think the distribution curve for YC batches probably shifts a minute amount to the right each batch but mostly fattens. Then, the range of investor opinions coupled with a growing positive bias creates what seems like a uniform larger shift to the right. Just speculation though.

Some of this may have to do with the quality of startups going in to the batch. As the reputation of YC rises, more and more good-quality startups want to join (rather than go on their own?). This of course doesn't take away the fact that YC has done a great job mentoring and helping early stage startups.

This makes me think (not totally seriously, though) that a YC prep-school for startups might get a lot of attention. There seems to be some room for startups at an even earlier stage than YC would accept now.

That's what early stage pre-incubators like Founder Labs are aiming at.

How can you judge a class so quickly? I think it's impossible to evaluate a class until 3+ years after YC.

Depends what you're judging. It's true you can't judge returns till much later, because those usually depend on a single outlier, and whether they turn down the $50m offer.

The sort of test I'm talking about is a preliminary one: how promising the companies look on Demo Day. Investors who've been to a lot of past Demo Days (and there have now been 12) have a pretty good sense of what percentage of the companies they want to talk to further.

Do you feel close to the limit you can handle? You've said several times you're not there yet, but maybe you have a feeling about it.

60 should be ok, because we have 3 partners advising startups about product stuff now (me, Harj, and Paul Buchheit) and I know 1 partner can easily handle 20 startups. We probably could not get up to 80 without more partners though.

I still wonder, at what point, would other variables come into play, and what those variables might be? Let's say you get a couple more partners, then would you take 100 companies?

The point is similar to this one: we often have long wait times for specialist doctor's appointments, and there are many young people wanting to become doctors, then why can't we just train more specialists?

The American Medical Association and state governments tightly restrict the market supply of physicians. This keeps physicians salaries high in the name of safety and quality. However, the market is prevented from trying different models.

That sounds reasonable, but I'm curious if you have data on satisfaction as reported by the founders? As your reputation grows, you may be getting better results partially through having higher quality startups, which would cloud the issue.

When the food served every Tuesday is the biggest complaint, I think you're running a pretty tight ship. There was no shortage of attention for those who needed it.

The last couple batches I've started asking each batch at the last dinner what we need to improve. As of last batch there were no major gripes about not getting enough attention from us.

I'm curious. What are the major gripes?

The biggest were about Preview Day, where we introduce the startups early to a few top investors. We did it too late in the cycle, didn't give startups long enough to present, and some of the investors were lame. We're not sure if we'll even keep doing it.

There were also complaints about the food. It's getting better but it's still not that great.

It's getting crowded in the room after Demo Day. That will be partly alleviated by a 2x bigger room this summer, and we'll also probably write some software to let investors register their interest in a startup instead of assuming they'll connect afterward.

Some of the speakers were boring. There are always a few.

I'm sure you guys have enough time to dedicate to the companies, but it does lose a little of the "exclusivity" for the companies themselves. It used to be I could name every company that came out of a YCombinator class, now it's likely a bunch will go through that I've never heard of.

This would also be explained by a higher quality applicant pool

Yes, the applicants do get slightly better each batch. But not as fast as the batches grow. So if we were going to get hosed by growth, odds are a single batch's increase in quality would not be enough to conceal it.

Obvious retort: YC is not just about the attention of PG and Friends. The summation of prestige, money in the present and, likely, in the future, and YC network alone still put a company in a different world.

Question: Will YC pick people who need heaps of attention just to succeed? (My) Answer: Probably not. YC people are called "relentlessly resourceful" for a reason. They will succeed somehow with or without hours of coaching and mentoring by the YC folks. This probably translates into less office time than you would expect from 60+ companies.

But I've never been through it, so I don't really know the answer. There are some brilliant people I've met who couldn't do the simplest of things. What little attention a group needs in one area might be made up for in another.

edit: a sentence had no subject.

But they don't have to compete - each startup can all have all the face time they want.

"There's no limit to the number of office hours they can book; startups talk to us as often as they want... The office hour software is designed not just to book slots, but to show us if startups are getting enough time with us. We set aside blocks of time for office hours, and if not all the slots get taken, we know demand is satisfied. (If all the slots get taken, we set aside more time.)" http://ycombinator.com/atyc.html

Actually we've improved the system since then. Now startups that want to talk can also request office hours, and people who are in the queue get slots first whenever we create a new block. I should add something about that.

I think its just a mark of the times. Top talent is focusing on creating new ventures. I think if the quality of businesses is on par with the last batches – and I am sure it is, if not higher – it's not a problem. I assume the talent level is equal to previous years so if the teams submitted their ideas two years ago they would have been accepted. I would focus less on the number and more on the increasing success of Y Combinator.