Hacker News new | comments | ask | show | jobs | submit login
Ask PG: Will you post an update to "Startup Ideas We'd Like to Fund"?
231 points by awwstn on July 9, 2013 | hide | past | web | favorite | 72 comments
Here's a link to the original one from 2008: http://ycombinator.com/ideas.html

Much of it still seems relevant and proved over the ensuing 5 years to be quite prescient. I'd love to hear what new ideas/themes have emerged in the time since this was written.

I think some people tend to take pg's startup ideas too seriously. I mean, I've heard talks by him in which he explicitly says not to take these too seriously. They are examples and only examples. So right now, for example, it's pretty obvious security is a big thing. It doesn't (shouldn't?) take pg to tell you that. All you have to do is read the news.

Find a big problem (it just happens that dating is one) and fix it. Don't plan on getting in YC (or any other incubator) or being successful simply because you picked something off one of pg's lists. I'd say the most important qualities of an entrepreneur is to be in tune with these problems without getting any sort of external guidance. It's not a good thing when someone comes on HN and posts a post-mortem about their YC experience beginning by citing an RFS or one of pg's "idea" essays. Personally, I think that some RFSs from 2009 (http://ycombinator.com/rfs.html) are downright silly -- 5 and 3 in particular. Moral of the story: find something you're passionate about -- a problem, a niche, etc. -- and build something to fix it. If it happens to fit into one of the RFSs or pg's ideas, oh well, good for you I guess.

I totally agree with most of this. I'm actually not really looking to start a company or apply to YC in the near future, so I mostly posted this because I found it interesting how spot on he was 5 years ago and I'm curious what his more recent thoughts are on this. Also, I'd like to know if my own predictions and thoughts are aligned with those of the guy who's gotten it right so many times before.

Trust me, I wasn't planning to request a list of startup ideas from him so that I could pick one of them out of a hat and apply to YC with it. :-)

In light of a couple of recent stories on HN, including the one on the passing of Doug Engelbart, I'll offer this up as a source of ideas for startups:

Everthing old is new again... and some ideas that seem old haven't actually been completed yet. Other ideas seem to fail, but only because they are "before their time".

With that in mind, I'd suggest going back and reading the writings of people like Doug Enbelbart, Ted Nelson and other computing pioneers (and not just computing pioneers, really, but any great thinkers. How about Nicola Tesla, for example?) and look for places where they proposed amazing things years or decades ago, that still don't exist, or don't exist as fully as they could. On that same basic note, go back in time and re-read some old issues of Infoworld, Computer World, Information Week, Business 2.0, Red Herring, Fast Company or Wired from the late 90's or early 2000's and mine for "before their time" ideas that might be ripe for a second shot.

Hanging around hackerspaces is also a great way to gain exposure to a constant stream of interesting ideas and approaches.

Excellent points, particularly with respect to looking outside of hackerdom.

Yeah, I should do that.

I thought I had read in one of those that PG really wants to replace banks and credit cards. Or maybe it was someone else. Either way, that's one of the things I'd love to get into as far as "big picture" stuff (along with a cloud-based "space station" on venus, but the bank thing might be slightly less difficult). I'd completely revamp the entire purchase process, as much as legally possible (and realistically doing that would also require lobbying to change any outdated laws we can that aren't being protected by the other banks. You'd have to go after all the peripheral laws first I feel like, or else the banks would crush you with their influence). The entire transfer of funds and the documentation of the purchase would be re-done based on what we can actually do with this kind of data. It's insane that my bank only tells me things like "AMAZON MKTPLACE PMTS AMZN.COM/BILL WA 06/26" for an amazon purchase. Mint tries to parse that and just has no idea how to classify it. The truth is, the bank should only get sufficient information for facilitating the transfer, and there needs to be a simple system of reporting detailed purchase info directly from the store to the user. Getting this to work with many vendors would be a pain, but it doesn't help that currently I don't know anything about those systems. I have some ideas on how to kind of abstract that out...

Just responding to the specific example, I've been very happy with square / square wallet in this regard. The system recognizes my card, and offers to send me an itemized receipt via email for any transaction. I can use this to clearly see the merchant, date/time, and what I paid for. Simple (the bank) is making inroads in this as well, and even PayPal's receipt emails are more descriptive. Maybe a bit too complex of an issue for a startup to hit now (look at what happened to Dwolla) but certainly still a lot of "pain points" to solve left.

Maybe Mint should scrape Amazon and other retailers in addition to bank websites? Would you trust Mint to keep your Amazon password in addition to your bank's? (Maybe Mint should pivot slightly to include password database management? :p )

Right now, our system is pretty much how you describe it, where the financial institution only gets the basic information of who and how much. And if you want more detail, the retailer provides it.

So I guess you're looking for a service that centralizes that info into the same place as your other financial details?

(paragraphs, please)

What if retailers/banks gave you a separate account login that can only be used to read data? There wouldn't be nearly as much risk in giving this out to third parties.

What if retailers/banks gave you an API (with authentication of course) that can only be used to read data?

You know, Google reads all your email and uses it in combination with your search history to populate Google Now cards. When you buy something online, you typically get a detailed email receipt. I bet if anyone knows exactly what you buy, it would be Google. That is, if you are using Gmail.

Mint is now owned by Intuit. Are the original founders still with the project? If not, I'd be surprised to see them take such a precient direction.

I believe the founder is still head of product development for all of Intuit.

Not the case. And we are all gone. None of the original team remains.

Did you see this one from 2012 ?


From personal experience, it seems like YC may no longer have an interest in funding RFS #5. Our entire interview was spent trying to explain why someone would possibly want to code on a handheld.

Full disclosure: Our demo and Altair BASIC was a way to develop native iPhone apps in the cloud, from a tablet or other device. We wanted to expand to Android apps, and eventually be a general cloud compile/debug/run service.

Not sure that your example proves that YC is no longer interested in RFS #5. I think RFS# 5 covers companies that could be huge, but I'm not convinced that I would ever want to code on a tablet.

RFS #5 has nothing to do with the essay linked. See http://ycombinator.com/rfs.html for all the "Request for Startups" bullet points.

Personally, I think it's a dumb RFS to begin with. From a professional standpoint, I couldn't handle developing on anything else but two 24+ inch monitors (if not three) and a mechanical keyboard.

Not that I couldn't do it on a laptop or tablet, it just makes me about 300x more productive -- and that's what matters.

I don't think thats the point.

It should be possible to code on a tablet because thats the device most kids will have access to as their first computer.

My first (and most people on HN) computer I turned it on and the OS simply waited for me to give it an instruction.

That is important in lots of ways, not just in the IPO in two years ways.

There would definitely be some pains going from huge monitors to a small tablet. The bigger side of our plan was that it wouldn't just be possible to develop from a tablet, but really, from anywhere. Develop iPhone apps on your Windows box, or Chromebook, for example.

I think that in the future it might be possible that developing on a tablet might just mean hooking your tablet or other sub $300 device up to two 24+ inch monitors. Think developing iPhone apps from a $25 Raspberry Pi, thanks to the cloud :-)

>I'm not convinced that I would ever want to code on a tablet.

Well when I type on a on a 10 inch tab - two thoughts cross my head A) This is not working B) This could work given a bit more size and refinement.

So I would not discount tablets as dev devices. Perhaps not as a hardcore workstation, but as a starbucks dev station sure - soon.

You shouldn't have to use a QWERTY-like keyboard to enter text into a tablet. The very idea of a flat space with a list of letters which has you individually selecting letters (or moving your finger around on a full list of letters) is a huge waste of space on an already small screen.

>You shouldn't have to use a QWERTY-like keyboard to enter text into a tablet

Why not - If I can smoothly transfer touch typing skills then thats a pretty big win.

> huge waste of space on an already small screen.

Hence my comment about slightly larger screen. I can almost touch type on a 10 inch...bit more and it could work. The capacitative aspect might even work better than a physical keyboard for "touch" typing.

> Well when I type on a on a 10 inch tab - two thoughts cross my head A) This is not working B) This could work given a bit more size and refinement.

A bit more size and refinement, a nice keyboard, and a stand gets you a 11-12" laptop.

>gets you a 11-12" laptop.

No it doesn't. No hinge & typing space is also display space under some conditions e.g. browsing.

I'd classify those under "refinements" ;)

Its a different form factor...especially the no hinge part.

I could see it with a graphical programming language -- one where a FOR loop is a large box into which you'd drag other commands, where each variable and equation is an icon you drag out from the side. Almost like a physics flash game.

We mostly expected people to attach a bluetooth keyboard to their tablet or phone. For people who insisted on using a touch screen, we were working on a graphical context sensitive system like you describe -- except for existing languages (Objective-C)... because nobody wants to keep tapping semicolons and parenthesis :-)

Something like http://scratch.mit.edu/ ?

What if the tablet was as big as a desk? or I saw a demo of some microsoft tech where the monitor level was transparent and you could reach behind it and interact with 3d objects.

Sounds to me like they're very interested in funding RFS5 and that they were grilling you to see your response. Were you exhilarated and raving about all the advantages in a confident, persuasive manner, or did you stumble on why you might even want such a thing yourself?

It's possible. The given reason for our rejection was the market size for our product. Really, we were caught heavily off guard by having to explain why it would be useful. We thought, "It's an RFS, so they already know it could be huge!" and spent zero time thinking about it.

Full disclosure: Our demo and Altair BASIC was a way to develop native iPhone apps in the cloud. We wanted to expand to Android apps, and eventually be a general cloud compile/debug/run service.

Having to explain why something could be "huge" shows how well you understand the market.

I haven't read the 2008 list but just went there based on some of the comments here. This #29 stood out:

"What's the best way to make a web site if you're a real estate agent, or a restaurant, or a lawyer? There still don't seem to be canonical answers. "

There is a huge opportunity here not for a site builder so much as a way for those types of businesses (and other businesses) to keep their site current and fresh once it is built. None of the existing options work that well with this demographic. Either to juvenile or to many features and options and a learning curve. Fix that and the world will beat a path to your door.

Start small with this. Solve the issue of how I know what the specials are from the local sushi takeout restaurant that I visit a several times per month. Then move to other restaurants. Make it dead simple for the busy owner to get me that info. Even if it means simply shooting a picture of the special board and getting it to a single page site that you host in the cloud.

The thing with these types of people (realtors, restaurant mangers, etc.) is that you can make something extremely easy to use but there will always be effort in producing new content; effort they are not willing to expend. Time and time again I've worked with people who couldn't be bothered to enter the data for their new listing or upload photos. They always feel like it is busy work and better farmed out to an assistant or done by myself at a unsustainable cost. Maybe the problem is that most of these people are technically inept and/or can't type more than 10wpm, I don't know. I've completely given up on anything real estate related and refuse to take any projects in that space because the failure to produce new content and maintain a web property always gets reflected back on me, as if I made something too hard for them to do. Plenty of dead-simple options exist to create and maintain a site, I find the issue is lack of cooperation with the site's owner to produce new content.

Now in the food service space, you could easily auto-generate an update based on the POS inventory. The problem there is many restaurants use a system like Micros, which has data access, but refuses to open it up to third-party vendors. If Micros ever wised up and opened an app store where customers could install approved and signed apps, the person who bridges inventory to web will cash in and retire early.

"Maybe the problem is that most of these people are technically inept "

First let me say that your comments are dead on from my experience (and I'm even talking outside and before the internet dealing with small business in general as well as the people you mention.)

I don't think it's inept I think small business people are "shiny ball" driven and they react to things (deadlines, fines, customers, tax filings, supplies running out) they are used to going at their own pace and as a result have a hard time having the discipline to do things like this on any type of regular basis.

Perhaps the way to handle this is more or less some kind of mechanical turk.

If you tell the realtor "guy from India will be calling every 10am Tuesday" that creates the shiny ball that forces him (like taxes that are due) to get his house in order. Just a thought. I know when I was in a different business we did something similar and had decent results. We also did everything we could to make it easy for someone.

Obv. your time is worth more. But perhaps there is someone who could do the "teeth pulling" at a lower rate.

After all with video chat and all of that it's not a non-starter.

To tackle just this problem, we created this: http://wigwamm.co.uk - would appreciate feedback.

A restaurant get's more hits on their Yelp page than their actual website. When shit get's super fragmented it helps everyone involved to have an aggregator. It's why the App Store works so well.

It seems to me that LinkedIn would be well positioned for something like this. Allow pro users to create microsites.

I think the canonical answer to this for restaurants is Facebook page + Google Maps listing + Yelp listing.

too! Seriously, it's not that hard.

"getting it to a single page site that you host in the cloud."

you mean, like a website?

There are a million canonical answers to building a web site for a real estate agent, everything from Pinterest to a simple Wordpress with a domain centric theme.

Too me this space is about lead generation. Something like channel advisor for services.

News will morph significantly in the more competitive environment of the web. So called "blogs" (because the old media call everything published online a "blog") like PerezHilton and TechCrunch are one sign of the future. News sites like Reddit and Digg are another. But these are just the beginning.

I'll make the bold claim that we haven't yet seen this "morph" in media. The blog concept is now staid yet Buzzfeed is the most innovative publisher we have. And they are just re-branding the advertorial as "native advertising."

The only changes I've seen lately are larger and more diverse media to accompany text. Not a substantially new product as far as I can tell.

The morph happened, but it's got nothing to do with media format. It has to do with source and accuracy.

The big change is that we don't get our science from OMNI, our glimpses of new cars from popular mechanics or our poll forecasting from CNN. We get specific stories from subject matter experts who can communicate without oversimplifying into nonsense and who have a true need to deliver truth above advertising or team-affiliation concerns.

(As if they deliver nonsense, or are transparently-in-the-bag for a monied interest, they'll be quickly sidelined and ignored.)

I haven't seen the last part. Many popular blogs are delivering nonsense and not being sidelined. The blog-news industry is full of poorly researched speculation sourced mainly from Twitter and Wikipedia (not always with attribution), reblogged images from Reddit, sensationalized summaries of university press releases, top-10 lists, etc., etc.

"Buzzfeed is the most innovative publisher we have"

Strong disagree. I am not sure I would call the New Yorker the most innovative and yet buzzfeed has nothing comparable to strongbox[1]. Vice? Quartz? Medium?

Its possible that I am letting my opinion of Buzzfeed's "content" cloud my judgement of their innovation. What do you think BF is doing that is innovative?

[1] http://www.newyorker.com/strongbox/

I see Buzzfeed's innovation as largely around the continued development of the "Contributor Model" for content, which drives down the per-unit cost of content by turning to non-fulltime outsiders to help create content along side full-time staffers. Medium taps into this, as did Quora before it.

Vice is pretty innovative around web content but no new formats that I know of and a typical "stringer" model with high editorial touch. Quartz is visually different (super slick JS based webapp) but the underlying news is the same that you get elsewhere.

This is something I've been thinking about recently. If you boil down the 'news problem,' the dominant solutions are traditional news channels (TV networks, newspapers) and sites like Reddit and Digg. I would categorize bloggers as very small-scale traditional news sources, and RSS feeds as a way to combine these traditional and nontraditional sources.

The strength of traditional news is partly cultural, but also due to the nature of its content as being curated and vetted by some (hopefully) informed party prior to publication. Whatever is there, can very generally be expected to be of a higher quality than any random given 'social news' post on the same subject. The drawback, of course, is that the only news available is that which the source determines is 'newsworthy,' which is a very subjective and questionable standard.

Conversely, the strength of 'social news' sources such as Reddit is in their inherent chaos. The crowd decides what floats to the top, with no real vetting process or curation except within the most strictly moderated subcommunities. This ensures that the most socially relevant topics get 'published' to the front page and into the public eye, but those stories have little to no quality control. For those of you unfamiliar with the concept of the "Eternal September," there has been an established trend wherein as the size of a loosely controlled network grows, the quality and content increasingly caters to the lowest common denominator, driving the more invested (and usually the more intellectual) users out, and eroding the value of the network. This has been seen over and over on Reddit, as the mainstream subreddits get overrun with a colossal stream of new users, and any niche interests get sidelined and marginalized.

I would propose a hybrid approach. Why not create a news source - whatever category you decide to label it - that publishes content that is vetted for quality (bias, sensationalism, thoroughness, etc.) but not subject area; in other words, where the articles are all good, and any subject is fair game? Layered on this could be some categorized and 'social' voting algorithm, much like Reddit with subreddits, and users could subscribe to those channels which they cared most strongly for, thereby creating their own high quality news source.

This is just a thought I had today. It's obviously not fully baked yet. What do you think?

Is this what medium.com is doing?

I'm not sure. Sounds like I'll have to check it out, though, because I'd love to use a news service like that.

An idea I've been interested in seeing is a regional paywall for online media. Instead of doing it for a single site like the NY Times (who is uniquely positioned to do that), regional newspapers and TV stations could band together and have a shared paywall. The big challenge would be figuring out the revenue share.

I'm guessing that online dating will still be on the list.

it probably always will be - so long as the man / woman imbalance with dating doesnt even out!

I'm not sure what imbalance, exactly, that you mean, but I don't see anything of the sort as the reason that online dating isn't going anywhere. The fundamental problem I see is that meeting new people is hard. Particularly after school, with a full-time job, if you recently moved to where you live, and if you don't live in a city (being close helps, but it's not the same).

Part of my problem with online dating is it's too much like looking for a job. You have to fill out forms, put in all of your data again, and perfect your wording in multiple locations. Yet, you're underqualified for who you want to be with and the only offers you get are those which can't take seriously.

Im referring to the fact that on most dating sites you'll find "n" male profiles - and something like 0.5n to 0.7n female profiles. Dating for LGBT works brilliantly - and the only location based dating/hookup apps that I know of that actually took off are targetted at LGBTs. The heterosexual space on the other hand still suffers from the same old - problem as high school.

How do you know this is true? Your own experience, or data? I live in a big area, so perhaps I can't see the effect even if its there.

I run a consultancy that helps companies do stuff in big data.. one of our recent clients was one of the largest dating portals in europe :)

I don't know if there is an imbalance either, I do know that in certain circles it's culturally unaccepted. Many people I know consider online dating to be something for losers.

Interestingly, this can easily be turned around: cherry pick applicants, make it invite-only, or make it exclusive in any other way. When 50 "cool kids" are in, the rest will be lining up.

On what basis are you saying that "online dating isn't going anywhere?" One could call Facebook the most successful online dating site, one-third of married couples last year met online, and couples who meet online are less likely to divorce. That sounds pretty impressive to me!

I mean that online dating is not leaving us; online dating is here to stay; online dating is obviously a service that people like using and get something out of. I was disagreeing with the parent poster about why online dating is here to stay, not that it is here to stay.

ahhh. Idioms at work.

Perhaps some hardware ideas will be on there. It seems like all those savvy Kickstarter trinkets are the next big thing (Lockitron comes to mind).

Have any of these even been solved yet?

  Does the email replacement sound like a support ticket system to anyone else?

Not to be snarky or anything but in short: we're currently working on the new ideas. Please give us a little bit of time before you ask Mr. Graham to give us a synopsis of the most logical projects to address. Trust me, whatever he wants done: there are people working on it. "What do you want to contribute?" is the more relevant question. What are your ideas for "Startup Ideas We'd Like to Fund"? and start working on those.

An even better question to help is: Who is working on the new greatest technology, and how can I help?

Paul Graham and Steve Jobs weren't magical people... they are/were people who addressed the needs of the masses. Start looking at things that annoy you and figure out a way to fix them. That is way to better society.

A little more recent:


I like RFS #7.

http://ycombinator.com/rfs7.html (in case you'd like to know)

Not sure if you've seen it, but RFS #7 was about Facebook's instant personalization service. ycombinator.com/rfs7.html

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact