Billion Dollar Startup Ideas

1. A better Google

Say I’m looking for a way to learn html. I google learn html.

My search results include a bunch of links that may or may not be good. This isn’t what I want. I just want the best one (or maybe a few of the top ones to choose from).

I think that for common things, like “learn html”, you should be told what the best stuff is. For example, w3schools, codeacademy and treehouse are all pretty high quality material and should probably be given to you as the tier 1 content. On the other hand, html.net and wikihow articles are not good content. When I first got started learning html, I spent way too much time on html.net, and other shitty sites. I should have been directed to the best options immediately.

Basically, there are 3 ways to provide content to people when they query for something: algorithmically, socially, or editorially.

• Google uses algorithms to show you what you query for.
• You could show people the sites that have gotten the most upvotes.
• Or you could hire people with good taste to filter the good stuff from the bad stuff for you. This is basically what I’m proposing. It’d take a lot of resources to do this, but the upside is tremendous.

Part of this “better google” would include search queries, but another part would be browsing for content using nested categories. This is what allows you to currate the content (you can’t curate the results to each search query, but you can curate the content in a certain category). Also, I think that this type of browsing is underrated from a UI perspective. I definitely acknowledge that there’s a place for search queries, but a lot of time I think that it’s helpful to be guided to what you’re looking for by clicking through categories.

The way I envision this is sort of like “consumer reports for the web”. You would be shown the “tier 1 results”, and there’d also be tier 2 and tier 3 results for you if you’d like. When applicable, there’d be short summaries of the strengths and weaknesses of each of the results (A.com is good at this, but B.com is good for that).

2. Better Health Care

The allocation of resources in health care seems very inefficient to me. Simple things like check ups and stuff could be automated, or given to junior practitioners. Allocating resources this way would allow for health care to be much less expensive. You don’t need 12 years of schooling to x-ray and splint a broken finger. If you suspect some sort of complication or difficulty, you could always go to someone more well trained.

The “user experience” of seeing doctors sucks. For one, they’re always late. In no other profession could you consistently be 30 minutes late for appointments and get away with it. The bad “user experience” would drive people to leave.

Doctors are also often very brief. A lot of the time I don’t know what they’re doing, why they’re doing it, what’s wrong with me, what the alternative hypotheses of what can be wrong with me are, what symptoms I should expect, how long until I recover, what all the treatment options are and their corresponding up/downsides etc. I’m not saying that it’s an efficient allocation of resources to have doctors explain all of this. It probably isn’t. I’m just saying that these are all unmet needs, and that unmet needs are opportunities to do better and make a profit. I think these needs can be met pretty comprehensively in a quick, practical and cost-effective way.

As far as implementing a better system, I’m not sure how it’d be done. One difficulty is that you’d need people with the appropriate training, but that’s difficult given how college and medical schools work. Maybe you could start by using nurses or EMTs do to routine things cheaper than a doctor will, along with doing the things that will provide a good “user experience”. Or maybe you could invest a lot of money, and train people appropriately. However this gets implemented, I think it’d involve some pretty big upfront costs. Still though, I want to emphasize that although the upfront costs would be pretty huge, I think the payoffs would be many many times more huge.

*A sort of sister problem is treatments and drug companies. Solving this is probably pretty straightforward, but would just involve huge upfront costs, and figuring out governmental regulations.

3. A Cheap and Healthy Fast Food Place

The title says it all; it’s very straightforward. Cheap. Healthy. Fast. Convenient. Taste pretty decent. I think this will be one of those “duh” ideas when it’s implemented.

What would it serve? I don’t know. There are tons of quick, cheap, healthy and decent tasting options out there. For example, pastas with variations of frozen vegetables and spices/sauces. Simple smoothies with frozen fruit and frozen vegetables. Etc.

4. A Reliable Retail Store

This is sort of a physical version of “A Better Google”.

One time I went to Rite Aid to get some pencils. There were all these random brands that all seemed the same, so I just picked one. Later on I realized that the erasers on the pencils I bought completely suck! Stores shouldn’t sell items like this.

Picking out the right product is difficult. A store should make this as easy as possible. One aspect of doing this is just researching all the options and selling only the best ones in their store. And stores should only give you one/few options, because too much choice is bad.

Another aspect of making it easy to choose the right brand is navigation and browsing in the store. Often times stores are big and complicated to navigate, and you have to walk around a lot and bump into people. I think a good solution to this would for stores to have tablets and/or touch screens you could use. Maybe have touch screens located throughout the store so you could search and browse for items you want rather than walking around to find it. And next to the items, there could be tablets with everything you’d want to know about the products. Reliable ratings and analysis, customer reviews, demonstrations etc. (Note: this is sort of an independent idea in itself.)

5. A Better Startup Accelerator

a) A good startup accelerator should handle all the little things for a founder, so he could work on his product. The founders have a comparative advantage in building, not cooking and cleaning.

What do I mean? What should it provide?

• A place to live (no cleaning, laundry, searching for an apartment, dealing with landlords, dealing with bills, commuting etc.).
• Food (no cooking, shopping, dishes etc.).
• + things I didn’t think of

I’m thinking that the best way to do that is something like a college dorm.

In addition to “living stuff”, a good accelerator should obviously provide “startup stuff too”. Legal stuff, introductions to investors, advice, dinners, alumni network etc.

b) A good startup accelerator should select for good ideas. This is obvious, but true, and I don’t think people do a good job of analyzing ideas right now. See http://lesswrong.com/lw/bc3/sotw_be_specific/. I don’t know how to prove it, but I think if you went through peoples thinking about ideas, you’d find a lot of rationality errors, and I think that if you did a better job with rationality, you could do a much better job of analyzing ideas.

c) The accelerator would fund startups according to a different structure.

Founders have high elasticity of utility with respect to income. The marginal value of money drops off a lot more at tens of millions of dollars for founders than it does for investors. Therefore, investments should be structured accordingly to take advantage of this.

My idea is that the investor would get like 90% of money past $50 million. Making $50 million is more than enough of an upside for a founder I think. (There’s a lot of details I didn’t go into with this structure because they aren’t the point. The point is the structure. It gives investors most of the “really big upside”, and gives founders “big upside” + all the resources they need to succeed.)

6. A Better Social Network

I’m pretty unsatisfied with social networks. It feels like it’s all pictures of people partying, and buzzfeed articles. The quality of conversation/interaction is very low.

Social networks also don’t do a good job of helping you meet new and cool people.

I’m not sure how to satisfy these demands, but I think that if they were satisfied, there’d be a huge reward for doing so. And they don’t seem hard enough to figure out where the reward doesn’t justify trying to figure it out.

The idea I have is to make it all about reputation. If people upvote your posts and comments and you have a good reputation, your future posts will be shown prominently in other peoples’ feeds. If you don’t have a good reputation, they won’t be shown prominently in other peoples’ feeds. There’s obviously lots of other things to account for, but the idea is that reputation would be the gist of how to get better quality interaction.

Another idea is to foster better relationships. One idea I sorta like is to delete friendships if the two people don’t interact with each other. That way you could get a good sense for who the persons friends actually are, and deliver them content accordingly. Another idea is to really emphasize/incentivize creating a thorough profile. Users would really talk about who they are, what they like, what they do, and what they think. With that information + accurate information about who your friends are + reputation, you could probably do a better job of letting people meet each other. And you could even iterate from there into a dating website…

7. A Better Sports Website

Sports websites, like espn.com, suck right now. The analysis and discussion isn’t very good, the features aren’t very comprehensive, and the UX isn’t very good.

Here’s a hint of what I mean with features and UX. Player profile pages should have a quick analysis of a player (like this), a more in depth analysis (like this), and some video analysis and highlights (like this). And a quick estimate of how good he is (6th best power forward, 30th best overall).

As far as analysis and articles go, I think that there’s a pretty strong correlation between how good of a predictor you are, and how good of an analyst you are. The stuff I read from analysts seems terrible, and I often feel that I know more than them and could prove this with by pairing our predictions head to head. For example, I thought there were a lot of stupid things in nbarank, and I think I could have done a better job than them.

So here’s my idea: let anyone sign up to make various predictions => if you do a good job predicting, you could get money + a chance to write articles that will be “lightly featured”. If your “lightly featured” articles are upvoted and liked, they’ll be more heavily featured. You could feature content based on how well the writer is liked and based on how good of a predictor the writer is. Something along those lines. The idea is that this will lead to much better analysis, because the analysts will be proven predictors.

8. Better Education

See this. Basically, I think that education is much better if it’s self-paced, and I think that the instructional material could be much much better than it is.

I plan on fixing education to make the world a better place (not to make money off of it; money to me is just a means for me to do this), but I include it because you could make a big profit off of it along the way if you wanted to.