So, Amazon bought Whole Foods. In other words, Amazon instantly grew its grocery supply chain by something like 100x.

Americans still overwhelmingly shop for groceries at the store, but this could be eaten away by integrations like Alexa. Soon, the staples from your Amazon shopping list will be ready when you walk into Whole Foods, and you can pick out the last few things, like fresh produce, yourself. 

Once consumers are used to starting their grocery experience online, small food creators like Justin’s get a new playbook. In today’s grocery stores, the way you sell is by getting distribution, a.k.a. shelf space. But shelf space is unlimited online. And so a new model makes sense, based on superior experience and superior SEO. Expect reviews and word of mouth to matter far more than they do today. There’s never been a better time to create a niche food product.

Lastly, I wouldn’t be surprised to see Amazon Prime discounts for grocery items as a way to encourage loyalty. Perhaps the discounts are only available for people who start their orders online. 

I wish I had an expert to help guide me through the complicated healthcare industry. This expert would know which doctor I should visit, has my medical history, and is an advocate for my wellbeing and wallet.

This expert would work with every type of insurance, every doctor, every state.

When I get new health insurance this expert would make a handy list of nearby doctors that take my insurance and collect their yelp reviews, estimated cost, and other information to help me make a great decision.

When I need to schedule an appointment, this expert would do it for me, calling the office during business hours and finding something that works with my schedule.

When my doctor asks you about the immunizations I got a decade ago, the expert will help me find it by calling previous doctors.

When I get referred to a specialist, this expert will show me their reviews and cost, and even offer to show me other nearby specialists that I might prefer instead. If I choose one of them, this expert would handle the introductions and scheduling.

When I get back test results and my doctor gives me a cryptic 10 second explanation, this expert would be there to help explain. And when I’m long gone from the doctor’s office, this expert would still be there to fill in the details.

When I need to pick up prescriptions, this expert could show me the best nearby pharmacies if I haven’t chosen one yet. And if I have, this expert would show me exactly how many refills I have left and help me schedule an appointment when I run out.

If I break a bone and need surgery, this expert could show me hospitals and surgeons that are top of their field. If I’m willing to drive or fly to another city for care, this expert would expand the search across the state and country.

All in all, sounds pretty great. As far as I can tell, this expert doesn’t exist yet, but I really wish he or she did.

Personal transit–car, lyft, uber–is valuable. It picks you up where you are and drops you off where you want to go, no walking or transfers required. It’s great for people with mobility challenges, and can change bit by bit as people move, get jobs, and switch their transportation needs.

However, personal transit doesn’t solve congestion. At a certain point, to get more people to a desirable destination or through a busy corridor requires something with more capacity. Like a bus, or a train, mass transit. The problem is, all other things equal, mass transit is less convenient than personal transit. The way to address this is to ensure that mass transit is faster than personal transit for congested routes.

The corollary is that mass transit shouldn’t exist if it isn’t faster than individual transit along the same route. In practice, this means fewer lines, dedicated lanes, underground rail, and infrequent stops. This provides the incentive to use this more efficient type of transit.

Unfortunately, mass transit is often seen as the option that should only be chosen when there is no other option. For people too poor to afford something else, and for accessibility needs like wheelchair access. This is the wrong way to think about transit.

It’s crazy to think that someone who is blind, uses a walker, or has other mobility challenges should be forced to take mass transit, the type of transportation that should be designed with speed and efficiency in mind. For this person, personal transit is a great fit, and anyways is probably both faster and easier than mass transit when the walk to the stop is factored in.

A city’s responsibility to provide equitable access to transportation can be fulfilled in other ways. Require every personal transit operator to offer a minimum quality of service for people with mobility challenges so that someone with a wheelchair can get a ride at their door in less than 15 minutes. Subsidize both personal and mass transportation for low-income people. And use mass transit for its strength, adding more capacity to an already congested route.

30 minutes and 1 transfer. Beyond that, a transit experience degrades quickly. Some big cities with great transit still fail this test, partially due to their sheer size. For example, Beijing has a great subway system but can still feel remarkably inaccessible because the city is so large that many places worth going still take more than 30 minutes and 1 transfer. San Francisco, by comparison, has an almost non-existent subway, yet feels more accessible because of the natural constraints of the ocean and bay.

But even San Francisco isn’t consistently easy to access. Compare these two maps to see how long it takes to get places by transit. Living in the center of the city means it takes at most 45 minutes to get almost anywhere. Living in the northwest means an hour or more for large and interesting parts of the city like the mission & dogpatch. 

Michael Healy recently published a book about the history of San Francisco’s bay area rapid transit system (BART).

Some highlights:

• Even in the 1960s, the cost of BART seemed crazy. On top of that, it used new, unproven technology. The cost reminds me of today’s mega-projects like the California high-speed rail.

• BART required some complicated political maneuverers to get off the ground. For example, the BART board got the California legislature to pass a measure that reduced the required voter support from 66% to 60%. BART didn’t have enough voters for 66%. This reminds me of the soda tax in SF didn’t pass until it was reworked to require only 50% rather than 66% of the vote.

• Once the initial stake was driven, nothing could kill the project. Even though it eventually required far more funding, it still got finished.

• After the initial project was completed, it lacked any coherent vision other than “operate, and maybe grow some day”. Consequently, nothing much happened to BART over the past 40 years. Contrast this with San Francisco’s Muni which keeps expanding at a slow but steady rate.

Highways get built by selling a vision of faster travel. But for most places, congestion will only get worse, and the question becomes how to make the best of congested roads. 

Beeline is an experiment in Singapore that creates new bus routes based on the actual trips that people are taking, rather than an abstract measure like population or job density. Beeline is designed with congestion in mind by promising a commute that is equivalently fast (or more accurately, equivalently slow) and at a similar price point as driving, but without the extra hassle of owning and maintaining a car.

The overall goal is to get people who are currently driving to consider higher-density alternatives that help the city as a whole have more trips with the same road capacity. 

Compare this to Uber or Lyft, where congestion often leads to surge pricing. These are best at off-peak times when they’re clearly faster or cheaper or both. They’re the worst at highly congested times, where infrastructure that gets better at rush hour will win any day.

 

Which compromises are necessary, and which go too far?

Consider BART, the San Francisco Bay Area’s rapid transit system. Started in the 1960s, It took 15 years before the first trains started running, and the path was full of compromises.

• Many stations were moved based on community feedback

• San Francisco got an entirely new station late in the process

• Berkeley forced BART to change plans and build its three stations underground

• Overhead handles were added once it became clear that the dream of only having seated passengers was unworkable

• Opening day was pushed back again, and again, and again

• San Mateo county dropped out, leaving a large funding gap

• The BART directors decided to kick Marin county over fears of the funding measure not passing the popular vote

• Service started with only a handful of 2 and 3 car trains because of supplier issues (trains now run as many as 10 cars)

But even more interesting are the commitments that weren’t compromised.

• The system was going to get built, whatever the cost. This became especially painful as unexpectedly high inflation ate away at the original funding measures.

• Automation of the trains. Rather than kill the dream of automation, BART decided to run trains with a mandatory 2 station gap while problems with the automated control system were solved.

• The tunnel underneath the bay, a vital interconnect between west and east bay.

There were many opportunities to compromise on these commitments, but to do so would be to fundamentally change the project. Moreover, from the perspective of 40 years later, many of the original compromises seem unimportant. Useful to remember for anyone trying to get a project out the door today.

San Francisco’s Muni has a system that lets people know when the next bus is coming. Unfortunately it’s full of gaps and difficulties that make it hard to use in practice. The good news is that Muni is designing a new system. Here’s how I’d redesign it:

###Why have real time information?

The point of real time displays is so that people can gather and understand information faster and better. The 4 key times & places when people need information are:

• At a stop (how long will I have to wait?)

• Before leaving home/work/etc… (when should I leave home so I don’t need to wait outside?)

• Making a transportation decision (how long will it take by bus compared to by Lyft?)

• On a bus or a train (how long until I need to get off?)

An effective system should have the following:

###A display at every single stop

Every stop needs a display, not just major stops. No display, no stop. This is an essential part of a modern transit experience. This display should display wait times for every line that serves that stop. This information should fit on a single display so that it doesn’t have to rotate. (helps #1)

In locations with fare gates, an additional display should be placed before the fare gate so that riders can make an informed decision (helps #3).

For underground or elevated locations there should be an additional display placed at the entrance to the station so that people can avoid rushing when it’s not their train that’s arriving. (helps #2)

Here’s what this display could look like:

###A display in every vehicle

Buses and trains need a display that helps people solve #4. This display is different from the display at each stop. It needs to show the upcoming stops and time estimates for each stop. Just like the display at each stop, the information should all fit onscreen at the same time and not require rotation. 

Here’s what this onboard display could look like:

###An open API

A dedicated Muni app is not the correct place to show transit information because Muni is only one small part of a larger transportation system. Apps like Google Maps and Lyft are better places to show transit information because they can solve #2 and #3 better than a standalone app can.

An open API will allow these integrations to flourish and will provide the best experience for potential Muni riders. This API needs to have both real time predictions, but also crucially the real time and historical locations of every vehicle in the system so that 3rd parties can develop better algorithms for prediction that incorporate additional sources of data such as traffic. Allowing for additional sources of predictions will also result in better validation of Muni’s own predictions.

Ideally this API has centimeter level and minute-level precision.

###No additional services

Services like the standalone Muni app, nextmuni.com, SMS, and phone predictions should be discontinued. These are better provided by 3rd parties using platforms like Google Maps and Lyft. Any cost savings should be redirected into the addition of displays at every stop and on every train. 

Right now, almost every home has a kitchen, even in the city where other options for food are plentiful. Consider what life would look like without the kitchen:

• Homes cost less to build and renovate. For example, renovating a kitchen in the US can cost $10,000 to $100,000. 

• Every meal is eaten “out”, which means getting outside more often.

• The time spent cooking can be used for other things, like spending time with the kids.

• Going out of town becomes easier because there’s no food left at home to spoil.

• Grocery stores become less of a thing, clearing space for homes and businesses.

• Some of the most common causes of fire are eliminated.

• Cafes, food trucks, food carts, and other inexpensive options open up to serve people who are eating all of their meals out. 

• 10% of a home’s space is a kitchen (probably more in cities), which can be used for other things, including entirely new homes, restaurants, public space, or parks in the long run. 

• Jobs are created to serve all of the extra meals needed, perfect for people who love to cook. 

• New types of businesses become possible, such as kitchens that can be rented out for events

Getting rid of kitchens isn’t really feasible when the nearest restaurant or cafe is a 30 minute walk, but it makes a lot more sense when considering how to use the very limited space available in our biggest cities. For example, in Taiwan I noticed that many people never cooked, and instead ate at inexpensive corner shops (often moble, on a cart). A similar thing already happens for many people in the US who order delivery 7 nights a week. And yet, every home still has a kitchen.

Seth Godin has an interesting post about different types of motivation, the thing that keeps you going when it’s hard. Go read it, I’ll still be here when you finish.

What’s my fuel? A few components jump out:

• Dissatisfaction (because it’s not good enough as it is)

• Engineer (because there’s a problem to be solved)

• Possibility (because we can, and it’ll be neat to see how it works in the world)

I’d really like it if I could add a few more to that list:

• Becoming a better version of myself

• Big dreams (because I can see it/feel it/taste it)

• Connection (because others will join in)

And then there are motivations that I identify less with. I know many people who thrive on competition. Others who have a creative itch, and still others driven by professionalism. All valid, all valuable. 

What’s your fuel?

Social networks tend to be free to use, because every additional user strengthens the system. A paid tier with exclusive features often don’t work because holding back features makes the free product less useful, and every user lost is very bad for the network. 

Products that are built with machine learning are similar, except the network effects are hidden. The data generated by each user can be used to make the system better, and, in general, whoever has the best data, wins.

A paid-only subscription to one of these products will typically result in a worse product than a freemium model, and it’s likely that a totally free model will be even better. The exception is when an important group of users feels compelled to pay to get the best product. For example, business users purchasing software. In this case, freemium is the way to go.

Angela Duckworth talks about the KIPP teaching thesaurus in Grit. Turns out that the wording teachers use can dramatically affect how students see themselves, and how they learn or don’t learn in the future.

Undermines Growth Mindset

Promotes Growth Mindset

“You’re a natural! I love that.”

“You’re a learner! I love that.

“Well, at least you tried!”

“That didn’t work. Let’s talk about how you approached it and what might work better.”

“Great Job! You’re so talented!”

“Great job! What’s one thing that could have been even better?”

“This is hard. Don’t feel bad if you can’t do it.”

“This is hard. Don’t feel bad if you can’t do it yet.”

“Maybe this just isn’t your strength. Don’t worry-you have other things to contribute.”

“I have high standards. I’m holding you to them because I know we can reach them together.”

A similar shift in messaging might work for teams:

Undermines Growth Mindset

Promotes Growth Mindset

“We don’t have the skills to do that”

“We can do that if we first learn these skills.”

“We’re going as fast as we can.”

“What’s one thing we could change to help us go faster?”

“Our launch went perfectly!”

“Great launch! How can we do even better next time?”

“So and so isn’t a good fit for this type of project.”

“So and so doesn’t have the right skills for this yet, but we can teach so and so by…”

“This isn’t our strength.”

“We need to hold ourselves to a higher standard with this because it isn’t yet our strength.”

Read more about growth mindset on wikipedia.

It’s great that self-driving cars will eventually reduce fatalities, but even today there are things that cities can do to make streets safer.

Research has demonstrated that making a street feel more dangerous causes drivers to increase their awareness which leads to fewer accidents. This is really interesting, and counterintuitive.

Cities can also make pedestrians and bikers more visible, for example with curb extensions (a.k.a. bulb-outs), trees, and parking protectedbike lanes.

Personally, I’d be down with wider sidewalks and narrow streets across the board, but that might have to wait until parking is less needed.

One useful way to think about self-driving cars is to imagine them as more perfect than human drivers, eliminating many of the inefficiencies that reduce the throughput of roads. For example, making intersections more efficient or reducing the compounding negative effects of traffic jams. 

It’s nice to imagine that these efficiency gains will lead to less traffic. Unfortunately, it’s also wrong. To understand why, consider a well-studied change has also increased capacity of roads over the past 50 years: the addition of more lanes and new highways. And the conclusion is that more capacity induces more demand such that overall traffic is not reduced.

From this research we should expect self-driving cars to increase total miles traveled by cars but without any noticeable reduction in traffic. 

If only that were the end of the story. Alas, self-driving cars not only increase the efficiency of the road, but also make the ride more comfortable for the driver who no longer has to drive.

A quick thought exercise: How much longer are you willing to commute if you could spend that time reading, watching netflix, or with your kids rather than focused on the road? In my own experience, I’m willing to almost double my commute.

This willingness to spend more time will translate directly into more induced demand and longer trips. Get ready for more gridlock.

My housemates, perhaps obviously, have keys to my home. Clearly I trust them not to walk off with my laptop while I’m out. So why don’t my best friends also have keys?

Even though it’s unusual, it makes a lot of sense. With a key to my place, a friend can

• wait inside when they arrive before I do

• leave their bike at my place when they are headed to an event nearby

• stop in to use the bathroom and see if anyone’s around

• use my living room as an “office” when I’m not around during the day

In the past this has been a bit of a hassle because keys need to be duplicated, and it’s very hard to revoke access. But with an internet-connected lock, this becomes far easier, not to mention trivial to log every access.

Containerization changed the world of shipping. But another insight happened at the same time: the realization that trucks, trains, and ships are all in the same business, the transportation of goods.

This might seem obvious now, but it wasn’t then. At the time, it was unthinkable to imagine building or buying an end to end string of trucks, railcars, ports, and ships that make it possible to guarantee a shipment’s arrival date in the hands of consumers on another content. 

Shipping containers helped reduce shipping costs by 50 to 90% by eliminating most of the work that was required to load and unload cargo from ships, trains, and trucks. But even after this technology was first proven to be effective, it took over a decade before it was clearly on the way to dominating the industry.

A similar concept is at work with modular construction, where walls and even entire structures are produced in a factory and are then placed on the construction site. The messy and time consuming alternative is to build and assemble everything locally, which is labor-intensive and harder to speed up. For now, modular construction hasn’t yet made a dent in building costs or times, but perhaps we are in a similar decade of uncertainty. 

A city is by definition big (in terms of density) compared to the surrounding area. And yet, paradoxically, cities are better at supporting smallness than suburban and rural areas. 

A small employer in the city doesn’t need to offer a parking lot, because the city has ample transportation. It doesn’t need to offer onsite meals, haircuts, or help with taxes. It can focus.

Further away from the city the big one stop shopping experience is king. First came malls, and then walmart. 

This notion of big enabling small can also be seen with aggregators like Google & Facebook which enable individual niche publishers in a way that newspapers never did. But, unlike a city, the leaders at Google & Facebook aren’t voted in by the people. 

 

When an afterschool program ends I often see parents wait in their cars to pick up their kids. With cars lined up one after each other parents don’t get the opportunity to see and chat with other parents they know. This problem simply wouldn’t exist if everyone was on foot.

Imagine how much more connected a school or afterschool program might be with parents that have a chance to chat every day or every week rather than hardly at all.

Here in San Francisco it’s common to see shops with bike parking, but rare to see bike parking on residential streets. 

Recently a home on my block renovated their garage and in the process added bike parking. I love it. What better way to tell guests and the public that bikes are welcome than to explicitly make space for them.