Death Race 2000: How Safe Will Autonomous Vehicles Need To Be?


The technology behind autonomous navigation is developing quite rapidly. From our early investment in Cruise, we’ve gotten a vantage point on just how much progress has been made in the past three years alone. But if I’m confident about the tech, I’m less sure about the target. That is, how “safe” do autonomous vehicles need to be? And what is today’s reality? 35,092 automotive deaths in the US in the year 2015 (on a per capita basis, down significantly from the 1970s).

So is the target number for autonomous vehicles 35,091 deaths per year? That seems rational but would require us *not* overreacting to headlines which blame errors by the technology for deaths.

Or is the number some smaller percentage of current fatalities – say 10,000 deaths per year – owing to the emotional reaction that the benefits will need to be quantitatively significant in order to make drivers (and regulators) comfortable with giving up control to the machines?

Maybe controversially we should be able to tolerate more deaths per year in the move to autonomy. Certainly if you look at the deaths per capita average over the past 100 years, there were eras nearly 3x today! And if autonomous vehicles add value in aggregate because they support faster, more efficient, more relaxed shipping and travel, shouldn’t we tolerate more danger in return?

Besides wondering generally about the psychology of safety and autonomy, where we end up on this “how safe” spectrum will also influence two important factors: how quickly we attempt to move from 0% to 100% autonomous, and the role of insurance.

In terms of rollout velocity, it’s fairly noncontroversial to suggest that if every car was autonomous (and “talked” to each other continuously while all maintaining a shared or similar “safety” ruleset), the road would be safer than a 50/50 mix. Has anyone seen a graph that takes autonomous vehicles as a percentage of active vehicles and correlates against projected automotive accident rate? These forecasts will certainly be influencers to the regulatory and financial incentives that accelerate to autonomous density once the technology crosses mainstream viability. It’s a classic “it’ll happen slowly, then very quickly” dynamic.

The second question I have is about the role of insurance. As distasteful as it may seem, we do today have ways to value a life based on number of factors including age, race, vocation, geography and so on. The “how safe” question might be moot (excluding corporate negligence) if we’re comfortable with allowing insurance to fill the gap between desired safety and actual safety. That is, let’s say autonomous vehicles were twice as unsafe relative to today. Do we need as a society to wait until safety improves or can we use financial compensation as a mechanism to bridge the gap. And whose insurance? The driver, as it sits today. Or will the manufacturers (Tesla, GM, etc) need to carry liability insurance as a passthrough if it’s decided the driver wasn’t at fault in an accident but instead the algorithm was?

Sometimes The Secret Is To Just Show Up & Do Your Job

This is definitely the only blog post I’ve written from a Milan airport lounge, returning from my first trip abroad since we started Homebrew. So what got me out of my San Francisco routine? The chance to work with a group of global mayors to help push thinking forward on how cities and technology companies can work together productively. Cities are my DNA and where I feel happiest – raised in New York and now for the past 20 years in San Francisco. They are the hubs of population and of innovation. I’ll write more about the CityXChange event once it’s settled in my mind a bit more, but this post is about how I got invited.

Screen Shot 2017-05-17 at 2.40.31 PM

There were six VCs attending from US and global funds. All of those funds are substantially bigger than ours – as much as 25x larger! While obviously bigger isn’t always better, I did want our hosts to understand the difference in resources that those sorts of funds have compared to a seed firm like ours, so as the attendee list was solidifying, I checked in with the organizers just to make sure they wanted to use an attendance slot on lil’ ole me.

“Hey, I’m totally honored to be invited but you know we don’t quite have the team or dollars that some of these other partners have at their access,” I told my friend who had invited me to the summit.

“Yeah I get that,” he replied, “but you sometimes have good ideas and you follow through on what you commit to doing. You’d be surprised how much that matters.” [note: *sometimes* ?!? 🙂 ]

So that’s what stuck with him. And I think it’s a reminder that sometimes nailing the basics of being someone people want on their team opens doors wider and broader than putting endless energy into window dressing.

While there’s a lifetime of good fortune, work and, yes, some privilege, that I benefit from, if you show up, have good ideas, and do the work, you might end up with opportunities you didn’t expect. Some may even be in Italy!


Why Trump’s Victory Caused Alec Ross To Run for Governor of Maryland – Five Question Interview

My friend Alec Ross is running to be Governor of Maryland. I support him and hope you’ll join me, just go to his website to learn more. I’ve known Alec since early in Obama’s presidency. He was working for Secretary Clinton on an initiative on “21st Century Statecraft.” This project focused on ensuring America was sending more than just “guns and butter” abroad (ie military intervention or economic aid). The currency of the modern world was ideas; how could the US add that export to its foreign policy toolset? One effort under that umbrella  was their Tech Delegation trips, which sent groups of technology executives to foreign countries for week-long working sessions. In April 2009 I was invited to visit Iraq as part of the first delegation. It was an amazing experience for me and I’ve been close with Alec since that point. Thought the best way to share my enthusiasm would be a Five Question interview. Here you go:

Hunter Walk: “Just when I thought I was out, they pull me back in” — was it the 2016 Presidential election outcome which started you towards running for Governor of Maryland or had you been considering it prior?

Alec Ross: I had no plans to run for office any time soon. It was Trump’s election that drew me in. I grew up in coal country, the white hot center of Trump support and I spend a lot of time with millennials who “care” but don’t vote. I decided to run because I think America needs a new wave of leaders with new ideas.

HW: Technology startups tend to work from outside the system before, ironically, if they’re successfully then becoming the establishment they were trying to disrupt. Why do you personally believe that government is still the best system to catalyze change, versus working from the private sector?

AR: I think the private sector can be a powerful driver of change, but the power of government to impact peoples’ lives for the better or the worse remains unrivaled. The budget for the State of Maryland is $41 billion, with a $17 billion operating budget. Those resources impact peoples’ health, education and transportation in a way that exceeds what the private sector does in these fields.

HW: Did you give your family a “go or no go” vote before declaring your candidacy? How did your kids react?

AR: My wife is a 6th grade math teacher in Baltimore’s public schools so she sees jarring examples of inequality and lack of opportunity everyday. She was excited for me to run because she knew I would be a champion for our public schools.

The kids’ only sort of get it. They’re 14, 12 and 10 years old. They’ve seen the spotlight of public attention before because of my time in the Obama Administration, but it’s a little different with their Daddy being the one out there under his own banner. I think they’re excited about it if a little weirded out at times.

HW: Silicon Valley isn’t demographically or economically representative of the “average American.” Do you Mark Zuckerberg’s “50 State Listening Tour” is an example of the tech industry breaking out of the echo chamber? How can the average startup employee get a sense of what America is going through right now.

AR: I’m usually a defender of Silicon Valley but I have to agree that it is isolated from what most of America is going through. The economic recovery was really a recovery along the Atlantic and Pacific coasts. Much of middle America is still stagnating. I think it’s all for the good that Mark went out on a listening tour and I would only encourage others to do the same. And here’s the thing — I think it will actually help make people better investors and businesspeople. Reconnecting with middle America can’t help but open your eyes to problems that need solving, and entrepreneurs are inherently problem-solvers.

HW: No Fake News! Which journalists and pundits are worth reading, watching, listening to?

AR: I’m really cranky about most journalism these days. Most of the reporting of the U.S. presidential election made me want to puke. The media needs to take responsibility for the monster they helped create in Donald Trump and some of what they over-reported and some of what they under-reported still makes me red-faced angry.

Okay — now that I’ve gotten that off my chest — here are some folks worth following IMHO:

David Ignatius from The Washington Post

Jon Lovett from Pod Save America

Gideon Rachman  and Edward Luce from the Financial Times

Matthew Bishop and everybody else at the Economist

Fareed Zakaria

Julia Ioffe and James Fallows from The Atlantic

Susan Glasser from Politico

Why You Shouldn’t Loan Money to Someone Who Thanks God

Crunching data from Prosper, the peer to peer lending site, researchers correlated default rates with words most likely to appear in a loan application. From an article in New York magazine:

Screen Shot 2017-05-10 at 9.50.38 PM

Somewhat intuitive when you look at the types of words (the former suggest a degree of financial planning; the latter hopes and prayers) but still fascinating. And what about using data like this to reverse-engineer “better” loan applications? Of course at scale that would start to skew the data overall, but done selectively you could create “false positive” loan applications by mimicking the language from “good” loans. Sorta how I’ve always wondered if successfully-funded Kickstarter projects have project descriptions that are substantially different than those which fail.

Four Years of Homebrew: Notes From Our 2017 Annual Investor Meeting

“The days are long but the years are short” applies to parenting and venture firms. Satya and I raised Homebrew’s first fund Q1 2013 and now, four years later, we held our 4th annual meeting last week, just days before our official birthday (anniversary of fund close).


For VCs, annual LP meetings are combination Board Meeting + Community Event. We spent an afternoon with our investors, providing updates on fund performance, market dynamics and future plans. Several founders gave 10-12 minute updates on their companies, and then, a dinner with our LPs, Homebrew founders and fund advisors. It’s of course not the only time our investors hear from us – we send our quarterly letters with financial statement, and given that our investor base is largely institutional, see them individually once or twice a year in-person (in addition to ‘as-needed’ conversations when they’re referencing other fund managers, discussing direct investment opportunities in our portfolio companies and so on).

If you’re interested in what gets covered and how these meetings evolve, here are summaries from our first three:

Year One – 2014

Year Two – 2015

Year Three – 2016

Satya wrote up Year Four – 2017 so rather than repeat what he covered, let me add just two personal observations.

  1. Founder < > Market Fit Is Beautiful – Watching five founders present their companies we were struck by how strongly each startup was an extension of their founders, even as some scaled to teams an order of magnitude larger than they were when seed funded. Homebrew founders care deeply about the problems they are solving and treat their company as their first product, in the care they apply to building diverse teams, mission-driven cultures and responsible growth strategies.
  2. I Don’t Generically Like VC. I Like Satya, Our Founders, Our Investors and Homebrew’s Model – If I wasn’t doing Homebrew, I likely wouldn’t be the 3rd, 7th or 11th partner at someone else’s fund. I wouldn’t be the consumer guy at a $1b fund. And I wouldn’t be making 100 small investments a year. Maybe I’ll feel differently in three, seven or 10 years, but just as we care about Founder < > Market fit in our investments, Homebrew was designed with the same care. That’s one reason why, when people ask me about interest in “getting into venture,” I try to help them understand that different roles, different models, different firms can lead to very different experiences and they shouldn’t generically seek to “become a VC.”

On to Year Five!


We Value Conversion More Than We Value Trust: Wading Into The Unroll.Me Discussion

So the moment I read Mike Isaac’s TK profile it jumped out that was going to have a problem on their hands. Their consumers were surprised at what was being done with their data and no one likes surprises. Regardless of legality (it appears 100% legal) and regardless of uniqueness (what is doing isn’t unique), their PR team’s initial response was “read the ToS.” Uh, no, wrong answer. Their CEO then published a blog post which was either sincere or insincere depending on your parsing of “sorry” and general cynicism. Others chimed in with “what did you think, it was free” vs the “my data is my data” purity test. Then today, a friend of the CEO wrote a passionate essay basically saying the CEO is a good guy and tech industry sucks overall.

But let me parse a bit why faced a backlash despite having the right to aggregate and sell anonymized data in exchange for giving you a free product:

Surprised Users! It seems that somewhere along the way evolved how they were monetizing their understanding of your inbox. At first it was to serve you targeted ads in their notification emails to you, but then it became about aggregating and anonymizing data to sell to third parties (this is what Uber purchased). They may have updated the ToS to give them this right, but they seemingly never communicated proactively in plain language what this meant. Why? Because it’s another friction in conversion and, as an industry, we measure funnel dropoff to a second decimal but don’t measure trust often or well. And they put trust at risk.  If they’d proactively notified users and provided an explainer, perhaps they could have mitigated the surprise proactively:

“Hi. In order to keep this service free for you we’re doing something new to help us pay the bills. In helping you manage your inbox we do analysis on which businesses are sending you email and what they’re notifying you about. This information – in aggregate and anonymized – is useful as a market data product we make available to third parties. For example, we may create research reports which help airlines understand trends in travel based on email receipts. Your individual info is never shared, your personal email data NEVER leaves our secure servers and you can [opt out of this panel and upgrade to our premium product; delete your account and all info we have stored; etc] at any time.”

Net-net, the expansion of their business model was perfectly legit but sensitive enough to the average consumer who thought the service was just helping them identify and unsubscribe from mailing lists. Which brings us to….

Ad Targeting vs Data Selling. But was always targeting you based on your data! Yes, they were serving you targeted ads in email notifications based on their understanding of your interests. However, this still feels like they’re keeping your data within their walls. Of the thousands of ads you might see, they’re selecting one that’s relevant to you. This isn’t at scary as the idea that your data (even anonymized and aggregated) is being bundled and sold outside of their corporate systems. Again let me emphasize this post is about why people felt angry, it’s emotion not logic. There was a toxic shock reaction to the idea of one’s data being decoupled from their use of the service and sold. It’s why a company using your data to improve your experience (and their monetization) inside of their product is perceived differently than your data being used by that same company to inform a 3rd party.

But Her Emails. The value prop is simply stated – here’s a picture from five minutes ago

Screen Shot 2017-04-25 at 8.58.41 PM

From this simple screen you are prompted to give full access to your email account. Obviously many people gave that access (I did a while back but a few years ago decided no startup should have access to my inbox – too risky). The average person doesn’t think it’s scanning the contents of all your email, just recognizing sender and subject lines. Even a technically aware person might think they’re working off some collaboratively filtered white list and black list of sender addresses, not scanning your Lyft receipts and recording metadata. was building a much more sophisticated view of you as a user than they may have needed to deliver much of their core value proposition. Again, not illegal, but ultimately invasive.

I’m writing this up because nearly a decade of work at Google gave me some insider perspectives on how consumer react to different assumptions around privacy (I was there when Gmail launched and was ‘reading’ your emails). And I’d urge startups to not be too cute but evading the discussion with their users but instead follow a set of best practices around ensuring users have control of their data and understand the tradeoffs they’re making in how you are using it to their benefit as well as the company’s needs.

From an outsider perspective did nothing illegal, or even that uncommon, and every consumer should be proactive in learning how the services they trust maintain or abuse that trust. But did fall short of best practices and potentially even crossed into a gray area ethically depending on the nature of any internal conversations around the level of consumer disclosure they should make once their business model changed.

Seed Investing & Dark Dealflow: You Lose 100% Of The Deals You Don’t See


Folks outside of seed stage venture often think “winning” deals is the most important part of being a great investor, but I’d argue it’s “seeing” the best opportunities that’s actually the strongest indicator of future fund success. Or to put another way,

I rather see 100% of the top seed opportunities and win 50% than see 50% and win 100%

What outsiders don’t realize is that much of the most intriguing seed startups go through a financing process that’s “dark” – not seen by most investors because it’s either competitively privileged to a few firms or so against traditional patterns that the walk is more random.

Satya and I track “misses” (NYC and SF/BayArea startups in our areas of interest) as rigorously as the other phases of a traditional “See -> Pick -> Win” funnel. Four years into Homebrew, we’re fortunate to feel really good about our relationships and reputations — you live and die by those in our business.

When we do “miss” an opportunity there are two main causes:

1. Founders With Strong Pre-Existing Venture Relationships – for example, repeat founders going back to their same investors and not expanding the pool. What can we do in these situations? Make ourselves tangible in our practices and areas of expertise and build case for founders that these are worth adding to their cap table and will best serve them in their current company. Also seek to build familiarity pre-funding process so that we too can have a “pre-existing” relationship.

2. New Areas of Interest That Our Portfolio Doesn’t Yet Represent – SaaS, Marketplaces, Commerce, Autonomous Robotics, Financial Services/FinTech. These are some of the verticals that our portfolio suggests we have interest and a POV. But at any given time there are areas that we’re exploring actively without a current investment, or we’ve made investments but they’re not yet public. When we “miss” an opportunity the feedback we sometimes hear is “oh, we didn’t know you cared about XYZ.” So we plan to be a little more open and ‘learn in public’ about areas of interest. Given our culture – and scale – this is less like to be via large scale “content marketing” and more hands-on learning and sharing of information publicly. Maybe it helps us, maybe it helps founders, maybe it helps other investors too.

Coming out of our 4th Annual LP Meeting last week, we’re thinking in particular about #2 and how to implement some new ideas there.