Tesla, Autopilot and Data Collection


f I were going to build a self-driving car, I’d want to have a large corpus of field data. What situations come up most commonly? How do humans handle them? How might our computer react in a similar situation?

Google’s self-driving car efforts are well known and they’ve spoken publicly about how all the miles their autonomous cars drive are carefully logged and analyzed. When a human takes over, presumably something has gone wrong or is at risk of going wrong and those situations are carefully scrutinized. Unfortunately this data is not what I’d call a clean sample of real human behavior. Everyone driving one of Google’s car is either an employee of Google or closely affiliated and most importantly, they know they’re driving one of Google’s special and expensive cars. No doubt they’ve signed a bunch of confidentiality and other forms. And so even if they’re driving themselves, they’re likely to be highly cautious. Those cars can only be used in certain controlled circumstances and the data Google can collect will be constrained.

When Tesla announced the autopilot features it really struck me that the hardware installed seems much more capable than what is really required for the features they’re offering. Radar for adaptive cruise control? Seems like overkill. But hundreds of thousands of Tesla cars with these sensors, all collecting data on their drivers and the situations they encounter seems like an amazing opportunity to build a corpus of real world situations encountered by human drivers and what they do. We already know that Tesla has the ability to connect to their cars remotely. What if Tesla is already deploying their self-driving software to their cars and running it in a mode where it’s just not hooked up to the actuators in the car? At every moment the car software could be simulating what it might be doing in the present situation and logging what happens when its choices differ from what the human actually does. Tesla engineers can then analyze these logs, adjust their software and re-simulate the car encountering that situation. At some point they’ll have it down to where the only places the human and the computer diverge is where they’re convinced the computer is making better choices. At that point, ship it! (Modulo lots of regulatory and insurance concerns.)

As an engineer that sounds exciting and cool. For me the ideal version of this would be that all the car data would be uploaded to HQ where I could analyze it indefinitely. Of course I’m sure customers and law enforcement would be interested to know if a full sensor download from all Tesla cars were being stored at Tesla HQ indefinitely. So it’s possible that the data would be anonymized somehow before being uploaded to HQ.

Does any Tesla owner out there want to share the privacy policy or the text of any opt-ins for the autopilot features on the new cars?

Catherine Rohr on Entrepreneurship

From “Life after VC“:

“Prison is America’s most overlooked talent pool,” she says. “Many gang members are proven entrepreneurs who built highly successful drug operations. The thing they were bad at was risk management.”

“We’re looking for leaders,” says Rohr. “If you weren’t good at selling crack on the street corner, you’re probably not right for PEP. We want the guys who know how to run an organization and get stuff done.”

Evidence Based Policy and an Obama Administration

That I’m a fan of Barack Obama is not a surprise to anyone who knows me. One of the things that appeals to me most about him is that he seems to be evidence based rather than ideology based. He is clearly very smart and secure enough in that intelligence that he can surround himself with other very smart people and actually listen to them. The thoughtful, nuanced positions he comes up with as a result may not play as well in a debate as pithy ideology, but they leave me with much more confidence that they are closer to “right” policy wise.

The outcome of the election is now looking likely enough that it’s not unreasonable to think about how the world might change under an Obama administration. NY Magazine just ran a lengthy, fascinating article on Obama’s transition planning process. Leading the process is John Podesta, former chief of staff to President Clinton and the guy who expected to be doing transition planning for Hillary Clinton. That Obama would tap a Clinton ally for the job (and that Podesta would take it) is refreshing compared to the political score-settling we have gotten used to over the last 8 years.

The article goes on at some length as to what the staffing and policy priorities of a hypothetical Obama administration might be and towards the end contains the following rather interesting quote:

Obama now informs Time’s Joe Klein that endeavoring to spark “a new energy economy [is] going to be my No. 1 priority when I get into office.”

Wondering what that may look like, I did a little Googling and came across the Center for American Progress, a think-tank helmed by the very same John Podesta. Front and center on the site is a link to “Green Recovery: A Program to Create Good Jobs and Start Building a Low-Carbon Economy”. There, Podesta outlines a $100B stimulus package that would invest in the following six things:

  • Retrofitting buildings to improve energy efficiency
  • Expanding mass transit and freight rail
  • Constructing “smart” electrical grid transmission systems
  • Wind power
  • Solar power
  • Next-generation biofuels

As a second year business-school student with strong entrepreneurial leanings and about to graduate into a rather rough job and capital market, understanding ahead of time where $100B in new spending is about to occur is welcome information indeed. More thoughts on the opportunities this presents shortly.

Microsoft Points and the Hillarity of Patents

It would seem I am now an “official” inventor. The USPTO has granted, ahem, Method and system for in-line secondary transactions, an element of the work I did on the system that ultimately became Microsoft Points.

For the record, I think the US Patent system is horribly, horribly broken, particularly when it comes to software patents. Does protection of this “invention” really foster innovation? I don’t think so.

Perhaps more poignantly, the incident reminds me of the fiasco of Microsoft Points, yet another project that could have, should have, would’ve been really cool, except that the drawn-out, risk-averse, by-committee, design process yielded a mushy product with deeply fatal flaws.

Ever wonder why Microsoft makes you do math to figure out how much something costs? It was based on the theory that pricing something in actual currency would bring unbearable regulatory scrutiny. An issue that Amazon, which also has a sizable market cap, brand and balance sheet to protect (not to mention Google, Apple and of course PayPal) seem to have circumvented nicely. That and a completely evidence-free conviction that users would spend more that way. (The so-called “casino effect”)

Of course, on the upside, participants in international, user-to-user transactions can quote each other prices over IM or email. So it’s got that going for it. How’s that scenario working out for you guys?

The (very) end of MSN Music

This week, Microsoft announced that they will be de-activating the DRM Servers that issue new keys for music from the (now discontinued) MSN Music service. In other words, starting June 1st, copying your purchased music to a new PC and expecting it to play will no longer work.

I bring this up because once upon a time, I was in charge of MSN Music and I built the system that Microsoft is now disconnecting.

Ian Rogers has described much more eloquently than I can how many of us got suckered, forced or seduced into spending our time and energy on systems incorporating DRM, convinced that it was inevitable or necessary or not that big a deal. (And let’s face it DVDs, XBOX video games and iTunes rentals are all proof that the right kind of DRM does not neccessarily equal market death.)

I continue to think that designing DRM must have been an interesting intellectual problem, and some pretty big brains came up with some rather clever stuff. It’s easy to see the temptation: The theory was that “big content” would not license their material for playback on a system that did not offer protection from copying and that customers had a strong demand for this content. So implementing DRM was a way to give consumers what they wanted and as a nice bonus, if Microsoft could implement acceptable protections in a way that Linux or Apple could or would not, then Windows would have a sustainable advantage.

It was this shift to “content owner as customer” rather than “customer as customer” that in my opinion started Microsoft down a long dark path from which it has not (and may never) emerge. Once the technology existed, and once Microsoft had displayed a willingness to put their customers second, Pandora’s box had been opened and content owners started to demand all sorts of outrageous things. And thus our own weapons were turned against us.

Unfortunately for Microsoft effective DRM on a PC is fairly obviously impossible without some rather intrusive hardware changes and even the approximate “hard for some people some of the time” roadblocks they managed to implement turned out to be rather cumbersome, never worked reliably and made the entire system less stable, less usable and less useful. One hopes that someone wondered about the wisdom of sinking so much time and effort into an “anti-feature” – work done to ensure that Windows could do less than the version before it, but once promises had been made there was likely no turning back.

Lessons learned for me from this:

  • If a large percentage of the work you are doing is for the benefit of someone other than your paying customer, think very hard about whether you’re doing the right thing.
  • Committing to do something “forever” is an obligation you should only encumber upon yourself and very thoughtfully – like in the context of a marriage. Committing someone else (like your successors at work) in this manner is almost certainly a mistake.

So my apologies to the customers who bought what I now see as a defective product. And apologies to Rob Bennett, General Manager of MSN Entertainment, who has had to clean up more than his fair share of messes rather than getting to focus on growing his business. There was a time when there was hope and optimism around this business.

Samsung on Failure

Chairman Kun Hee Lee of Samsung Electronics, as quoted in the HBS case on that company:

At Sumsung, we reward outstanding performance; we do not punish failure. This is my personal philosophy and belief. We need punishment only for those who lack ethics, are unfair, tell lies, hold others back or stand in the way of our unified march.

The media landscape

Although School doesn’t necessarily leave you with any more free time than working does, it does leave you with plenty of “mental space” – room to think about new things, how the world might be changing around you and what opportunities that may present.

One thing that’s been on my mind lately is the writers strike and what it means. A couple of articles have leapt out at me:

  • The international herald tribune had an interesting article about the now bleak economics of movie making has become for the major studios.
  • Marc Andreesen’s rather excellent blog had a post about Hollywood reforming in the image of Silicon Valley.
  • Patrick Goldstein picked up the thread with an excellent article in the LA Times that describes the “entrepeneur artist” and cites Spielberg, Jackson, Lasseter, and Lucas as ahead of their time archetypes.

The theme of all this is that falling production costs, proliferating distribution channels and generally crummy economics weaken the grip big studios have traditionally exerted on the movie business. There may be a coming wave of entrepreneur-artists who make modest budget movies outside of the studio system, take greater artistic risks and get paid like owners rather than hired guns.

More powerful artists suggests that the vertical integration of the industry is going to start cracking. Funding, production, promotion and distribution are all separate functions that may not be operated by the same entity.

A few observations:

  • Raising $7M doesn’t seem like a big deal if you think you’ve got the right project. Attaching a known director or actor seems like a good way to convince an investor that you’re on to something. I can imagine funds which invest in a diversified portfolio of projects or “angels” who invest in a single project that appeals to them.
  • Production at a lower budget means that costs that used to be insignificant will suddenly start getting scrutiny. Unionized labor that gets paid at least 8 hours a day regardless of how much they work and have all their meals catered seem like low-hanging fruit.
  • Setting up and ripping down a production company every time you work on a new project seems incredibly inefficient. I can imagine standing companies that know how to work together and move from project to project.
  • The perhaps-apocryphal Disney executive who asked his team to “only make the hits” may have in fact been onto something. In order to keep the machinery of a studio running, it needs to have a pipeline. Rather than funding potential hit movies, studios are really funding the best N projects they can find and perhaps this leads to poor funding choices. The Last Boyscout, The Sixth Sense and The Matrix are all movies that apparently “sold themselves” from the script, suggesting that at least in a few cases a script is so compelling as to suggest that in the right hands it will be a hit.
  • Promotion and Distribution still seem like the core difficulty. “On the net” is still not a great venue for watching visual entertainment longer than 10 min and has the nasty side issue that nobody wants to pay. Certainly creating celluloid prints does not seem cost effective.

Things I am thinking about:

  • How do we connect investors with projects? How do they sort good projects from bad ones? This process already happens through an informal social network, can one be systematized?
  • How does investing in low cost movies work? What financial structures do we use to share the upside without removing incentive? Who has succeeded here before?
  • What is really happening to production costs? What are the real cost drivers? How do we accelerate that drop? Not every movie is the Blair Witch Project.
  • How do we find an audience for these new works and how do we get our work out to them?
  • Many artists enter the business motivated by fame. Some movies really do call out for big budgets and huge promotion. Any re-imagination of the business that eliminates either of these things will not work.