Abductive Intelligence

Dear Twitter, Please Just (Un)Wrap the Links for Me

As long as Twitter is changing the handling of URLs, why don’t they just:

1. take my super long link as part of the Tweet
2. automatically shorten it to some fixed number of characters (t.co + hash)
3. automatically unwrap it for display in “new” (upgraded) Twitter clients, including the web site

???

I had to read a couple of pages of email and blog to figure out that you’re working on this, but what is taking so long?

Smart Investors Evaluate How You Start a Company (or Project)

Many of the people that approach me with projects or new companies have ideas about getting feedback, money, effort, or something else from me.  As a person with an entrepreneurial spirit, I give out a lot of “free advice” and help with matchmaking with the expectation of little in return.  I do this for many reasons that I won’t explore today.

But, as an investor, I am watching your intuition.  When you negotiate deals, is your aim to maximize your benefits at the expense of others?  When you setup projects, do you consider the best interests of everyone in addition to your interests?  If the answer to these questions is ‘no’ then I can see you in a career as a transactional trader on Wall Street but I’m probably going to shy away from investing in your project.

Your negotiating intuition is my window into whether you have an exploitative personality. Turn the tables: if you are investor in my company and I need to work 100 hour weeks to make the firm successful, if you want to own 99.9% of the firm (and I get virtually nothing) then why would I do this?  Thus, as a rule, investors need to let founders keep a large percentage of a company so that they have incentive to keep working at the company and using their talents to build the firm.

Similarly, if your hiring plan indicates that you expect to get a lot of work out of skilled employees for nothing.  Or, worse, you intend to trick them into believing they have something when they have nothing, then what should I think of you?  I know of many cases where this has happened, but I prefer deals that aren’t dependent on the founders being pigs. 

This investing strategy isn’t really about idealism.  It’s evolved over time as I realized that deals aren’t purely transactional, they are part of a repeated game.  Stronger people do not need to exploit.  And, as a founder, if it means that you only end up with a BMW instead of a Ferrari, I expect you to suck it up and realize that you probably made the big picture mistake that resulted in that outcome.

A Difference Between University and Startup Mentality

It seems like I’ve been writing a version of this email once every two weeks all summer, so it is time to commit it to the blog.  This post is for students and professors in academia (not a business school) that are trying to communicate to me what their business is doing.  It is relevant to Fall CS 5150 students engaged in the joint Entrepreneurship / Computer Science program.

When you work in an interdisciplinary program (and you have some notion of the definition of successful communication), you quickly realize that academic fiefdoms speak very different languages and, therefore, communication between them is more difficult than you might expect.  Students are expected to learn to mimic the communication protocols of their instructors (regardless of how efficient they are) and this is one of the primary standards for whether students are judged as successful (whether they receive high grades).  Since this post isn’t really a commentary on this situation, I won’t dwell on it.  This backdrop is important because when students leave the nest of the university and enter the business world, the communication paradigm changes completely (because the vast majority of people in academia have little business success and the vast majority of people in business have little academic success). 

Despite the illusion of precision in methods, these communication problems translate into an immediate problem for a student (or faculty member) entering the work force (or vice versa).  People need to understand what you are trying to tell them.  Ironically, because business people interact so frequently across a broad range of skill sets, a lot of the terminology in business seems more standardized (more commonly understood and frequently defined the same way) than it is in academia.  As a business person, I am not willing to read 10 pages of nonsense to understand that you are trying to invent a new term to replace the phrase “capitalization table” in a startup as it applies to your type of business. (Note that this doesn’t mean that cap tables are standardized.  It just means that it is difficult to function in an executive position and not know what a cap table is.)

Back to the situation at hand.  You are a student/faculty member and you are trying to start a business.  Step 1 is to have an idea.  Step 2 is to learn to communicate it effectively to someone else. The simplest way that I can break down how to do that is to ask you to answer three questions:

1. What are the two most critical experiments that you need to run to determine whether your idea may uncover “something important”?
2. If the experiments go well, how would you characterize the “something important” from question 1?
3. What is the cost of the team and resources to conduct these experiments (per month and for the duration of the experiments)?  (in your answer, you must address how much buffer you have for unexpected problems)

The combination of the answers to these questions form the basis for the hiring plan for a business team, the budget for their work, the product positioning statement, and the problem they are attacking.  I cannot stress how critical the answers are to understanding the scope and focus of a team.  These questions actually have a parallel in academia, but the expectations for the answers are (in my experience) strikingly different.

If you are a student in academia, there is an expectation that you can link your work to the research interests of others, past and present.  When you cite theorems, related work and established methods, you are positioning your work inside of a research paradigm.  Similarly, when you build a business you situate your enterprise in the landscape of related companies, past and present, and the market trends that people believe in.  In the startup world, each company (failed or successful) is an experiment.  Venture Capitalists and entrepreneurs learn the experiments associated with different startups and the outcomes.  This learning is similar to the ability to conduct and communicate a literature review in academia.

A second parallel between academia and business exists in experiment selection.  Similar to higher quality entrepreneurs and business leaders, higher quality PhD students, faculty, and program administrators at NSF/NIST/etc. give solid advice about “useful” experiment selection.  The definition of “useful” experiment selection itself deserves an entire blog post, but an important distinction between academia and business is the definition of useful experiment.  A useful experiment in business is, at its core, one that can lead to a profitable revenue stream.  In academia, successful experiments have a definition that is dependent upon the academic domain’s norms and standards, which are set by the ruling political class of the domain.  Understanding your domain’s political ruling class and norms/standards is always a key gauge of success.  Students prove this by graduating.  Faculty prove this by having papers accepted and grants approved.  Business leaders demonstrate this by getting funding for their projects.

As I stated in the introduction, this post is for students and professors in academia that are trying to communicate to me the primary activities of their business.  You are trying to move into a world that has a different communication paradigm and different political values.  You must adapt to it because, unless you are a super nova superstar, it will not adapt to you.  If you don’t understand the paradigm, have some humility and ask.  The advice may cost money and you get what you pay for.

Salmon Sous-Vide Experiments

We’ve been experimenting with fresh king salmon cooked sous-vide and I have a few notes that I want to record.  These methods work well in a beer-cooler sous vide setup because the cooking times are short.  For the beer cooler sous-vide, you’ll want to practice a few times to get the starting water temperature correct (the water temp will decline after you insert the cold fish).

First, the fillet cuts should be about 1.25” to 1.75” inch thick and weigh approximately 0.5 to 0.7 lbs.  Thinner, smaller fillets results in mushier outcomes with these procedures.

Second, 20 minutes in an iced salt-water brine solution prior to cooking substantially helps with maintaining the color of the fillets.  The cold salt-water helps prevent the discoloration of the salmon with albumen.  To brine the fish:

1. Pour 2 cups of kosher salt into 2 quarts of distilled or tap water in a container that will allow the salmon to be completely submerged.  The water should be room temperature when mixed with the salt to encourage suspension.
2. Add 2 to 3 lbs of crushed ice.  Put the mixture in the refrigerator and wait until the water temperature drops to 35 degrees F before proceeding.
3. Add the salmon to the brine.
4. Wait 20 minutes.
5. Remove the salmon from the brine.
6. Rinse if desired.
7. Pat dry.
8. Vacuum seal the fish for sous-vide cooking.  We seal the fish with maple syrup.  Yesterday it was BLiS bourbon barrel aged maple syrup.

While the salmon is in the brine, prepare the water bath.  To cook the salmon to rare (but just beginning to flake well), the water temperature should be set to 116 to 118 degrees F.  To cook to medium rare, the water temperature should be 126 degrees F.  To cook to medium, the water temperature should be 140 degrees F.  If you want to cook the fish above medium, stop reading and just buy canned tuna. ;)

In the past, we’ve cooked the salmon to 122 degrees F and then immediately consumed it from the water bath.  Yesterday, we pan finished it to encourage reduction and glazing from the BLiS maple syrup, so I set the bath to 119.5 degrees F.  But even a quick sear in the pan brought the final fish temperature up too high.  Next time we recommend trying 115 or 116 degrees F for the water bath if you intend to quick sear pan finish or crisp it with a kitchen torch.

Finally, the amount of time that the fish stays in the bath is very important.  12 minutes in the water bath is just on the edge of being enough.  We typically cook to just shy of 20 minutes.  40 minutes is the absolute maximum and every minute over 20 minutes begins to make the fish lose its crispness.

Happy cooking!

Johnson Business School Integrates with CS 5150 to form Startup Incubator

Exciting news!  We’ve tied the Johnson Business School’s entrepreneurship course with CS 5150 for the Fall Semester so that students in both the business school and computer science can build startups at Cornell while receiving guidance from the venture community and the computer and information science community.

Up to 3 joint MBA - CS/IS teams will be allowed this year.  The teams will consist of up to 3 MBA students and 5 to 7 CS/IS/MEng students.  David BenDaniel will coordinate the MBA side while William Arms and I oversee the work with the CS/MEng/IS student teams.

The teams will benefit from NYC VC guest speakers, startup CTO guest speakers, and participation in 1 to 3 startup incubator competitions.

See past articles 1 and 2 on this topic …

Machine Learned “Initial Temp” Calculator for Beer Cooler Sous-Vide

The summer sous-vide machine learning project has paid off.  I now have enough data to estimate the desired initial starting temperature for the water, given the amount of water in the cooler, a history of the cooler’s temperature logs, the starting mass of the food, the desired stasis temperature and the desired cooking time.

The machine learning algorithms estimate the hidden temperature loss function of a beer cooler (sous-vide cooking vessel) well enough so that “correct” cooking can be approximated without a heating element.  There are downsides to version 0.1 of this technique — it only works well in a small temperature range (100 to 140 degrees F) and it hasn’t been tested on a lot of different vessels.

Once I have some more success with the calculator on predicting out of sample (in real cooking situations), I will publish the calculator on the web for others to use.