45 comments posted · 4 followers · following 0

11 years ago @ Redeye VC - Why First Round Capita... · 0 replies · +1 points

When I hear the "real world", I suspect that maybe they are talking about NSFW content or believe that anything from research is just intellectual self-abuse and a waste of time, money, and effort. Here I will assume you mean the second.

Sure jump through all the hoops you want, and call them 'reference customers' instead of thieves, but don't give away the secret sauce.

For what's in the "real world", I'm not sure you are understanding and accepting the power and value of good secret sauce. Rare? Yes. Impossible? No. And, for what's rare, since any very good success is rare, we have to be willing to consider what's rare.

With such entrepreneurship, we're talking the valuable part of the future here, so can't see all of it in commodity products in the rear view mirror.

Still, there are plenty of ways to see the potential of this future: Likely the easiest way is to look at some of the better US DoD projects since, say, 1941.

E.g., I long wondered what would happen if a good WWII army came up against the US. Well, in Gulf War I, we found out. There was a bombing campaign, and then it took exactly 100 more hours to blow away the Iraqi army. Remember the "Highway of Death", that is, the result of some Iraqis leaving Kuwait and found by some US airplanes, maybe an A-10.

In Gulf War I we may have had more injuries from R&R than enemy action.

Some Iraqi tanks started to threaten, so we dropped one of our special bombs that come all apart into little pieces with each piece smart enough and with control enough to home in on its own target. Net, one bomb and all the tanks dead.

An Iraqi tank was parked between two buildings, so the US hit the tank but saved the buildings. In response an Iraqi officer said "American military technology is beyond belief.".

The gun on the Iraqi tanks had a range of about 1 mile, and the tank had to stop to aim to shoot. The gun on the US tank had a range of about 2 miles, and the tank could aim and shoot accurately while moving. So, when some US tanks were moving forward and just saw some Iraqi tanks, guess what happened? I'll give you three guesses, but you will only need one.

Once I was programming at the JHU/APL in Howard County, MD in the group that did the orbit determination for the Navy's version, the first version, of GPS. There was a unit on the roof, and the system reported the position within a foot. [I was writing software for passive sonar adaptive beam forming.] Now we have GPS in smart phones!

Well, there were Friday seminars, and I went to one on Ga-Al-As Hetero-junctions. Right, sounds like gibberish and not the "real world". But what it was on was solid state lasers, each about the size of the head of a pin. Why? To send laser light down an optical fiber! Why? Chop the laser light and, thus, send bits. At the time, the entire US voice long distance bandwidth was well under 30 Gbps, and now can get 40 Gbps or so from one pin head of a Ga-Al-As hetero-junction laser on one optical fiber. By picking different wavelengths for the lasers, can put some dozens of such signals on one fiber. Then put maybe 144 fibers in a cable, and put the cables along high voltage electric power lines, railroads, highways, pipelines, lakes, rivers, coastlines, etc. Might want to amplify the signal: Right, can do that just by passing the optical fiber through a cute device that doesn't really even touch the bits and leaves them just in the optical signal. With good fiber, can go a good fraction of the way across the landscape before need to amplify. What to do with such things? The Internet! So, get to 1 billion users and 100 million Web sites quickly, grab enough ad revenue to sink much of old print media, create a problem of how to find stuff in the millions of Web sites, trillions of Web pages, and build, say, Google, as essentially a computer based library card catalog subject index worth ballpark $200 billion. You've noticed such things, right?

Close enough to the "real world"?

There's more, BELIEVE me, there's MUCH more. How to know? It's sufficient just to build on work in math, science, engineering, and technology, which have not yet been nearly fully exploited.

11 years ago @ Redeye VC - Why First Round Capita... · 2 replies · +7 points

Congrats on winning the law suit. Thanks for fighting back at the big 'suits'.

First mistake: You signed an NDA, that is, a full employment arrangement for lawyers. Instead, the best way to keep something secret is just, and here's the amazing answer, don't tell anyone!

Second mistake: You didn't appreciate the high power of some powerful, valuable 'secret sauce' that is difficult to duplicate or equal. Eventually you learned from the e-mails that without the disclosure there was little chance the work could have been duplicated or equaled. Now you have a better appreciation for the potential of secret sauce.

Should'a known that: Teach some technical courses in, say, applied math, to some undergrads and grads and begin to see what fraction of the students can actually work the more challenging exercises. Hint: Nearly none.

Lesson: If you've got some secret sauce that actually took some good thinking to construct, then the chances some competitor will be able to duplicate or equal that are slim to none.

Here are four strong reasons why:

First the employees at the competitor just won't go to the trouble to think hard enough. E.g., to do that thinking about have to get into a quiet room, keep the door closed, refuse to answer the phone, get things quiet and KEEP them quiet ("noise is the mortal enemy of thought"), pop open a cold can of caffeinated soda, put feet up, think carefully, slowly, and deeply, leave the TV off, don't go out for a beer, etc. Only a tiny fraction of the population has ever done such a thing, and only a tiny fraction of that tiny fraction is good at doing such things.

Second for most good secret sauce, those employees didn't take the right, challenging, advanced courses in grad school (hint: they are NOT in 'computer science') and, thus, don't have the prerequisites and most definitely will NOT just reinvent those grand results on their own.

Third, the competitor likely has a 'hierarchy', and in that hierarchy no manager short of the CEO wants anyone working for him stirring up secret sauce he doesn't understand or couldn't stir up on his own. I.e., such hierarchies are still like the factory floor at Ford 100 years ago where the supervisor knew more, the subordinate was there just to add muscle to the work of the supervisor, and any subordinate who might have known more than his supervisor was a threat to the supervisor (look up 'goal subordination'). So, really, in a competitor, no one short of the CEO will be much of a threat in stirring up secret sauce.

Fourth, for the three reasons above and much more, big organizations don't innovate. Steal? Maybe. Buy? Sure. Innovate? Nearly never.

Exercise: A lot of examples in VC funded startups? Nope. Conclusion: Secret sauce stirring is unpromising and has poor ROI or, for people who know how to do it, is very promising and has fantastic ROI?

Answer: In ROI, secret sauce stirring can totally blow the doors off Moore's law and, really, nearly anything else.

Lesson: When an entrepreneur comes in, don't evaluate his project just by looking at the UI/UX which is like judging a book just by its cover, sticking with just what's obvious and tangible, ignoring the crucial secret sauce, and straining over gnats and forgetting elephants.

Exercise: A lot of secret sauce behind the biz plans you receive? Nope. So, this means that secret sauce is unpromising or, in a few, rare cases, a grand, unexploited opportunity?

Exercise: Venture capital should set aside the few, special cases and go with the broad trends or set aside the broad trends and look for value in the few, special cases?

Lesson: Good secret sauce can be finger lick'n good stuff if you just don't tell anyone.

Lesson: Go for SaaS or a Web app so that your secret sauce and its code can stay locked up inside your server farm.

Third mistake:

> the company had invested well over a year’s effort to get the Best Buy deal underway.

No, clearly the company did no such thing! Instead the company just let its fast talking salesmen, marketing, and bisdev types talk the company into believing that the selling efforts were accomplishing something. No they weren't! All the selling was doing was spending "well over a year" being romanced out of the secret sauce. Marketing people are wont to do things like that!

Final Lesson: Look, if the customer really wants it, then he will be willing to pay a reasonable price to you for it. All that NDA stuff just meant that he just wanted to get your work for free. And you went along with it. Uh, I'm 70 miles north of Wall Street and, due to some highly confidential information I just received, I've got a once in a lifetime opportunity for you -- you won't believe this one -- on a bridge over the East River. Act now!

13 years ago @ Feld Thoughts - Entrepreneurial Density · 0 replies · +1 points

"I’m curious if anyone out there has a real way to calculate this."

Sure: Your equation is just

ED = ((entrepreneurial_emps + students) / adults)

where presumably 'ED' abbreviates 'entrepreneurial density'. So you are just trying to estimate a 'population proportion'.

So, get something that is approximately a 'simple random sample' of the denominator and ask them one or a few questions to see if they are in numerator.

It's an exercise in Stat 101 to say how big a sample you need to get a 'confidence interval' on your estimate as small as you want. If getting the sample is really expensive, then might want to do that exercise. Else, ballpark, ask 500 people and call the result exact or close enough.

For how to get a simple random or representative sample in Boulder and how to ask the question, ask some marketing or political polling firms.

14 years ago @ Feld Thoughts - Cecelia Feld Exhibitin... · 0 replies · +1 points

I already know I like your mom's paintings!

Saint Julien?

With "food and wine"?

Uh, that would be the Haut Médoc with Saint Julien, Saint Estèphe, Pauillac, and Margaux?

Ah, heck: I know that the Haut Médoc is tasty, but I prefer Volney, Pommard, Beaune, Corton, Nuit Saint Georges, Echezeaux, Romanée-Conti, La Tâche, Richebourg, Clos de Vougeot, Morey Saint Denis, ..., Chambertin!

You may remember Nuit Saint Georges and Meursault from the first 'The Parent Trap'! At this point, of those, I have one Corton waiting for my Series A to close and a Meursault with too much dust on it. Ah, should replace the old Meursault with a Montrachet!

Uh, don't tell the customers that Mâcon Blanc is Chardonnay!

Naw, should write a few more Web pages!

14 years ago @ Feld Thoughts - The Power of Why · 0 replies · +1 points

For what he observed, there are better explanations just using well known 'causes':

Cause: As in E. Fromm, people are highly motivated to join groups.

The people who lined up to buy the latest shiny Apple product and the 250,000 who went to the Mall and heard King's speech were joining groups. The "I believe" part made the groups sound authentic.

The groups can be from the lunch table in 'Mean Girls' to the dinner table in 'Titanic' and political organizations up to cults. From a group, a person seeks membership, acceptance, praise, and approval. In some groups people in the group can reinforce their status via gossip.

With a shiny, new Apple product, can assert membership in the Apple cult and use the product for more gossip!

Related is the common motivation to join groups to save X for X one of the whales, the oceans, the rivers, some bird, some fish, the rain forests, the poor schools, the poor children, the planet.

So, maybe a company can assert that they "believe" that their 'inspired goal' is saving X.

A Web 2.0 company can ask if their company wants to be seen as leading a group of some kind, up to maybe a cult: Some people are eager to join such groups, and some people are eager not to.

For that decision, measure twice and saw once.

So, apparently according to Sinek, the CEO needs to be obsessed not with just the product but also with building a cult about the product! So we have one more entry on the long list of candidates for the number one thing the CEO must focus on!

We might test Sinek's idea: What is the "I believe" part of the success of Intel, Microsoft, Google, Yahoo, Facebook, Twitter, Goldman Sachs, JPMorgan, Toyota, BMW, Exxon, Wal-Mart, FedEx, HP, Dell, Rupert Murdoch, McDonald's, the Ford F-150, Pratt and Whitney, Boeing, Applied Materials?

14 years ago @ Feld Thoughts - The Bullshit of Govern... · 0 replies · +1 points


SUCH language, bovine scatology, indeed!

Yup, again, now that you've described your mother's accomplishments and shown her picture, no one's going to blame your language on her!

My language would be much worse than yours.

But, of course! Every business needs a 'business model, and the media has BS! They REALLY like it, and that's what they do: Everything at the 4th grade level except sex at the 10th grade level except anything having to do with numbers at the 2nd grade level or lower.

Some of the Fed stats may actually be okay, but the newsies are grotesquely incompetent at reporting them and don't know and don't care.

So, for

"The first clause says 'U. S. Economy Adds 290,000 Jobs in April.' This means to me that a bunch of people found new jobs in April. A bunch. Yay! Good economy."

Well the clause more likely means that in April the number of people hired was H and the number of people fired was F and H - F = 290,000. So, if F > 0, and I have to believe it is much greater than 0, then nicely H is much greater than 290,000. So, the number of people who "found jobs" is greater than 290,000.

But any such details are way, WAY beyond the 2nd grade and, hence, also the newsies.

In particular they just will NOT give definitions of their terms or quantities. And we don't know how the measurements are done. We don't get graphs of the quantities over time or in comparison with related quantities. Net, we get BS.

The failings of the media go on and on.

Yup, a nice solution is possible. And the solution solves this problem and many more. Working on it. Need to type in some more Web pages!

14 years ago @ Feld Thoughts - Give Your VCs Assignments · 0 replies · +1 points

Brad, you blew it: Showing your mother's accomplishments and picture, no one's going to blame her!

For what a CEO is supposed to do, let's see the ten top tasks, each of which should be at least 90% of the CEO's job:

(10) At the board meeting, have a good answer to the question, "You have $1.5 million in the bank. Now what are you going to do to make money?"

(9) "When you're out of money, you're out of business" -- make sure don't run out of money.

(8) Do well hiring people smarter than he is.

(7) Do well at the work unique to the company and also at 'the vision thing' for the future.

(6) Lead and motivate the workers.

(5) Do well at CEO Job One -- fund raising.

(4) ABC -- always be selling.

(3) Prepare the board meeting info stack and get it to the board members at least a week before the meeting.

(2) Be obsessed with the product.

And for the number one task of the CEO:

(1) Give the board members assignments and give grades at the next board meeting.

Gee, how 'bout just make money? Is it enough to make, say, a lot of money?

For the assignments, how 'bout the board members helping with important work the board meeting was interrupting, uh, making money, e.g., selling ads, or if not that, then helping with the product, e.g., writing software? Uh, do board members write software? A server running SQL Server quit: Repair it, get the data current, get it back online? Shop for dirt cheap commercial space with electric power from two grids and 10 GbE? Plug together servers from stacks of motherboards, processors, memory sticks, disk drives, power supplies, CD readers, and mid-tower cases?

Can't write software? Okay: How 'bout holding board meetings via Twitter?

Lacking that, how 'bout board meetings starting at about 11 PM in a private dining room in the back of a bar, with lots of beer and Victoria's Secret models? SI Swimsuit models? With voting by the board on the better collection of models? Or, with just pictures but extra beer? Or have the board vote on who's prettier, your mother or Sarah Palin? With still more beer!

14 years ago @ Feld Thoughts - Sawyer on The Notion o... · 0 replies · +1 points

Nice review. I wondered about some of that.

Okay, for one step more detail, suppose Brilliant Joe, fresh from 24 hours of hard coding, sitting on his porcelain rest chair, has a brilliant idea for a new do everything super fast algorithm and patents it.

Now Hacker Tom is bringing up a Web site, and on his server uses Brilliant Joe's algorithm: Tom reinvented it, saw it in someone else's code, read about it in an article and reinvented the rest himself, saw the patent itself, downloaded it from a Russian hacker site, or whatever. But, the algorithm, all 200 lines of it, definitely is buried somewhere in Tom's 75 KLOC confidential server side code.

Now, how can Brilliant Joe (1) detect Tom's usage of Joe's algorithm or (2) get evidence of infringement? That is, it looks like Tom is free to bury Joe's algorithm inside his server farm without risk.

14 years ago @ Feld Thoughts - Annual Escalating Pate... · 0 replies · +2 points

Theater of the Absurd: JPEG. MPEG. So, there should be patents on everything Fourier, all low pass filters, all Hilbert space projections (would you believe Banach space projections?), the bass tone control on old hi-fi sets, principle components, all idempotent transformations, all transformations, 2 + 2 = 4?

Software Patents: The whole situation is a FUBAR SNAFU, legal mud wrestling illuminated by burning money.

Entrepreneurial Resolution: Try to bob, weave, stay out of the view of the other guys' lawyers. If they can't see you, then they can't target you. Hope that somehow software patents get tossed into the dumpster. Remember: Don't have to outrun the bear; just have to outrun enough other people trying to outrun the bear.

Short Term Approach: Turn the whole subject into a new river of lawyer jokes. I don't have any just now; don't feel like laughing.

General Situation: Can't expect everything to be tidy; history shows that big things were done, and everything tidy was at best fantasy. Should still be able to do big things.

Ironic Retaliation: All software people should file patents on everything having to do low temperature, air-tight, controlled temperature, water-bath cooking as in posts of user nathanm at!

14 years ago @ Feld Thoughts - Patents Are A Weak Mea... · 0 replies · +1 points

Okay, I read the Myhrvold NYT piece and the HBR article, etc. He's putting lipstick on his pig.

He writes a lot that he should know is nonsense about research, applied research, patents, and business.

E.g., "Invention’s stepchild status is reflected in the way it’s typically funded, which I call the charity model. The entities that provide the vast majority of research funding to U.S. universities -- mostly government agencies like the National Science Foundation, the National Institutes of Health, and the Department of Defense, along with private donors -- do so without any expectation of a financial return. In other words, research grants are gifts, not investments."

Nonsense: It's NOT "charity": Most of the funding is for US national security, and there Congress is getting what they want -- a solid base of research and labor for US national security. The 'applied' part is in layers of additional, large RDT&E organizations, FFRDCs, Lockheed, etc. The NIH funding is, uh, for responding to voters who want progress on major diseases, etc.?

E.g., "The shrinking handful of corporations that still fund long-range research have the same mind-set. Their leaders rarely run research as a business in its own right; instead, they fund it as an act of faith that the ideas produced will somehow create value as they percolate through the product organization."

Nonsense: If they do much research, then it is for "business", mostly as in the current legal patent battles. Also the "act of faith" is mostly nonsense: The business typically severely blocks any contact with the "product organization" because they want to protect their existing, successful lines of business from any 'disruption'.

His goal, more useful, applied research to make strong businesses and improve the world is fine.

But he wants too many chefs in the kitchen: He wants separate groups for research, applied research, products, money making businesses and with 'markets' with patents and lawyers in the middle. Nonsense.

There is NOTHING to stop just one team, or even just one person, from the steps: (1) see market need, (2) do some research to meet the need building on past research, (3) build product, (4) start business, (5) make money.

To improve this, he is correct: Better funding would help. Computing is a grand exception since now a $1000 computer can do so much, but otherwise funding is usually just crucial.

Here's the real funding bottleneck: With the possible exception of parts of biomedical technology, nearly no funding source is able to look at results, say, just on paper, as from NSF funded work or ready for a patent application, evaluate the business potential (for other than patent trolling), write a check, and make money doing this.

Myhrvold implies that he CAN do such evaluations. Okay Nathan, then DO IT, and mostly f'get about the patent lawyers. Eat your own dog food. Uh, Nathan, QUALCOMM had Viterbi's work on coding theory, and he built a BUSINESS out of it. Your company is much different.

Uh, to check out your ability to evaluate, for something I'm not working on now, here's a take-home exam: For zero-day monitoring of server farms and networks, be both multi-dimensional and distribution-free (original research) and close to the best possible Neyman-Pearson result and, thus, save money on both false alarms and missed detections and, BTW, beat the Google, Microsoft, Sun funded Patterson RAD Lab monitoring research effort at Stanford and Berkeley. Build on the Hahn decomposition from the Radon-Nikodym theorem (with von Neumann's proof); right, NOT 'computer science'. Exploit k-D trees to make the computing fast. Sell to the most important farms and networks and then exit to Microsoft (their work in monitoring sucks more), HP, IBM, EMC, etc. Can you evaluate this direction?

Do you even know the Hahn decomposition? Uh, it's not physics. If you can evaluate such work, then post here and we'll talk, also about something much more valuable than zero-day monitoring. Else you are just a big legal effort, i.e., a patent troll wasting the time of the poor HBR readers, etc.

Net, except for patent trolling, no one can make a 'market' on intellectual property assets outside investors can't evaluate. Indeed, part of the crucial value is the ABILITY to evaluate.