Sunday, October 26, 2014

Rescale, a textbook Zero to One startup

I visited last week in San Francisco the offices of Rescale. About 30 people are cramped in tiny office in an old  nondescript two story building. The tiny parking lot asked for $50  and I kept driving to arrive late at my appointment after parking a mile away.

I chatted with Joris Poort, founder and CEO. He says he worked for four years at Boeing as a structural and software engineer on the 787 program, optimizing the design of the tail and wings. Yet this is not the end of the story. I read on the website,
Joris holds an M.B.A. with distinction from Harvard Business School, an M.S. in Aeronautics and Astronautics magna cum laude from the University of Washington, and a B.S. in Mechanical Engineering and minor in Applied Mathematics magna cum laude from the University of Michigan.
The is an unique combination of engineering skill, computer science and top business school thinking. This Harvard connection tells many things.

Ben Horowitz, in his book "The  Hard Thing about Hard Things"  tells about the Netscape recruiter who rejected him as an ideal candidate because he was not a graduate from Stanford or Harvard. It took a miracle to get the the job, directly from Marc Andreeseen, who was 22 at the time.

This was not Joris' challenge, as he was coding Rescale from home in San Francisco. before opening an office.

 Just look at the list of investors in Rescale . They are listed alphabetically, but see Richard BransonPeter ThielJeff Bezos, just a small sample from the 38 legendary names listed as investors  .
Rescale offers engineering simulations - traditionally an HPC domain  hard to use. Rescale has done a lot to make simulations easier to use - although there is lot more to be done in terms of ease of use in HPC.

Can Peter Thiel,  a  J.D. from Stanford Law School use Rescale? Probably not yet, not because Rescale hasn't done enough  to simplify the simulations, but because it is very hard  to simplify.

Sample Rescale Resources page

Yet Rescale is a major step in the right direction, clearly explained be Peter Thiel in his #1 entrepreneurship book on both Amazon and New York Times.  In the Atlantic, Derek Thompson writes "Peter Thiel's Zero to One Might Be the Best Business Book I've Read"

This is what Peter Thiel says right in the introduction.
the VERY MOMENT IN BUSINESS happens only once. The next Bill Gates will not build an operating system. The next Larry Page or Sergey Brin won’t make a search engine. And the next Mark Zuckerberg won’t create a social network . If you are copying these guys, you aren’t learning from them. Of course, it’s easier to copy a model than to make something new. Doing what we already know how to do takes the world from 1 to n, adding more of something familiar. But every time we create something new, we go from 0 to 1. The act of creation is singular, as is the moment of creation, and the result is something fresh and strange. Unless they invest in the difficult task of creating new things, American companies will fail in the future no matter how big their profits remain today. What happens when we’ve gained everything to be had from fine-tuning the old lines of business that we’ve inherited? Unlikely as it sounds, the answer threatens to be far worse than the crisis of 2008. Today’s “best practices” lead to dead ends; the best paths are new and untried.
He shows these pictures:

Peter Theil is credited to the idea of creating not only from nothing, but implicitely assuming highest risk portfolio startups. Yet Peter did not come out from a vacuum.

I learned from Professor Rashi Glazer from Berkeley -  Product Management Module III. Haas School Business, May 18, 2006 - that competition is everywhere, and the best way to neuter and minimize the competition is to select the highest risk, highest rewards portfolios. If we want to maintain the Silicon Valley leadership in the world, we must create companies no one else dares to start anywhere else in the world.

I summarized this ideas in My Product Management Manifesto entry in my blog. Quote:
The charts  above summarize what I learned in UC Berkeley seven years ago. Berkeley is always at least ten years ahead of other people time.  We have to liberate ourselves from the chains of numerical Return on Investment (ROI) and forecasts. Financial numerical decision makers do not work in Breakthroughs.
So Rescale is a textbook zero to one company. No one made money in HPC yet, Soon someone will makes tons of $. I think BoscoR falls in the same category

I will review the book in a future blog.

Wednesday, October 22, 2014

Value Propositions That Work

I came across a blog from Harvard Business Review with the title above. It is written by Anthony K. Tjan   I think this one of those no-nonsense blogs that are amazingly useful. This is what Tony says:

"Most people can’t explain what their company does — its value proposition. Consider that there are only four types of consumer benefits that matter and by extension only four categories of value propositions that work."
  1. Best Quality: Think of brands that set a standard
  2. Best bang for the buck:  Some consumers will always buy on price. Best-in-class value doesn’t always mean lowest price, however, but rather the best quality-to-price ratio. 
  3. Luxury and aspiration:  On the other end of the spectrum from bang-for-buck players. In High Tech this needs a translation. Apple tablet are fantastic, expensive but largely unnecessary . The latest entrant in Tesla Motors, with a $100k Ferrari killer electric car. 
  4. Must-have: Again in high tech it needs translation
No one can say this is new. What is new is combining all together, or as many as possible from the four value proposition in your company and product.
"Figure out how to reposition your offering. Stop being stuck in the middle and aim to set a new standard."
Tony Tjan with Arriana Huffington

Monday, October 20, 2014

My reading of The Hard Thing About Hard Things, by Ben Horowitz

 I read Ben Horowitz book like a Roman à clef  to learn its secrets from what he says and what he doesn't say. The real message, is in the white spaces between the letters. The essence of what Ben writes is not at the beginning, nor at the end of  of the book. It is in the middle, on page 212:
In all the difficult decisions that I made through the course of running Loudcloud and Opsware, I never once felt brave. In fact, I often felt scared to death. I never lost those feelings, but after much practice, I learned to ignore them. That learning process might also be called courage development process....
Over the past ten years technological advances have  dramatically lower the financial bar for starting a new company, but the courage bar for building a great company remains as high as it has ever been
Here are some samples sub-titles of chapters to see what Ben means by courage development process:
  • The right way to lay people off
  • Preparing to fire an executive
  • Demoting a loyal friend
  • When smart people are bad employees
  • ...
  • Old People
The last is a sub-chapter in Chapter Six, page 170 which brought me shivers. Ben uses the word "Old" only in the title, The rest he calls them "senior" people. This is what Ben writes on page 175
Aw Man, you sold your soul: hiring your first senior people into your company, may feel like selling your soul, and if you are not careful, you may well end up selling the soul of your company. You take risks and you have to win your race against time. This means acquiring your best talent, knowledge and experience even if it requires dealing with some serious age diversity.
71 year old woman skydiving 
Ben's  grand-father Phil Horowitz was a "card carrying Communist" and lost his teacher job during McCarthy era.

I am from Romania and  grew up with Communism as a state dogma. Most of my "communists" friends were also one of the finest people and were as disillusioned with the regime as we were.

Ben Horowitz met Marc Andreeseen when he applied for a job at Netscape. He was already married with  three children, including daughter Mariah who is autistic (like my son).

Figure 3: Ben and his wife Felicia Horowitz. They met in a blind date in 1986
The recruiter rejected Ben as an ideal candidate because he was not a graduate from Stanford or Harvard. Then he met Marc who was 22 at the time. The two developed natural  bond and Marc got the job. This is the miracle #1.

Marc was ahead of his time. Just read the Employer Creed to see what I mean.

Marc changed Ben's life for ever.
The launch of Netscape did no go smooth. Here are Ben and Marc email exchanges, under stress.

Figure 4: Ben email to Marc
Figure 5: Marc email to Ben
This would worry anyone and got worried Ben. He was 29 and had a family and many responsibilities.  He  showed these emails to Felicia and she said: "You should start looking for a job right away."

But the miracle #2 happened. Ben was not fired. He writes: "more shocking, Marc and I became friends; we have been friends and business partners ever since."

These two miracles can never happen on demand.

Chapter 9 The End of the Beginning

The last chapter finally reveals the key of this book: how the idea of Andreeseen Horowitz (  was born.

While a founding CEO of Loudcloud,  David Beirne a partner at Benchmark Capital jumps in a meeting and asks Ben this question: "When are you going to get a real CEO?". This was in front of all co-founders, humiliating him.

Ben says:
Not a day went by, when I didn't think of the interaction with David Beirne...  Marc and I discussed this paradox often We wondered aloud why as founders we have to prove our investors that we can run the company, rather than our investors assuming that we would run the company we created
The revolution was partly made possible by Marc and Ben experience with a VC like David Beirne. I looked him up on The Funded 

Figure 6: See David Beirne rating as a partner (1.8)
Personal Specialty: None
See Andreeseen Horowitz as 4.1 top rated fund
Screen shot taken October 20, 2014
Marc and Ben looked at over 800 venture capital firms. Only about 6  (0.75%) of them had consistent returns. This means the overwhelming 99.25% of the VCs had no consistent returns

Marc and Ben realized that it is easier to make a CEO from a technical founder, rather than bringing  in the so called "professional CEO". When Adreeseen Horowitz started the great majority of VCs "were better designed to replace the founder, rather than help the founder grow and succeed"

Therefore  identified key deficits and should address, while creating itself as a new breed of capital funding technology company - new and refreshing:
  1. The CEO skill set
  2. The CEO network
    (Large companies, executives, engineers, press and analysts, investors and acquirers)
I did write before about romanticizing the startup in San Francisco The best and the worst companies are startups, but the best are usually remembered. The others disappear.

In a way, The Hard Thing about Hard Things is written as a response to a generalized David Beirne mentality. But it is also written for the technical CEO skill set. For example, DataBricks CEO is Ioan Stoica a distinguished researcher, I suppose this book is written for him.  They support Apache Spark and soon probably Mesos. This is a big data, not social network field, with some serious scientists involved. I am familiar with this field, as you see from my blog on BoscoR. University professors are, as a rule, are notorious for their reluctance to risk. Most keep their academic position in parallel

Can one teach them entrepreneurship?  Ben Horowitz warns us right at the beginning in the introduction. Business book recipe-like books are mostly useless. The real life challenges have no recipes:
I do not attempt to present a formula in this book. Instead I present my story and the difficulties I had faced.
Just patterns that one discovers.  However each time Ben had a challenge as a CEO, he had no fall back alternatives. So it is hard to have a CEO completely succeed, when one cherishes the safe, lukewarm office of the tenure for life as a plan B,

To me, being a mensch, should be first and foremost goal. Here is the definition of a mensch, for easy reference.:
Mensch (מענטש) a Yiddish word that means "a person of integrity." A mensch is someone who is responsible, has a sense of right and wrong and is the sort of person other people look up to. In English the word has come to mean "a good guy." Menschlichkeit (מענטשלעכקייט) is a related Yiddish word used to describe the collective qualities that make someone a mensch.
Andreeseen Horowitz is in essence, an example of Menschlichkeit.  This is the real treasure underneath the apparent toughness  the book projects to a superficial reading.

Figure 7: The Rainmakers (click to enlarge)

Thursday, October 09, 2014

BoscoR: Extending R from the desktop to the Grid - accepted at SC'14

A pleasant surprise: The paper BoscoR: Extending R from the desktop to the Grid has been accepted to 7th Workshop on Many-Task Computing on Clouds, Grids, and Supercomputers (MTAGS) 2014 to be held at SC'14 in New Orleans, November 16 to 21, 2014

I am one of six authors, - Derek Weitzel is the lead.  Dan Fraser, Marco Mambelli, Jaime Frey and David Swanson are the others

We did focus on end users,  A key design goal  was to have a flat learning curve for any laptop R user who tries BoscoR. We decided to do an evaluative analysis of R researchers,  in addition to the quantitative analysis. Otherwise we risked to have an almost perfect software, that nobody uses.

In High Performance Computing (HPC), ease of use is not usually the main goal. I want to thank the peer reviewers who accepted this new, gate opener focus in high performance / throughput computing, We care about users unfamiliar with grid, clusters, clouds and dynamic data centers. We want them to be  productive, while enjoying and creating new habits with our high throughput computing applications.

What is R project for Statistical Computing?

The  R is a free software environment for statistical computing and graphics. It compiles and runs on a wide variety of UNIX platforms, Windows and MacOS. Quoting from our paper's abstract, BoscoR
... is a framework to execute R functions on remote resources from the desktop using Bosco.  The R language is attractive to researchers because of its high level programming constructs which lower the barrier of entry for use. As the use of the R programming language in HPC and High Throughput Computing (HTC) has grown, so too has the need for parallel libraries in order to utilize computing resources.
Fig. 1 Rexer Analytics reports explosive usage of R among data miners
This explosion of R usage made headlines even in New York Times, Data Analysts Captivated by R’s Power. Here is an interesting quote
R first appeared in 1996, when the statistics professors Ross Ihaka and Robert Gentleman of the University of Auckland in New Zealand released the code as a free software package.
According to them, the notion of devising something like R sprang up during a hallway conversation. They both wanted technology better suited for their statistics students, who needed to analyze data and produce graphical models of the information. Most comparable software had been designed by computer scientists and proved hard to use.
Lacking deep computer science training, the professors considered their coding efforts more of an academic game than anything else. Nonetheless, starting in about 1991, they worked on R full time. “We were pretty much inseparable for five or six years,” Mr. Gentleman said. “One person would do the typing and one person would do the thinking.”
Right from the beginning, we learn that the statistics professors are in general not computer scientists. One can not be the jack of all trades.

A web page,  CRAN Task View: High-Performance and Parallel Computing with R,  shows that existing R parallelism is mostly confined to single server or laptop. Quoting from our paper
Data mining requires computational resources, sometimes more computational resources than can be provided by their desktop. In a recent study, “Available computing power” was the second most common problem for big data analysis.
In addition, the respondents stated that distributed or parallel processing was the least common solution to their big data needs. This could be attributed to the difficulty of processing data with the R language on distributed  resources, a challenge we set out to solve with BoscoR.
A reason that distributed computing is not seen as a popular solution to big data p rocessing is that scientists are more familiar processing on their desktop than in a cluster environment. R is typically used by people that have not used distributed computing before and do most of their analysis on their local systems with IDEs such as RStudio . Users are unaccustomed to the traditional distributed computing model of batch processing in which there is no interactive access to the running processes.

What R researchers say after using BoscoR

Another  quote from the paper:
During follow up interviews with users after using BoscoR, we received many positive reviews of the framework. Improving the user experience of using R on distributed resources was a primary goal of BoscoR. One example of a positive review was from a Micro-Biology researcher from the University of Wisconsin:
I have a huge set of data, which I have to split into pieces to be handled by each node. This is something I can do with the ”grid.apply” function. This reduces the submit time from several hours, to several seconds... it is a phenomenal improvement.
This will greatly increase my use of grid computing,  as right now, I only use grid computing when I have no other choice. 
I was one of the authors who took this interview . Here are more details:
So,   here is deal. Currently, if I want to do something in super-computing high throughput, I use the classical resource manager and I have a several hours barrier. There is something I want to run right now and I have been postponing it. 
Days go by and I don't do anything.
Now I am able to do with "grid apply" BoscoR the exact  same thing I am doing with the classical resource manager currently,  I will use BoscoR all the time, every single day. 
So here is what "easier-to-do" means: new habits, pleasure, less tedious work, more creativity, more satisfaction, more joy and feeling more power to do things I, the user, never been able to do before.

This is where technology transforms into sentiments in people When I drove a Tesla car, all I felt was a driver exhilaration and forgot about all the technical marvels inside

How BoscoR does it

It all started with the work of Derek Weitzel who created a software called Campus Factory in 2011. This is a distributed high throughput computing framework to link clusters without changing security or execution environments. These were ones of the main stumbling blocks in accessing and doing work on remote clusters.

This Campus Factory lead to the idea of Bosco,  designed to be run without administrator intervention. It automatically installs the Bosco client on a remote cluster. SSH was chosen as the protocol since it is used nearly universally for cluster access. Jobs are submitted to Bosco, which then submits over ssh to the remote cluster. Input files are transferred over the SSH connection as well. Bosco then monitors and reports the status of the job on the remote cluster as idle, running, or completed. Once the job is completed, Bosco will transfer output files back to the submit host.
Fig. 2: Architecture of Bosco
This is great, but Bosco at the beginning aimed to be coupled with any application. We did customer interviews face to face and we noted that researchers do not want to access command lines. They  looked straight in our eye and said clearly through words and body language: "We want to use the same graphical user interfaces we are accustomed to use in our day to day work".

Looking at various applications, we discovered R. It is open source and has an estimated two millions plus users. Another author, Dan Fraser traveled to Spain at the R-users conference and discovered an unusual interest for a Bosco version specially made for R users.

We combined Bosco with another module that we updated, called GridR
Fig. 3: GridR input generation
We explain this in the paper:
GridR was modified to detect, and if necessary install, R on the remote system. This was performed by a bootstrap process.
In the GridR generated submit file, the listed executable to run on the remote system is not R, but the bootstrap executable. The bootstrap executable detects if R is installed on the remote system. If it is installed, it simply executes the user defined function against the input data. If R is not   installed, the bootstrap downloads the appropriate version of R for the remote operating system. The supported platforms are identical to Bosco’s, CentOS 5/6 and Debian 6/7. R is downloaded from a central server operated by the OSG’s Grid Operations Center
Now you tell me; isn't it cool?

Job Submissions. Direct versus Glidein

You can read the paper to see the exact definitions, but here is an interesting plot
Fig. 4 Comparison of job submission methods
Direct and glidein submission methods have approximately similar workflow runtimes. But, for short jobs, glidein has significantly shorter workflow runtimes.This can be attributed to the advantages of using a high throughput scheduler over a high performance scheduler.

A future startup?

I am convinced BoscoR can become a great tool to bring science beneficial to society and creating wealth very much like the many successful Silicon Valley startups.

However this depends on the commitment of the engineering potential founders, a delicate issue in Academia


  • To all the Bosco team members, who made such a difference in working and enjoying my life at the same time
  • To Open Science Grid for financing this work
  • To all other wonderful people I met at OSG , XSEDE, FermiLab

Wednesday, October 01, 2014

The amazing Ello: a new style of social network

There is a new breed of social network, coming from Vermont. The founder is Paul Budnitz, who also owns a bicycle shop Here is their manifesto, written by Aral Balkan

Ello Manifesto 

Your social network is owned by advertisers. 
Every post you share, every friend you make, and every link you follow is tracked, recorded, and converted into data. Advertisers buy your data so they can show you more ads. You are the product that’s bought and sold.
We believe there is a better way. We believe in audacity. We believe in beauty, simplicity, and transparency. We believe that the people who make things and the people who use them should be in partnership.
We believe a social network can be a tool for empowerment. Not a tool to deceive, coerce, and manipulate — but a place to connect, create, and celebrate life.
You are not a product.
Aral Balkan though left Ello, because he discovered that Paul accepted VC money.  "What I didn’t know at the time was that they’d taken $435,000 in seed funding from FreshTracks Capital, a Vermont-based Venture Capital firm. This had apparently been announced by FreshTracks Capital in March.
Paul Budnitz, Ello CEO
Here is some more information from the web site 
Ello doesn't sell ads. Nor do we sell data about you to third parties.
Virtually every other social network is run by advertisers. Behind the scenes they employ armies of ad salesmen and data miners to record every move you make. Data about you is then auctioned off to advertisers and data brokers. You're the product that's being bought and sold.
Collecting and selling your personal data, reading your posts to your friends, and mapping your social connections for profit is both creepy and unethical. Under the guise of offering a "free" service, users pay a high price in intrusive advertising and lack of privacy.
We also think ads are tacky, that they insult our intelligence and that we're better without them.
According to article from Canadian media, Ello was " created by a group of seven artists and programmers, it’s still in a beta phase as it works out bugs and introduces new features. It’s also only available through personal invites, and is fielding a reported 31,000 join requests per hour."

Per hour!!! No wonder my request to join from yesterday got no answer

Some call it a Facebook killer as" Ello differs from Facebook and Twitter in its “no ads” policy, and it doesn’t require you to register with your real name–contrary to Facebook’s recent decision to enforce such a policy."

It remains to be seen how it will keep its pure innocence Meanwhile the manifesto produces up to 50,000 new subscriptions an hour. One day they must figure out how monetize without alienating their idealistic users, if the company is not sold before that.

Note added January 28, 2015

Money talks. The $435K seed investment grew to $5.9 million three month ago  Was Ello bluffing? There is no way to survive in the United States of America without funding, and without profits to justify the investors hopes. Revisiting the web site, this is what we read:
Ello is completely free to use.
We occasionally offer special features to our users. If we create a special feature that you really like, you may choose to support Ello by paying a very small amount of money to add that feature to your Ello account.
Oh well. This is just the beginning

Blog Archive

About Me

My photo

AI and ML for Conversational Economy