Category Archives: Product Development

A Grand Circle Jerk

I have neglected including a song at the beginning of each of my blog posts. For this post please consider this tune.

The rise and recent fall of the bio-tech firm Theranos is yet another example the challenges inherent in the valley’s startup business culture and the media echo chamber which supports it.

Theranos, founded in 2003 by Elizabeth Holmes while she was a student at Stanford became  a poster child for both disruption and all that that is insanely great about the valley. Theranos was going change the world. Hmmm where have we heard that before ?

Under Holmes’ leadership Theranos was developing a more efficient blood test. The test could reportedly provide the requisite chemical analysis in support of 30 or more  different blood tests using only a single drop of blood obtained through the pin prick of a finger. This test was supposed to enable blood testing services in just about any location. These efficiencies in turn were going to empower the average consumer with information which could ultimately lead to better health and wellness outcomes.

There were early signals that Theranos’ technology may not have been delivering the results originally claimed. The company, mid to late last year admitted to backing away from the technology upon which it’s original value proposition was predicated. For some reason these early signals did not result in any serious re-evaluation of the company.

Entrepreneurs by nature, are  optimistic and fiercely committed to their ideas. Holmes was no exception. Unfortunately she also believed her company could ignore common practices around peer reviewed work in the health and science fields. This thinking finally caught up with Theranos. Earlier this year, the FDA and The Center for Medicaid and Medicare Services (CMS), the organization which certifies providers for the Medicare and Medicaid determined the company’s claims and lab practices questionable. CMS has taken actions to de-certify the company. The FDA and CMS  have gone as far as to ban Ms. Holmes and others in the company from running any sort of lab for two years effectively removing Theranos from any sort of relevance in the near term.

Going forward it’s not clear what is to become of Theranos. Maybe lady luck will strike and Theranos will survive to prove the value and efficacy of its original technology. Maybe not. Ms. Holmes in all likelihood will lay low and find herself a gig as an  EIR at one of the VC firms here in the valley. Certainly this is not the last we will see of Ms. Holmes in the valley based on her ability to raise capital alone.

What is certain is the more Silicon Valley (and the industry friendly press which writes about it) claims to be the purveyor of only the best,  ideas, best in breed  “power of the market and merit” community, and ethically evolved in contrast to the rest of the U.S.  business sector the less it seems to ring true.


So what exactly did Facebook Open Source for Deep Learning HW

Facebook recently announced it was open sourcing it’s AI/Deep Learning HW design, Big Sur.

I am not a member of OCP so it’s not clear I can get all of the documentation. However, from the way the industry media is falling over itself heralding this act of open sourcing it must be something compelling right.

In all of the articles I have read so far I have seen nothing of new processors, switch fabric ASICS etc… That is not to say that motherboard level design is trivial however I would have expect something more significant based on the column inches I have pawed through in the last two days

Maybe someone can help me understand what I am missing

Building and SELLING pschidt is not easy

First, the song, Devo’s rendition of “Working In A Coal Mine”.

I don’t know Elon Musk, nor have I ever worked at Tesla. Most interesting to me about Musk and Tesla is how they have managed NOT to take over the car industry, despite the quality of their product. In fact, over the summer an article from Yahoo News indicated that Tesla might be losing 4K/ copy of the Model S.

I point this out NOT to criticize Musk, Tesla or their efforts, instead I just want to underline how difficult it is to design, build, distribute and sell a product. Especially a product of this scale. Nonetheless and despite the challenges Tesla forges on.

I am not sure what their end-game is, and I am not necessarily a fan of electric cars. It’s not clear to me that e-cars are the most innovative solution for transportation and managing global warming. However , I do admire their trying.

In the immortal words of “Lonesome George: Ride on Josephine, ride on”




Wither the MVP, rejoice in the old fashion art of product development

Music: Why must I chase the cat

Read this article  by Bill Aulet which was posted in Venture Beat. While phrases like “Minimum Viable Product” and “Lean Startup” are in vogue these days. Aulet takes us back to yesteryear. Back to a time where a product was a combination of assessing the market, querying potential customers, and most importantly, having some significant vision.It seems that lately, the “product” cycle is more about rapid iteration. I won’t deny that this approach facilitates some forward progress. However I need to also comment that it does tend to drive things in the direction of micro-iteration, which then consumes development resources in ways which generate lots of lines of activity in support of “features” while  at the same time taking resources away from deliberations and thinking which might foster the development of a truly innovative product and/or additions to them.

I liken it to the hardware world when the first PALS (Give me a 22V10 or give me death) and FPGA’s (Field Programmable Gate Array’s) came on the scene. Both of these were a type of hardware device which could be programmed to function in whatever way the designer desired. If, after programming, you didn’t like what it had, you could “erase” it and try again. This was a boon for hardware developers. Gone were the multi-month waits after taping our chip to get an actual piece of physical hardware back.You could literally try two or three implementations of a design out each day.You could fix bugs rapidly. You could add a “feature” in as fast as you could think it up. Oh my, “Slap me, I must be dreaming”, or so we thought.

That’s where the problems began. It was precisely because you could iterate rapidly (in hardware development terms) that features were added which maybe weren’t necessary even thought they added to the overall feature count of the product. Worse still, there was a marked drop in quality as the game became about spinning and getting something into the lab quickly to “show”. What is now called an MVP I supposed

In Mr Aulet’s piece, he points out how  becoming enamored with getting something built, coded etc.., can get in the way of the real work of true product development.To wit, you become a slave to the release cadence.

It’s great to see the software side of the business get some exposure and lead the charge forward. I hope part of the charge forward includes reflecting on lessons learned from others. One can only hope…