The Eigencast

Justin talks to Alan Trefler, Founder and CEO of Pegasystems, about what it means to write code without actually writing code.

They talk about how Alan believes writing code at the level we do today is ‘absurd’ and how he thinks we should be operating computers more like cars. They reminisce about 4GL systems in days of yore, and why they failed.

They dig into what AI means for future computer and human systems, and whether the current crop of systems claiming AI are anything truly great, or just fancy if, then, else scripts.

They talk about how great computers systems, at their best, are tools to aid human understanding and decision making, and that a lot of the grunt work is already automated, and that much more of it really should be.

All models are wrong, but some are useful. What are you using to model your processes?

Chapters

* 00:00:00.000 Intro

* 00:00:15.856 Episode Intro

* 00:03:05.352 Interview

* 00:03:25.163 What is Pega?

* 00:08:34.268 More than a 4GL

* 00:11:06.726 Coding is absurd

* 00:13:06.300 What does using Pega look like?

* 00:16:02.142 Virtual Robots

* 00:19:00.124 All abord the AI roundabout

* 00:23:19.399 AI helping humans

* 00:32:08.691 Computer as support system

* 00:34:33.830 The perils of technology

* 00:36:21.990 Systematic Luck

* 00:41:27.127 Powerpoint Compiler

* 00:43:53.771 Outtakes

Links

* The Pegasystems website.

* Alan Trefler’s Wikipedia page. He really was co-champion of the 1975 World Open Chess Championship!

* The start of the AlphaGo DeepMind rabbit hole. See you in a week!

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

 

Justin talks to Yadin Porter de León, Head of Content Marketing at Druva Inc.

They discuss how technology companies like to push the productivity cost savings from using their technology, and yet those cost savings never seem to involve firing people. Thus they are really about enhancing the value side of the equation, and now about saving cost, but that’s not what they talk about, which is odd.

They talk about how there are lots of average managers, and when they go to cut costs, they often choose highly visible internal things to cut instead of more substantial (and more difficult) savings. A company in a death spiral will be removing free snacks, buying cheap coffee and scratchy toilet paper, rather than getting rid of under-performing managers or making hard business decisions.

The discuss how management is a skill, which requires both theory and practice to get good at. People aren’t just born as great managers, any more than people are born being great at programming in C.

They also talk about the cargo cult of startup management, where people blindly copy the practices of other successful companies without any understanding of whether or not they actually work. Worse, many startups attempt to abandon management altogether in an attempt to avoid the mistakes of others, and then missing out on all the benefits that good management provides.

There’s are reasons for organisations to have managers. Not knowing those reasons isn’t a good reason to try to do without management.

Chapters

* 00:00:00.000 Intro

* 00:00:15.856 Episode Intro

* 00:02:29.295 Interview

* 00:06:14.900 Productivity vs Firing People

* 00:10:04.650 Death Spiral

* 00:11:57.186 Bathroom Judging

* 00:15:31.575 Management is a Skill

* 00:19:55.865 Startups and Management

* 00:24:40.866 Outtakes

Links

* Yadin works at Druva, who make data protection software. This episode was recorded just before Druva presented at Cloud Field Day 1.

* An article in The Atlantic about holacracy at Zappos, and another one in Fortune.

* A collection of Peter Drucker’s best writing on management in The Essential Drucker. Worth the read.

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

 

Justin talks to Lars Herrmann, GM, Integrated Solutions Business Unit at Red Hat, about containers and enterprise automation.

In this wide ranging conversation, they cover how Lars originally got into open source (he was trained as an economist!) and how Red Hat’s business model attracted him to the company.

They discuss how containers aren’t new, and compare and contrast the various ways they’ve been used in the past, and talk about what’s different this time around. It’s not so much about the technology as the business processes that the technology forces you to use.

They talk about the differences between virtualisation and containers, particularly things like duplication, overhead, and security isolation. They discuss how enterprises are stuck using old processes, despite virtualisation allowing for new ones, and how this has created a management headache.

Car analogies are used with much relish, and there’s much discussion of the Toyota Production System (one of Justin’s favourite topics) and how the current IT industry changes mirror those of the automobile industry in the 1980s.

Finally they talk about the rise of the IT factory, and how the IT machine will require very different organisational structures and processes to what enterprises have used before. It’s an exciting time to be part of the industry.

Chapters

* 00:00:00.000 Intro

* 00:00:15.856 Episode Intro

* 00:03:09.881 Interview

* 00:07:13.470 RHEL7

* 00:10:36.528 Containers Aren’t New

* 00:16:47.676 Duplication and Isolation

* 00:20:05.951 Change is Constant

* 00:23:20.151 Technology as Forcing Function

* 00:28:51.166 ITIL for Cloud

* 00:30:59.221 Germans and Car Analogies

* 00:34:29.919 Non-malicious Snowflakes

* 00:39:05.470 Rise of The Machine

* 00:49:59.433 Outtakes

Links

* Get tickets to my webinar Introduction to Financial Analysis here. It’s be recorded, so if you can’t make the time, you can still watch it.

* Check out what Red Hat are doing with containers.

* We talk a lot about Docker containers.

* Learn about Solaris Containers/Zones that predate Docker by a decade.

* There are also BSD jails that are worth learning about.

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

 

Justin talks to Ben Uretsky, CEO and Co-Founder of DigitalOcean.

They talk about how DigitalOcean is one of the only pure play cloud providers in the world, and how it wants to be the default provider for Software-as-a-Service companies worldwide.

They discuss containers, and how we’re about to make a bunch of mistakes with them. “If we thought that VM sprawl was bad,” Uretsky says, “now container sprawl is even worse.”

They also cover how maybe servers and storage, and maybe even network, is possibly an out-of-date way of thinking about how to build applications. Maybe service composition is the way of the future? But then, what do the services run on?

Finally they cover the world of Open Source, and what it takes to make money when you give away your code for free.

Chapters

* 00:00:00.000 Intro

* 00:00:15.856 Episode Intro

* 00:02:50.933 Interview

* 00:07:16.121 Why Add Block Storage?

* 00:13:43.070 What Makes Your Community Special?

* 00:17:53.211 Word Of Mouth

* 00:20:25.081 Software Defined Control

* 00:22:36.419 Autonomous Systems Are The Future

* 00:25:28.662 Levels of Abstraction

* 00:30:01.912 Open Source World

* 00:33:25.127 Community

* 00:36:14.179 Google Platform For All?

* 00:38:43.855 Closing Remarks

* 00:39:45.916 Outtakes

Links

* The DigitalOcean website.

* The new DigitalOcean droplet block storage offering.

* The DigitalOcean Community website, which is chock full of helpful articles on how to configure open source stuff.

* The book Site Reliability Engineering: How Google Runs Production Systems that I refer to.

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

 

Justin talks to Meredith Whalen, Senior Vice President, U.S. Insights and Vertical Business Units, IDC about IDC’s alternative to Gartner’s bimodal approach.

They talk about why integration is a vital piece that is missing from Gartner’s bimodal approach, and why that’s going to cause problems for organisations that try to use it. Justin refers to Simon Wardley’s PST model as an alternative.

They discuss how digital transformation is really just another form of organisational change, so many of the existing tools and techniques will work. Putting digital in front of words doesn’t fundamentally change the universe.

Justin points out that IT departments have been forced to compete with cloud, and for many this is the first time they’ve ever had to operate like a real business. Meredith shares IDC research that indicates some IT departments are beginning to get it, and their businesses are noticing the improvement.

Chapters

* 00:00:00.000 Intro

* 00:00:15.856 Episode Intro

* 00:02:50.933 Interview

* 00:04:14.666 Simon Wardley’s PST Model

* 00:07:04.304 IT Transformation To Help Business Transformation

* 00:07:58.407 What Even Is Digital Transformation?

* 00:10:58.529 The Business Controls The Money

* 00:13:08.929 Cloud Competition

* 00:13:42.186 Are You Ready?

* 00:15:04.301 Where Are The Marketers?

* 00:20:25.638 Change Management

* 00:22:25.759 IT Is Getting Better

* 00:24:35.314 Closing Remarks

* 00:25:38.952 Outtakes

Links

* An IDC community blog by Meredith Whalen about IDC’s alternative to Bimodal.

* Simon Wardley’s blog, Bits or Pieces.

* The Pioneers, Settlers, and Town Planners model.

* The book Accidental Empires, by Robert X. Cringley.

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

Justin talks to Sudheesh Nair, President of Nutanix, at the Nutanix .NEXT 2016 conference, held in Las Vegas, NV.

We discuss the delayed Nutanix IPO, and the Nutanix approach to financing. We talk about the market climate for startups generally in 2016, and how Nutanix may use the tight financing conditions as an opportunity to pick up complementary startups on the cheap. Nutanix PR make it clear these are all hypothetical situations.

Sudheesh explains the underlying principles driving the Nutanix product evolution, and the company culture. We discuss the new Xpress product line, and what going down-market to SMB implies for the company.

We also talk about what Enterprise Cloud actually means, and what companies are really trying to do with infrastructure.

Justin travelled to Nutanix .NEXT 2016 as a guest of Nutanix.

Chapters

We’re experimenting with chapters in this episode.

* 00:00:00.000 Intro

* 00:00:15.856 Episode Intro

* 00:03:33.803 Interview

* 00:05:57.546 Explaining the IPO Strategy

* 00:08:23.640 Jeff Bezos Investor Letter

* 00:13:12.502 Coming to America

* 00:16:40.373 Storage is Central

* 00:22:21.631 The Xpress Experiment

* 00:24:37.085 Enterprise Cloud

* 00:28:58.530 Flattery will get you everywhere

* 00:29:27.601 Closing Remarks

* 00:30:47.462 Outtakes

Links

* The Nutanix website.

* The Nutanix .NEXT website.

* My article at Forbes.com where I analyse the Nutanix S-1 details.

* The blog by Nutanix CEO Dheeraj Pandey that Sudheesh mentions that I was critical of.

* Nutanix Xpress product line website.

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

 

Justin talks to Brian Cohen, CEO of Stratacloud, about the company’s vision to build a software-defined infrastructure platform.

Stratacloud has released its first product: SDI Install, designed to help VARs to deploy NetApp FlexPod.

They also discuss how customers are actually using cloud, and when they choose to deploy into the public cloud, when they use hyper-converged infrastructure, and when they want converged infrastructure, or something else. The actual buying process is quite different to what some analysts would have you believe.

Despite plenty of rhetoric to the contrary, there’s still a lot of money being spent on infrastructure that goes into datacentres.

Links

* The Stratacloud website.

* NetApp FlexPod.

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

 

Justin talks to Matt Benke, CEO and Founder of Spare5, and Andy Ganse, Principal Data Scientist, about how they use humans to complement machine learning and data analysis techniques.

Spare5 is so-named because people can take a spare five minutes out of their day to do a task allocated to them by Spare5’s system, a bit like Amazon’s Mechanical Turk, but a task that they are good at and enjoy. And unlike playing a game or checking in to Facebook or Twitter for the twenty-third time this morning, you get paid for your time.

But unlike the exploitative nature of other platforms (*cough* sharing economy *cough*), Spare5 doesn’t see this as an alternative source of a primary income. It’s very much geared towards people spending an odd minute here and there to do something they actually like doing, not something they feel they have to do just to survive. It’s an alternative to playing Candy Crush, but you get paid instead of paying to be entertained.

There are some interesting implications to Spare5’s approach, for how machine learning and modelling is done, the role of humans in a machine/human hybrid future, and how we design systems to provide end companies with the benefits of human labour but in a way that values the humans at least as much as the machines.

Links

* The Spare5 website.

* Amazon’s Mechanical Turk service.

* The original Mechanical Turk to which Amazon’s service owes its name.

* Coverage of Microsoft’s Tay bot incident.

* IBM Deep Blue.

* IBM Watson wins at Jeopardy! [YouTube video]

* Getty Images photo site.

* Google’s AlphaGo project.

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

 

Justin talks to Scott Dietzen, CEO of Pure Storage, about keeping a startup culture inside a public company. They discuss the storage market, and all-flash in particular, given the nature of Pure’s products.

They talk about why Pure chose to develop its newly announced FlashBlade product line as an “internal startup”, and how the group was organised, managed, and motivated. Scott shares his views on how product development inside large companies should be done, and there are some good lessons here for anyone interested in innovation, startups, and high tech.

Scott shares his experiences of entrepreneurship from four different companies, and suggests that Pure Storage will be the last time he plays a hands-on role in leading a startup. He provides a few tips for anyone thinking of trying the entrepreneurship route.

Links

* Pure Storage

* The Geek Whisperers, from which Justin stole borrowed Amy’s favourite question.

* Justin’s coverage of the FlashBlade announcement at Forbes.com.

Sponsors

This episode of The Eigencast was sponsored by PivotNine. Research, analysis, advice.

 

 

Justin talks to Howard Marks, Founder and Chief Scientist at DeepStorage about his predictions for the enterprise storage industry in 2016.

They talk about how research is performed, how it gets reported, and how to interpret what is written. They discuss how large analyst firms shape the very markets they research, and how vendors have little choice but to bend to their will.

Howard talks about how he believes Gartner and IDC have got it wrong about how the all-flash array market is defined, and how no one should be worried about where to put VDI any more.

Justin and Howard both whinge about comparisons between brand new technology and stuff that was bought five years ago.

Links

* DeepStorage.net, Howard’s company’s website.

* The Greybeards On Storage podcast.

* Chad Sakac’s blog on the four categories of storage system architectures.

* The Dunning-Kruger effect, named after the effect described in a paper titled Unskilled And Unaware Of It. Try this Google Scholar search for PDFs you can download for free.

* Previous Eigencast episodes mentioned in this show where Justin talked to Greg Kniereman from HDS: The Eigencast 007: HDS Goes All Flash and The Eigencast 008: Analysts and Decision Making.

Sponsors

This episode of The Eigencast was sponsored by PivotNine IT Consultants and Advisors.