How has the recent turmoil within the OpenAI offices changed your plans to use GPT in a business process or product in 2024?
Increased uncertainty means we are more likely to evaluate alternative AI chatbots and LLMs.
No change in plans, though we will keep an eye on the situation.
With Sam Altman back in charge, we are more likely to go all-in with GPT and LLMs.
What recent turmoil?
Cloud Native Ecosystem

IBM’s Angel Diaz: No Cloud is an Island

May 3rd, 2016 11:54am by
Featued image for: IBM’s Angel Diaz: No Cloud is an Island

As much as anyone we know within the ranks of Big Blue, Angel Diaz is the face of the new IBM. With seemingly boundless energy, Diaz is a new generation of IBM executive, one spirited and personable, tirelessly networking at trade shows, evangelizing about APIs, about the IBM Watson cognitive platform, and about the company’s new devotion to developers, which IBM recognizes are the new gatekeepers to the enterprise.

At the OpenStack Summit in Austin last month, we interviewed Diaz. He was enthused but sleepy, still a bit groggy from a late night of networking at The Container Bar, an establishment he had just heard about upon his evening arrival to Austin. It is a bar made from shipping containers. “I mean, you’d have to go to that, right?” he asked.

Diaz is IBM’s vice president of cloud technology and architecture. Diaz and his team oversee the technology, architecture, and strategy behind IBM’s hybrid, open and secure cloud. He is also IBM’s point person for open technology. He started out at IBM working at IBM Research and had a hand in developing many of the open standards we use today on the Web.

We wanted to find out more about IBM’s involvement in OpenStack and Cloud Foundry, about how IBM is helping companies grapple with the new development methodologies brought about by open source. And we couldn’t help asking about Diaz’s rich history with technology.

TNS: How did you first get into technology?

Diaz: So it happened this way, I grew up on a farm in Puerto Rico. And this was in the late ’70s, early ’80s, I was young, right? I had to follow the cows as they got milked, cleaning their poop, which is a standard job before you went to school. And I was like, “You know what? There’s got to be a way to automate this. I really should not have to be doing this, right?” I wanted to build a robot, an artificial intelligence. So I started getting really into that. I would go to the university as a sixth grader and read books about AI from Marvin Minsky. I didn’t understand them, but I was fascinated with this, with the technology and making our lives better.

I had some really strong mentors on the sciences side of things as I was growing up. I built my first computer, a Timex/Sinclair, and reprogrammed the entire operating system because I needed more RAM. Long story short, skipped a couple of grades, I went to RPI Rensselaer, got a Ph.D. in computer science and went right into IBM Research in ’97. I worked with a famous Austrian computer scientist, Erich Kaltofen.

At the time, it was the beginning of connected [computational] power. Computation will take forever with one machine, but I could have ten machines, a hundred machines, a thousand machines across universities, across the world. How do you connect these computers together? Around that time, I had been working with [CERN’s] Anders Berglund. We’re working on the web, HTML. I was doing the math, and he was doing physics. We started working on various projects that became web standards. The things that I always cared about was the semantics, the meaning of things because to compute with mathematics, you need to understand it. So I coauthored many of the original web standards, markup standards, Things like the DOM, which everyone uses in JavaScript. I was on that committee. I was involved with Cascading Style Sheets. I helped create XSLT and XSL-FO. I wrote the first use case for XML, which Jon Bosak was doing. We wrote the very first XML parser ever and put it out there. It was in actually C++, and we had to rewrite it in Java for people to use.

So I got really excited about the notion of multi-modal. How do you do markup for people who can’t see, right?  So how do you combine text and speech? So I ran IBM’s Natural Language Group, which is a big part of Watson now. Watson is a lot more than that, but you have to understand the language and convert it to a tokenized form, which then can be analyzed, right? That’s a very hard problem because you got to train systems and you have to understand the language.

Angel Diaz speaking at the IBM Interconnect conference in February.

Angel Diaz speaking at the IBM Interconnect conference in February.

Anyway, then Steve Mills was running software, and he had asked me to help him, along with a bunch of others, build the next generation transaction systems. And that’s kind of how it all happened.

Nothing in my life was planned, it just kind of happens. I’m just going where the flow is.

TNS: So what brings IBM to the OpenStack Conference? What’s the OpenStack story lately for IBM?

Diaz: Back in about 2000, we certainly knew that cloud computation, compute storage and networking were being democratized, and there is a potential for an internet-like explosion of innovation. Once you do that, stuff happens that you can’t imagine.

So we had a choice to make, around 2010, to start building out our cloud either on proprietary technology or open technology.  An IBM Fellow, Dave Lindquist, and I were talking to Steve and everyone said, “Wow, look at all these communities.” There’s a lot of good cloud stacks. We could start our own but why don’t we just join? Why don’t we help?” I might be able to galvanize the community because it was very fragmented back then.

It was like maybe ten potential projects. With OpenStack, so I met with Jonathan Bryce. He was at Rackspace at that time. What was great about OpenStack was the code maybe wasn’t the best, but the community had quite a lot of users then. You know, whenever you look at open source, users trump code. The code will get better. But the vibe, the way that we used to call the OpenStack way was great. It was a small community. So we helped stand up the OpenStack Foundation in 2012.

And by the way, we’re doing the same thing with Cloud Foundry. We’ve got the Cloud Foundry Summit coming up. We’re expecting between 2,500, 3,000 people there. It’s growing faster than OpenStack. We are creating these centers of gravity that make up a cloud architecture, an open cloud architecture.

So why do we do it? It is about choice with consistency. We want our clients to be able to build their applications whether they’re focused on infrastructure or platform,  and have that same thing work on any of clouds. No cloud is an island. It is a hallucination that there is just one public cloud vendor or one public cloud. Of course, there’s going to be private clouds. And of course, it’s going to be increasingly more public clouds and from multiple vendors.

TNS: Right. So you’re always going to have multiple clouds?

Diaz: What makes the web work? Interoperability, right? Interoperability through code, through APIs, through testing, right? That’s what makes it work. When you type in or ESPN or CNN or Macy’s, whatever, you don’t worry about whether it’s going to work. It just works, you don’t care what web service is on the other side, you don’t care what browser you’re using for the most part.

By the way, back in those days, I used to sit in a room with Netscape and Microsoft, and since IBM didn’t have a browser, I was kind of the monkey in the middle. They were fighting over the blink tag. I know it sounds funny but let me tell you, to them that was a differentiator back then.

OpenStack contributions by companies, from Stackalytics

OpenStack contributions by companies, from Stackalytics

So whether it’s Cloud Foundry or OpenStack, it is all about interoperability. We are one of the largest contributors to OpenStack. We have a large number of the project leaders. We have hundreds of developers who are 100 percent developing code for OpenStack.

TNS: That’s their full-time jobs…

AD: Full-time jobs. That feeds into our product portfolio, but what we’re trying to make clear at this conference is that for this to work really as vendors, we really need to show interoperability faster.

TNS: Okay. You talked about the community as a whole…

AD: Yeah, not just the community but also vendors, right? You walk in here, you see 9000 people, this is not an IBM conference but every single person here has IBM skills, because our cloud is with an OpenStack.

I can hire them; our clients can hire them. Think about it. If you’re a single vendor, and you’ve got your own proprietary software, you got to like build your own skill base and your own little fandom. We do that too very well.You can do that, but it’s a lot of work. Building a cloud is not maintaining a cloud. It’s not easy. There’s a lot of change in code.

But at the end of the day, all of our clients aren’t software providers. So what they end up partnering with the vendor like us and we deliver the technology, but they get the interoperability. They’re able to get the skill they need from a larger base, not just a single source.

TNS: So regarding what you’re offering for enterprises, Bluemix provides these cloud services behind the customer’s firewall.

Diaz: Bluemix is a Cloud Foundry container, right? We can provide that public or you can get Bluemix on premise too, on open source. We actually have the ability to stand up to the cloud and remotely manage it for our clients. We’ve got a thing called Relay. We will handle all the updates, all the patches, and everything. So you could focus on using your private cloud and that building a massive operations team, which is required.

Not all of the services come on premise. But a lot of the most common things are there. You could use them in combination. You could use stuff from outside, inside. You can use other providers, whatever. So our big push is interoperability.


TNS: We noticed a shift in the balance of responsibilities between developers and the system administration community with the new cloud environments. The developers in some of the larger shops are increasingly taking on operations duties…

Diaz: This is around the culture side and how you develop code. So the cloud has put a spotlight on developers. There’s no hiding. If something’s broken, you fix it. Fix. Fix. Fix. Iterate. Iterate. Iterate.

This is easy to do when you’re 28 people in San Francisco or San Jose. As soon as you got one person who lives in San Ramon, that developer is not in the day-to-day mix. So the question is, “How do you actually get that type of rigor and discipline about programming as it is internally? How do you do that at scale?”

Well, guess what, the open source community is doing it, right? So what we did is we started to transform how we build our code about three years ago internally. We have rejiggered and retrained about 4,000 or so cloud developers over the past four years, and we created methodology around that. We call it Bluemix Garage Method.

TNS: What is it?

Diaz: So there are about 25 or so base algebraic elements to this methodology. How do you do stand-ups? How do you meritocracy?  How do you do repair programming? How do you remote pairing? How do you do rotational samples? Basic things.

So we describe these and then we string them together through what we call paths. I don’t know if you remember Java Trails, but anyway, we call them paths. So if you are for example building a new application, well here’s a path that you go through. If you have an existing team trying to do this, here’s another path. If you’re building a mobile application, here’s another path for methodology. And it shows you to restructure teams and how to do development and how do you use an open source tool chain. And by the way, it’s abstract, but we also make it very concrete in Bluemix. So you can take that and make it a reality in Bluemix with our toolchain and everything else.

TNS: So there are a lot of companies where software development is now a core part of the business, and so Bluemix Garage could really help…

Diaz: Development is absolutely important part of the business. So the question of how you do it in this new world. And by the way, it’s not just for speed, right? You need speed; you need innovation and all that stuff, but [Bluemix Garage] is, frankly, a retention and skills-attracting mechanism.

Not every company has offices in California. Those companies can get whatever talent they need at whatever price they’re willing to pay, right? In New York, Seattle, Austin, there are some number of places where there’s a concentration of developers. But you’ve got companies everywhere that aren’t in these places. How do they retain and attract help? Culture.

If you want to get the new generation of developers to feel comfortable, you have to have this way of working. This is how they work today.

TNS: Because at the job interviews, when developers are asking, “Why should I work here?” then you can show them…

AD: You show them the way you work, that you’re not like stuck in some waterfall from hell where you spend 80 percent of your time building charts and not writing code. And this is also the way you organize your spaces, the way the labs are built.

So that whole culture aspect is very important. It has helped us probably obtain talent, retain talent, attract talent, and it also helps your clients.  I don’t know of any other vendor or big vendor, has this methodology.

IBM is a sponsor of The New Stack.

Group Created with Sketch.
TNS owner Insight Partners is an investor in: The New Stack.
THE NEW STACK UPDATE A newsletter digest of the week’s most important stories & analyses.