Highlights, mistakes and corrections, how did IBM, the former king of artificial intelligence, get back the “lost decade”?

IBM was once the king of artificial intelligence, and CEO Arvind Krishna is trying to reclaim that title.

On the day he announced his resignation, former IBM CEO Roryland praised Arvind Krishna (current IBM CEO) as “the CEO who is fit to lead the company into the next era.” Key technology areas such as artificial intelligence, cloud computing, quantum computing and blockchain have played an important role.”

In fact, Krishna has spent his entire career at IBM . As a Ph.D. in electrical engineering, he has been involved in the development of more than ten technical patents and has been with the company’s Watson Research labs since 1990.

After nearly two decades of training, he then became the head of IBM’s cloud and cognitive software division. Here, he helped orchestrate the company’s largest acquisition of 2018 and officially took over as CEO in April 2020.

Krishna can also be regarded as a “dangerous order”. After all, IBM has experienced business stagnation for a period of time and has lagged behind other technology giants to some extent.

As an early leader in artificial intelligence, IBM has invested a lot of money and resources in Watson’s division, and it also successfully let Watson beat human players on the 2011 TV quiz show Jeopardy. IBM has since explored other development paths, hoping to shape Watson as a medical tool to help doctors plan diagnosis and treatment recommendations. But unfortunately, Watson’s actual performance has not been able to meet expectations.

Earlier this year, IBM officially sold the assets of its Watson Health unit, but it has never stopped in AI exploration.

Today, IBM hopes to use AI technology to help corporate clients combat the climate crisis, and is once again on the TIME100 list.

Recently, foreign media TIME interviewed Arvind Krishna. In the interview, TIME and Krishna discussed the missteps in IBM’s early AI exploration and the company’s future development goals.

Interviews have been organized and revised for clarity.

TIME (Reporter): Arvind, my family lives in a town close to the IBM headquarters in Armonk, New York. My first after-school trip in kindergarten was to visit the IBM campus in the early seventies. Back then, IBM was synonymous with innovation, work style, and even image. How are these traditions reflected in today’s IBM?

Arvind Krishna: Good question, I’ve been thinking about this too. I think IBM does have a lot of advantages that will stand the test of time. You also mentioned, people, culture, temperament and innovation. I don’t think the innovation part has changed, nor has the excellent technical accumulation of IBM employees.

But in terms of image and temperament, IBM people are indeed different from the past. I joined IBM in 1990, when scientists still wore suits and ties, which are definitely not anymore. But this should be a trend culture at the national level, so it is not just IBM that has changed, but the entire United States. Now, we have a greater tolerance for how we dress, after all, this is the way to maximize people’s potential.

Of course, IBM has to keep pace with the times, and the biggest change is naturally the technology itself that we are concerned about. Thirty years ago, most of IBM’s business value came from physical computing equipment; but now, our main value comes from software development and consulting services, that is, technology deployment on behalf of customers. One constant of these changes, however, is IBM’s culture of helping clients grow their businesses, developing employee expertise, and emphasizing the benefits of speed and agility.

TIME: You’re also one of the tech executives to be CEOs during the COVID-19 pandemic. Before you took the helm, analysts say, IBM’s biggest problem was that leadership was in the hands of service and sales experts, not technologists. Do you think your promotion is also related to this problem?

Arvind Krishna: I feel like I can be CEO mainly because I have a clear vision for where the company is going – focusing on hybrid cloud and artificial intelligence . My desire is simple: to lead IBM back to business growth.

We’ve set a goal this year to achieve 5% business growth, not including the share from Kyndryl, the IT infrastructure services business. So I think what really matters is not just the specific subject background, but how to turn the wish into reality with a more responsible and pragmatic attitude.

I look at my role and think about what I should do every day. Of course I spend a lot of time on strategy, product, and related decisions, but I also have to devote energy to dealing with customers and making sure I get paid on time. I also spend some time working with partners.

In short, the CEO role is to take comprehensive positions across disciplines, and every CEO at IBM has done that. Of course, the tech industry is accelerating and moving faster. Under the rapid changes, as the highest decision-maker in the enterprise, the CEO’s own technical level and even technical influence are indeed facing more severe challenges.

TIME: When Watson won the Jeopardy quiz in 2011, the industry generally believed that IBM was the absolute leader in AI. But over the next decade, Watson seemed to fall short of meeting the expectations of corporate clients. In the end what happened? What does this period mean for IBM?

Arvind Krishna: Personally, I don’t see that period as a so-called “lost decade.” In our opinion, artificial intelligence is the only technological solution that can collect and use all the data that is continuously generated. We also know that there must be incredible value in this data. Without AI, humans can only interpret at most 1% of this data. Winning Jeopardy was certainly a great milestone, and it turned the world’s attention to AI technology.

More importantly, from this moment onwards, AI is no longer a toy in the lab, or the exclusive domain of top professors at MIT, Stanford, or Berkeley, it starts to play a role in the real world.

And what has contributed to all this is the tireless efforts of hundreds of scientists. But it must be admitted that when a technology moves from the laboratory to reality, it will always be accompanied by the shattering of many beautiful dreams. In 2012, 2013 and even 2014, the entire market was not ready to accept and trust artificial intelligence technology that was just “entering society”.

Of course, we also made mistakes, and we made more terrible things, and made more impact on the real economy.

Looking back on that time period, we can see that successful businesses used AI mostly in relatively simple and less critical areas — like recommending a book, a movie, or a website. Even if the recommendation is wrong, it is not life-threatening. So IBM should also start small, we are a little bit aggressive.

Maybe we can help fast food companies complete tagging automation first, we can help companies improve their business applications, in short, start with more inclusive areas. Once we get the initial results, we can build trust among our users and then scale up.

This is the way IBM is now, so it’s not that Watson is bad, but we were too eager to get to the sky at that time. Some things just have to be gradual.

TIME: After talking about AI for so long, can you make a simple definition of AI from your own perspective?

Arvind Krishna : Artificial intelligence is a technology that observes data and learns from it. In fact, the learning process of AI is very similar to human pattern matching learning. By the way, current AI is nowhere near as intelligent as a 9-month-old baby. **After all, the baby has seen a dog three times, and the fourth time he can recognize this animal. But the AI ​​had to look at millions of pictures of dogs to figure out what the hell it was.

In short, when the process of observing A, and the result is B, is repeated millions of times, such a pattern will be determined – this is the basic principle of artificial intelligence.

But today’s AI still doesn’t understand what gravity is. It also observes apples falling from trees, and can see which direction the apples are going after watching millions of similar videos. But with pears, it’s confused again. So I think it’s still going to take a long time to get this kind of general content together.

IBM is currently doing a lot of work on AI and climate change, can you be more specific?

Arvind Krishna: We believe that climate change is one of the core issues that our generation must address, and it is also a reflection of responsibility for the next generation. So we’ve made a commitment to achieve net zero emissions without relying on government subsidies, and not in 2050 or 2060, but in the upcoming 2030.

In order to solve this problem with AI technology, it is the actual quality of the data set that determines the result. Without access to the right data, we can greatly underestimate or overestimate our abilities. Can we find better carbon sequestration materials? Can we eliminate the current ubiquitous rate of energy waste of up to 30%? Can we optimize and break through inefficient energy usage habits? Can we harness and lead the open source spirit?

Can we build programs to motivate people to fight climate change with AI-created technologies? There are no definitive answers to these questions, but we’ve been rolling out tools and encouraging other startups to explore with them. I think these are the obligations we IBMers have to undertake as members of human society.

TIME: What is IBM currently doing to attract younger employees? The survey shows that most young people prefer to choose companies that have a more sense of mission and are more forward-looking. What is IBM’s experience in this regard?

Arvind Krishna: Let’s talk about statistics first. The last time I looked at the resume database, 3 million people had applied for IBM jobs. I think that’s a pretty good number, and we’ve been working hard to hire more people. That’s right, in order for employees to establish a real sense of belonging, we all have to break through the limitations of the thinking of “working workers”.

And IBM has succeeded because of our willingness to face really difficult challenges head-on, to take on tasks that take years to complete.

Take quantum computing, which has gone from science fiction to real-world applications in an experimental setting. We’ve been on this path for ten years, and it may be five to ten years in the future. So if you want to work on quantum computing, you can trust IBM, we are willing to work tirelessly for this difficult goal. After all, Watson, who won the Jeopardy Quiz Contest, took seven years from planning the project to attracting worldwide attention.

All in all, if you like technical work, you like working with smart people, and you like helping clients solve big, meaningful problems, then IBM is a good choice. On the contrary, if everyone just wants to develop mobile applications or krypton gold mobile games, then IBM is not suitable for you.

Original link:

https://time.com/6167753/arvind-krishna-ibm-ceo-interview/

The text and pictures in this article are from InfoQ