Feb 10 2024
This Chinese city wants to be the Silicon Valley of chiplets

This story first appeared in China Report, MIT Technology Review’s newsletter about technology in China. Sign up to receive it in your inbox every Tuesday.

Last month, MIT Technology Review unveiled our pick for 10 Breakthrough Technologies of 2024. These are the technological advancements that we believe will change our lives today or sometime in the future. Among them, there is one that specifically matters to the Chinese tech sector: chiplets.

That’s what I wrote about in a new story today. Chiplets—the new chipmaking approach that breaks down chips into independent modules to reduce design costs and improve computing performance—can help China develop more powerful chips despite US government sanctions that prevent Chinese companies from importing certain key technologies.

Outside China, chiplets are one of the alternative routes that the semiconductor industry could take to improve chip performance cost-effectively. Instead of endlessly trying to cram more transistors into one chip, the chiplet approach proposes that the functions of a chip can be separated into several smaller devices, and each component could be easier to make than a powerful single-piece chip. Companies like Apple and Intel have already made commercial products this way. 

But within China, the technology takes on a different level of significance. US sanctions mean that Chinese companies can’t purchase the most advanced chips or the equipment to make them, so they have to figure out how to maximize the technologies they have. And chiplets come in handy here: if the companies can make each chiplet to the most advanced level they are capable of and assemble these chiplets into a system, it can act as a substitute for more powerful cutting-edge chips.

The technology needed to make chiplet is not that new. Huawei, the Chinese tech giant that has a chip-design subsidiary called HiSilicon, experimented with its first chiplet design product in 2014. But the technology became more important to the company after it was subject to strict sanctions from the US in 2019 and couldn’t work with foreign factories anymore. In 2022, Huawei’s then chairman, Guo Ping, said the company was hoping to connect and stack up less advanced chip modules to keep the products competitive in the market. 

Currently, there’s a lot of money going into the chiplet space. The Chinese government and investors have recognized the importance of chiplets, and they are pouring funding into academic projects and startups.

Particularly, there’s one Chinese city that has gone all-in on chiplets, and you very likely have never heard its name: Wuxi (pronounced woo-she). 

Halfway between Shanghai and Nanjing, Wuxi is a medium-sized city with a strong manufacturing industry. And it has a long history in the semiconductor sector: the Chinese government built a state-owned wafer factory there in the ’60s. And when the government decided to invest in the semiconductor industry by 1989, 75% of the state budget went into the factory in Wuxi.

By 2022, Wuxi had over 600 chip companies and was behind only Shanghai and Beijing in terms of semiconductor industry competitiveness. Particularly, Wuxi is the center of chip packaging—the final steps in the assembly process, like integrating the silicon part with its plastic case and testing the chip’s performance. JCET, the third-largest chip packaging company in the world and the largest of its kind in China, was founded in Wuxi more than five decades ago.

Their prominence in the packaging sector gives JCET and Wuxi an advantage in chiplets. Compared with traditional chips, chiplets are more accommodating of less-advanced manufacturing capabilities, but they require more sophisticated packaging techniques to ensure that different modules can work together seamlessly. So Wuxi’s established strength in packaging means it can be one step ahead of other cities in developing chiplets.

In 2023, Wuxi announced its plan to become the “Chiplet Valley.” The city has pledged to spend $14 million to subsidize companies that develop chiplets in the region, and it has formed the Wuxi Institute of Interconnect Technology to focus research efforts on chiplets. 

Wuxi is a great example of China’s hidden role in the global semiconductor industry: relative to sectors like chip design and manufacturing, packaging is labor intensive and not as desirable. That’s why there’s basically no packaging capability left in Western countries, and why places like Wuxi usually fly under everyone’s radar.

But with the opportunity presented by chiplets, as well as other advancements in packaging techniques, there’s a chance for chip packaging to enter center stage again. And China is betting on that possibility heavily right now to leverage one of its few domestic strengths to get ahead in the semiconductor industry.

Have you heard of Wuxi? Do you think it will play a more important role in the global semiconductor supply chain in the future? Let me know your thoughts at zeyi@technologyreview.com.

Catch up with China

1. TikTok’s CEO, Shou Zi Chew, testified in front of the US Senate on social media’s exploitation of children, along with the CEOs of Meta, Twitter, Snap, and Discord. (Associated Press)

2. Mayors from the US heartland are being invited to visit China as the country hopes to find local support outside Washington politics. (Washington Post $)

3. A new class action lawsuit is suing the genetic testing company 23andMe for a data breach that seems to have targeted people with Chinese and Ashkenazi Jewish heritage. (New York Times $)

4. Tesla is opening a new battery plant in Nevada, with manufacturing equipment bought from China’s battery giant CATL. (Bloomberg $)

5. A new Chinese documentary shows the everyday lives of ordinary blue-collar workers by stitching together 887 short videos shot by themselves on their mobile phones. (Sixth Tone)

6. Baidu’s venture capital arm is planning to sell its stakes in US startups, as the US-China investment environment has become much more politically sensitive. (The Information $)

7. Huawei and China’s biggest chipmaker, SMIC, could start making five-nanometer chips—still one generation behind the most advanced chips today—as early as this year. (Financial Times $) 

8. A pigeon was detained in India for eight months, suspected of carrying spy messages for China. It turns out it’s an open-water racing bird from Taiwan. (Associated Press)

Lost in translation

Shanghai’s attempt to ban ride-hailing services from picking up passengers near the Pudong Airport lasted exactly one week before it was called off. From January 29 on, Chinese ride-hailing apps like Didi all stopped servicing users in the Shanghai airport area at the request of the local transportation department, according to the Chinese newspaper Southern Metropolis Daily. While traditional taxis are still allowed at the airport, passengers reported longer wait times and frequent refusals of service by taxi drivers. The raid-hail ban, aimed at ensuring smooth traffic flow during the Spring Festival travel rush, soon faced criticism and legal scrutiny for its suddenness and potential violations of antitrust laws. The situation underscores the ongoing debate over the role of ride-hailing services during peak travel seasons, with some Chinese cities like Shanghai frowning upon them while others have embraced them. In the early hours of February 4, the Shanghai government decided to reverse the ban, and ride-hailing cars were back in the airport area. 

One last thing

Lingyan, a panda living in a zoo in Henan province, could have been the first panda to get suspended on China’s TikTok for … twerking. The zoo hosted a livestream session on January 31, but it was suddenly suspended by the algorithm when Lingyan climbed on top of a dome and started shaking his butt. I don’t know if this means the algorithm is too good at recognizing twerking or too bad at telling pandas from humans.

A panda standing on top of a play den and twerking.

LUANCHUAN ZHUHAI WILDLIFE PARK VIA DOUYIN
Ecommerce MGMT 0 Comments
Feb 7 2024
Why China is betting big on chiplets

For the past couple of years, US sanctions have had the Chinese semiconductor industry locked in a stranglehold. While Chinese companies can still manufacture chips for today’s uses, they are not allowed to import certain chipmaking technologies, making it almost impossible for them to produce more advanced products.

There is a workaround, however. A relatively new technology known as chiplets is now offering China a way to circumvent these export bans, build a degree of self-reliance, and keep pace with other countries, particularly the US. 

In the past year, both the Chinese government and venture capitalists have been focused on propping up the domestic chiplet industry. Academic researchers are being incentivized to solve the cutting-edge issues involved in chiplet manufacturing, while some chiplet startups, like Polar Bear Tech, have already produced their first products.

In contrast to traditional chips, which integrate all components on a single piece of silicon, chiplets take a modular approach. Each chiplet has a dedicated function, like data processing or storage; they are then connected to become one system. Since each chiplet is smaller and more specialized, it’s cheaper to manufacture and less likely to malfunction. At the same time, individual chiplets in a system can be swapped out for newer, better versions to improve performance, while other functional components stay the same.

Because of their potential to support continued growth in the post–Moore’s Law era, MIT Technology Review chose chiplets as one of the 10 Breakthrough Technologies of 2024. Powerful companies in the chip sector, like AMD, Intel, and Apple, have already used the technology in their products. 

For those companies, chiplets are one of several ways that the semiconductor industry could keep increasing the computing power of chips despite their physical limits. But for Chinese chip companies, they could reduce the time and costs needed to develop more powerful chips domestically and supply growing, vital technology sectors like AI. And to turn that potential into reality, these companies need to invest in the chip-packaging technologies that connect chiplets into one device.

“Developing the kinds of advanced packaging technologies required to leverage chiplet design is undoubtedly on China’s to-do list,” says Cameron McKnight-MacNeil, a process analyst at the semiconductor intelligence firm TechInsights. “China is known to have some of the fundamental underlying technologies for chiplet deployment.”

A shortcut to higher-performance chips

The US government has used export blacklists to restrict China’s semiconductor industry development for several years. One such sanction, imposed in October 2022, banned selling to China any technology that can be used to build 14-nanometer-generation chips (a relatively advanced but not cutting-edge class) as well as more advanced ones.

For years, the Chinese government has looked for ways to overcome the resulting bottleneck in chipmaking, but breakthroughs in areas like lithography—the process of using light to transfer a design pattern onto the silicon base material—could take decades to pull off. Today, China still lags in chip-manufacturing capability relative to companies in Taiwan, the Netherlands, and elsewhere. “Although we’ve now seen [China’s Semiconductor Manufacturing International Corporation] produce seven-nanometer chips, we suspect that production is expensive and low yield,” says McKnight-MacNeil.

Chiplet technology, however, promises a way to get around the restriction. By separating the functions of a chip into multiple chiplet modules, it reduces the difficulty of making each individual part. If China can’t buy or make a single piece of a powerful chip, it could connect some less-advanced chiplets that it does have the ability to make. Together, they could potentially achieve a similar level of computing power to the chips that the US is blocking China from accessing, if not more.

But this approach to chipmaking poses a bigger challenge for another sector of the semiconductor industry: packaging, which is the process that assembles multiple components of a chip and tests the finished device’s performance. Making sure multiple chiplets can work together requires more sophisticated packaging techniques than those involved in a traditional single-piece chip. The technology used in this process is called advanced packaging. 

This is an easier lift for China. Today, Chinese companies are already responsible for 38% of the chip packaging worldwide. Companies in Taiwan and Singapore still control the more advanced technologies, but it’s less difficult to catch up on this front.

“Packaging is less standardized, somewhat less automated. It relies a lot more on skilled technicians,” says Harish Krishnaswamy, a professor at Columbia University who studies telecommunications and chip design. And since labor cost is still significantly cheaper in China than in the West, “I don’t think it’ll take decades [for China to catch up],” he says. 

Money is flowing into the chiplet industry

Like anything else in the semiconductor industry, developing chiplets costs money. But pushed by a sense of urgency to develop the domestic chip industry rapidly, the Chinese government and other investors have already started investing in chiplet researchers and startups.

In July 2023, the National Nature Science Foundation of China, the top state fund for fundamental research, announced its plan to fund 17 to 30 chiplet research projects involving design, manufacturing, packaging, and more. It plans to give out $4 million to $6.5 million of research funding in the next four years, the organization says, and the goal is to increase chip performance by “one to two magnitudes.”

This fund is more focused on academic research, but some local governments are also ready to invest in industrial opportunities in chiplets. Wuxi, a medium-sized city in eastern China, is positioning itself to be the hub of chiplet production—a “Chiplet Valley.” Last year, Wuxi’s government officials proposed establishing a $14 million fund to bring chiplet companies to the city, and it has already attracted a handful of domestic companies.

At the same time, a slew of Chinese startups that positioned themselves to work in the chiplet field have received venture backing. 

Polar Bear Tech, a Chinese startup developing universal and specialized chiplets, received over $14 million in investment in 2023. It released its first chiplet-based AI chip, the “Qiming 930,” in February 2023. Several other startups, like Chiplego, Calculet, and Kiwimoore, have also received millions of dollars to make specialized chiplets for cars or multimodal artificial-intelligence models. 

Challenges remain

There are trade-offs in opting for a chiplet approach. While it often lowers costs and improves customizability, having multiple components in a chip means more connections are needed. If one of them goes wrong, the whole chip can fail, so a high level of compatibility between modules is crucial. Connecting or stacking several chiplets also means that the system consumes more power and may heat up faster. That could undermine performance or even damage the chip itself. 

To avoid those problems, different companies designing chiplets must adhere to the same protocols and technical standards. Globally, major companies came together in 2022 to propose Universal Chiplet Interconnect Express (UCIe), an open standard on how to connect chiplets. 

But all players want more influence for themselves, so some Chinese entities have come up with their own chiplet standards. In fact, different research alliances have proposed at least two Chinese chiplet standards as alternatives to UCIe in 2023, and a third standard that came out in January 2024 zoomed in on data transmission instead of physical connections.

Without a universal standard recognized by everyone in the industry, chiplets won’t be able to achieve the level of customizability that the technology promises. And their downsides could make companies around the world go back to traditional one-piece chips.

For China, embracing chiplet technology won’t be enough to solve other problems, like the difficulty of obtaining or making lithography machines.

Combining several less-advanced chips might give a performance boost to China’s chip technologies and stand in for the advanced ones that it can’t access, but it won’t be able to produce a chip that’s far ahead of existing top-line products. And as the US government constantly updates and expands its semiconductor sanctions, the chiplet technologies could become subject to restrictions too.

In October 2023, when the US Commerce Department amended its earlier sanction on the Chinese semiconductor industry, it included some new language and a few mentions of chiplets. The amendment added new parameters determining what technology is banned from being sold to China, and some of those additions seem tailored to measuring how advanced chiplets are. 

While chip factories around the world are not restricted from producing less-advanced chips for China, the Commerce Department’s document also asked them to assess whether these products could become a part of a more powerful integrated chip, putting more pressure on them to verify that their products don’t end up being a component in something that they would have been banned from exporting to China.

With all the practical and political obstacles, the development of chiplets will still take some time, no matter how much political will and investment is poured into the sector. 

It may be a shortcut for the Chinese semiconductor industry to make stronger chips, but it won’t be the magic solution to the US-China technology war.

Correction: The story has been updated to clarify the classification of 14-nanometer chips.

Ecommerce MGMT 0 Comments
Jan 18 2024
Start with data to build a better supply chain

In business, the acceleration of change means enterprises have to live in the future, not the present. Having the tools and technologies to enable forward-thinking and underpin digital transformation is key to survival. Supply chain procurement leaders are tasked with improving operational efficiencies and keeping an eye on the bottom line. For Raimundo Martinez, global digital solutions manager of procurement and supply chain at bp, the journey toward building a better supply chain starts with data.

“So, today, everybody talks about AI, ML, and all these tools,” says Martinez. “But to be honest with you, I think your journey really starts a little bit earlier. I think when we go out and think about this advanced technology, which obviously, have their place, I think in the beginning, what you really need to focus is in your foundational [layer], and that is your data.”

In that vein, all of bp’s data has been migrated to the cloud and its multiple procurement departments have been consolidated into a single global procurement organization. Having a centralized, single data source can reduce complexities and avoid data discrepancies. The biggest challenge to changes like data centralization and procurement reorganization is not technical, Martinez says, but human. Bringing another tool or new process into the fold can cause some to push back. Making sure that employees understand the value of these changes and the solutions they can offer is imperative for business leaders.

Honesty toward both employees and end users—where an enterprise keeps track of its logistics, inventory, and processes—can be a costly investment. For a digital transformation journey of bp’s scale, an investment in supply chain visibility is an investment in customer trust and business reputability.

“They feel part of it. They’re more willing to give you feedback. They’re also willing to give you a little bit more leeway. If you say that the tool is going to be, or some feature is going to be delayed a month, for example, but you don’t give the reasons and they don’t have that transparency and visibility into what is driving that delay, people just lose faith in your tool,” says Martinez.

Looking to the future, Martinez stresses the importance of a strong data foundation as a precursor to taking advantage of emerging technologies like AI and machine learning that can work to create a semi-autonomous supply chain.

“Moving a supply chain from a transactional item to a much more strategic item with the leverage of this technology, I think, that, to me, is the ultimate vision for the supply chain,” says Martinez.

This episode of Business Lab is produced in partnership with Infosys Cobalt.

Full Transcript

Laurel Ruma: From MIT Technology Review, I’m Laurel Ruma. And this is Business Lab, the show that helps business leaders make sense of new technologies coming out of the lab and into the marketplace.

Our topic is building a better supply chain. AI can bring efficiencies to many aspects of an enterprise, including supply chain. And where better to start than internal procurement processes. With better data, better decisions can be made quicker, both internally and by customers and partners. And that is better for everyone.

Two words for you: automating transformation.

My guest is Raimundo Martinez, who is the global digital solutions manager of procurement and supply chain at bp.

This episode of Business Lab is produced in partnership with Infosys Cobalt.
Welcome, Raimundo.

Raimundo Martinez: Hi, Laurel. Thanks for having me today.

Laurel: So, let’s start with providing some context to our conversation. bp has been on a digital transformation journey. What spurred it, and how is it going?

Raimundo: I think there’s many factors spurring digital transformation. But if I look at all of this, I think probably the key one is the rate of change in the world today and in the past. I think instead of slowing down, I think the rate of change is accelerating, and that makes business survivability the need to have quick access to the data to almost not live in today, but live in the future. And having tools and technologies that allow them to see what is coming up, what routes of action they can take, and then to enact those mitigation plans faster.

And I think that’s where the digital transformation is the key enabler of that. And I would say that’s on the business side. I think the other one is the people mindset change, and that ties into how things are going. I think things are going pretty good. Technology wise, I’ve seen a large number of tools and technologies adopted. But I think probably the most important thing is this mindset and the workforce and the adoption of agile. This rate of change that we just talked in the first part can only probably be achieved in tame when the whole workforce has this agile mindset to react to it.

Laurel: Well, supply chain procurement leaders are under pressure to improve operational efficiencies while keeping a careful eye on the bottom line. What is bp’s procurement control tower, and how has it helped with bp’s digital transformation?

Raimundo: Yeah, sure. In a nutshell, think about old as myriad of systems of record where you have your data and users having to go to all of those. So, our control tower, what it does, is consolidate all the data in a single platform. And what we have done is not just present the data, but truly configured the data in form of alerts. And the idea is to tell our user, “This is what’s important. This are the three things that you really need to take care now.” And not stopping there, but then saying, “Look, in order to take that action, we’re giving you a summary information so you don’t have to go to any other system to actually understand what is driving that alert.” But then on top of that, we’re integrating that platform with this system’s record so that request can complete it in seconds instead of in weeks.

So, that in a nutshell, it’s the control tower platform. And the way have helped… Again, we talk about tools and people. So, on the tool side, being able to demonstrate how this automation is done and the value of it and being able for other units to actually recycle the work that you have done, it accelerates and inspire other technical resources to take advantage of that. And then on the user side, one of the effects that have, again, this idea of the ability mindset, everything that we’ve done in the tool development is agile. So, bringing the users into that journey have actually helped us to also accelerate that aspect of our digital transformation.

Laurel: On that topic of workplace agility. In 2020, bp began a reorganization that consolidated its procurement departments into that single global procurement organization. What were the challenges that resulted from this reorganization?

Raimundo: Yeah. To give you a more context on that. So, if you think about bp being this really large global organizations divided in business units, before the organizations, every one of these business units have their own procurement departments, which handle literally billions of dollars that’s how big they were. And in that business, they have the ERP systems, your contract repository, your process and process deviation. But you only manage the portfolio to that. Once you integrate all of those organizations into a single one, now your responsibility become across some of those multiple business units, that has your data in all of these business systems.

So, if you want to create a report, then it’s really complicated because you have to not only go to these different systems, but the taxonomy of the data is different. So, an example, some business will call their territory, North America, the other one will call it east and west coast. So, if you want a report for a new business owner, it becomes really, really hard, and also the reports might not be as complete as they are. So, that really calls for some tools that we need to put in place to support that. And on top of that, the volume of requests now is so greater that just changing and adding steps to process aren’t going to be enough. You really need to look into automation to satisfy this higher demand.

Laurel: Well, speaking of automation, it can leverage existing technology and build efficiencies. So, what is the role of advanced technologies, like AI, machine learning and advanced analytics in the approach to your ongoing transformation?

Raimundo: So, today, everybody talks about AI, ML, and all these tools. But to be honest with you, I think your journey really starts a little bit earlier. I think when we go out and think about this advanced technology, which obviously, have their place, I think in the beginning, what you really need to focus is in your foundational, and that is your data. So, you ask about the role of the cloud. So, for bp, what we have done is all of the data used to reside in multiple different sites out there. So, what we have done is all the data now has been migrated to the cloud. And then what the cloud also allows is to do transformations in place that help us really homogenize, what I just described before, North America, South America, then you can create another column and say, okay, now call it, whatever, United States, or however you want to call it.

So, all of this data transformation happened in a single spot. And what that does is also allow our users that need this data to go to a single source of truth and not be pulling data from multiple systems. An example of the chaos that that creates is somebody will be pulling invoice and data from Spence, somebody will pull in PayData. So, then you already have data discrepancy on the reporting. And having a centralized tool where everybody goes for the data reduces so much complexity on the system.

Laurel: And speaking about that kind of complexity, it’s clear that multiple procurement systems made it difficult to maintain quality compliance as well, and as well as production tracking in bp supply chain. So, what are some of the most challenging aspects of realizing this new vision with a centralized one-stop platform?

Raimundo: Yeah, we have a good list in there. So, let me break it into maybe technical and people, because I think people is something that we should talk about it. So, technical. I think one of the biggest things in technical is working with your technical team to find the right architecture. This is how your vision fits into our architecture, which will create less, let’s say, confusion and complexity into your architecture. And the other side of the technical challenge is finding the right tools. I’ll give you an example for our project. Initially, I thought, okay, RPA [robotic process automation] will be the technology to do this. So, we run a pilot RPA. And obviously, RPA has incredible applications out there. But at this point, RPA really wasn’t the tool for us given the changes that could happen on the screens from the system that we’re using. So, then we decided instead of going to RPA, going to API.

So, that’s an example of a challenge of finding exactly the right tool that you have. But to be honest with you, I think the biggest challenge is not technical, but human. Like I mentioned before, people are immersed in the sea of change that is going on, and here you come with yet another tool. So, even the tool you’re giving them might be a lot more efficient, people still want to cling to what they know. So, if they say, “Look, if I have to spend another two hours extracting data, putting Excel, collating and running a report…” Some people may rather do that than go to a platform where all of that is done for them. So, I think change management is key in these transformations to make sure that they’re able to sell or make people understand what the value of the tool is, and overcome that challenge, which is human normal aversion to change. And especially when you’re immersed on this really, really sea of change that was already going as a result of the reorganization.

Laurel: Yeah. People are hard, and tech can be easy. So, just to clarify, RPA is the robotic process automation in this context, correct?

Raimundo: Yeah, absolutely. Yeah. Sorry about the pretty layered… Yeah.

Laurel: No, no. There’s lots of acronyms going around.

So, inversely, we’re just discussing the challenges, what are the positive outcomes from making this transformation? And could you give us an example or a use case of how that updated platform boosted efficiency across existing processes?

Raimundo: Absolutely. Just quick generic. So, generic things is you find yourself a lot in this cycle of that data. The users look at the applications that said that data’s not correct, and they lose the appetite for using that, but the problem is they own the data, but the process to change the data is so cumbersome that people don’t really want to take ownership of that because they said, “Look, I have 20 things to do. The least in my list is updating that data.”

So, we’re in this cycle of trying to put tools out for the user, the data is not correct, but we’re not the ones who own the data. So, the specific example of how we broke that cycle is using automation. So, to give you an example, before we create automation, if you needed to change any contract data, you have to find what the contract is, then you have to go to a tool like Salesforce and create a case. That case goes to our global business support team, and then they have to read the case, open the system of record, make the change. And that could take between days or weeks. Meantime, the user is like, “Well, I requested this thing, and it hasn’t even happened.”

So, what we did is leverage internal technology. We already had a large investment on Microsoft, as you can imagine. And we said, look, “From Power BI, you can look at your contract, you can click on the record you want to change. Power App comes up and tells you what do you want to do.” Say, I want to change the contract owner, for example. It opens a window, says, “Who’s the new person you want to put in?” And as soon as you submit it, literally, within less than a second, the API goes to the system of record, change the owner, creates an email that notifies everybody who is an stakeholder in that contract, which then increases visibility to changes across the organization.

And at the same time, it leaves you an audit trail. So, if somebody wants to challenge that, you know exactly what happened. So, that has been an incredible outcome of reducing cycle time from days and weeks to merely seconds, at the same time, increasing communication and visibility into the data. That has been proved one of the greatest achievements that we have.

Laurel: Well, I think you’ve really outlined this challenge. So, investing in supply chain visibility can be costly, but often bolsters trust and reputability among customers. What’s the role of transparency and visibility in a digital transformation journey of this size?

Raimundo: I keep talking about agile, and I think that’s one of the tenets. And what I will add to transparent visibility, I would add actually honesty. I think it’s very, very easy to learn from success. Everybody wants to tout the great things that they have done, but people may a little bit less inclined to speak out about their mistakes. I’ll just give you an example of our situation with RPA. We don’t feel bad about it. We feel that the more we share that knowledge with the technical teams, the much more value it has because then people will learn from that again and not commit the same mistake obviously.

But I think also what honesty do in this visibility is when you bring your users into the development team, you have that visibility. They feel part of it. They’re more willing to give you feedback. And also, they’re also willing to give you a little bit more leeway. If you say that the tool is going to be, or some feature is going to be delayed a month, for example, but you don’t give the reasons and they don’t have that transparency and visibility into what is driving that delay, people just lose faith in your tool.

Where I think the more open, the more visible you are, but also, again, with honesty, is you have a product that is so much more well received and that everybody feels part of the tool. It’s something that in every training, at the end of the training, I just say, “By the way, this is not my tool. This is your tool. And the more engaged you are with us, the much better outcome you’re going to have.” And that’s just achieved through transparency and visibility.

Laurel: So, for other large organizations looking to create a centralized platform to improve supply chain visibility, what are some of the key best practices that you’ve found that leadership can adopt to achieve the smoothest transition?

Raimundo: So, I probably think about three things. I think, one, the leadership needs to really, really do is understand the project. And when I say, understand the project, is really understanding the technical complexity, the human aspect of it, because I think that’s where your leadership has a lot of role to play. They’re able to influence their teams on this project that you’re trying to… And then they really need to understand also what are the risks associated with this project. And also that these could be a very lengthy journey. Hopefully, obviously, there’ll be results and milestones along, but they need to feel comfortable with also this agile mentality that we’re going to do features, fail, adapt, and they really need to be part of that journey.

The second biggest, I think, most important thing is having the right team. And in that, I think I’ve been super fortunate. We have a great partnership with Infosys. I’ve got one of the engineers named Sai. What the Infosys team and my technical team says is, “Look, do not shortchange yourself on the ideas that you bring from the business side.” A lot of times, we might think about something as impossible. They really encourage me to come up with almost crazy ideas. Just come with everything that you can think about. And they’re really, really incredible of delivering all the resources to bring in a solution to that. We almost end up using each other’s phrases. So, having a team that is really passionate about change, about being honest, about working together is the key to delivery. And finally, data foundation. I think that we get so stuck looking at the shiny tools out there that seem like science fiction and they’ll be great, and we forget that the outcome of those technologies are only as good as the data that we are supporting.

And data, a lot of times, it seem as like the, I don’t know, I don’t want to call it ugly sister, the ugly person in the room. But it’s really people… They’re like, “Oh, I don’t want to deal with that. I just want to do AI.” Well, your AI is not going to give you what you want if it doesn’t understand where you’re at. So, data foundation is key. Having the perfect team and technology partners and understanding the project length, the risk and being really engaged will be, for me, the key items there.

Laurel: That’s certainly helpful. So, looking ahead, what technologies or trends do you foresee will enable greater efficiencies across supply chain and operations?

Raimundo: It’s not like a broken record, bet. I really think that technologies that look at our data and help us clean the data, foresee what items we’re going to have with the data, how we can really have a data set that is really, really powerful, that is easy, and it has reflects exactly our situation, it’s the key for then the next step, which is all of these amazing technologies. If I think about our vision, for the platform is to create a semi-autonomous supply chain. And the vision is imagine having, again, first, the right data, and now what you have is AI/ML and all these models that look at that internal data, compare that with external factors.

And what it does is instead of presenting us alerts, we’ll go to the next level, and it, basically, presents scenarios. And say, “Look, based on the data that I see on the market, what you have had in your history, these are the three things that can happen, these are the plans that the tool recommends, and this is how you interact or affect that change.” So, moving a supply chain from a transactional item to a much more strategic item with the leverage of this technology, I think, that, to me, is the ultimate vision for supply chain.

Laurel: Well, Raimundo, thank you so much for joining us today on the Business Lab. This has been very enlightening.

Raimundo: Thank you. I’ve been a pleasure. And I wish everybody a great journey out there. It’s definitely a very exciting moment right now.

Laurel: Thank you.

That was Raimundo Martinez, who is a global digital solutions manager, procurement and supply chain at bp, who I spoke with from Cambridge, Massachusetts, the home of MIT and MIT Technology Review.

That’s it for this episode of Business Lab. I’m your host, Laurel Ruma. I’m the director of Insights, the custom publishing division of MIT Technology Review. We were founded in 1899 at the Massachusetts Institute of Technology, and you can find us in print, on the web and at events each year around the world. For more information about us and the show, please check out our website at technologyreview.com.

This show is available wherever you get your podcasts. If you enjoyed this episode, we hope you’ll take a moment to rate and review us. Business Lab is a production of MIT Technology Review. This episode was produced by Giro Studios. Thanks for listening.

This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.

Ecommerce MGMT 0 Comments
Dec 29 2023
We need a moonshot for computing

In its final weeks, the Obama administration released a report that rippled through the federal science and technology community. Titled Ensuring Long-Term US Leadership in Semiconductors, it warned that as conventional ways of building chips brushed up against the laws of physics, the United States was at risk of losing its edge in the chip industry. Five and a half years later, in 2022, Congress and the White House collaborated to address that possibility by passing the CHIPS and Science Act—a bold venture patterned after the Manhattan Project, the Apollo program, and the Human Genome Project. Over the course of three administrations, the US government has begun to organize itself for the next era of computing.

Secretary of Commerce Gina Raimondo has gone so far as to directly compare the passage of CHIPS to President John F. Kennedy’s 1961 call to land a man on the moon. In doing so, she was evoking a US tradition of organizing the national innovation ecosystem to meet an audacious technological objective—one that the private sector alone could not reach. Before JFK’s announcement, there were organizational challenges and disagreement over the best path forward to ensure national competitiveness in space. Such is the pattern of technological ambitions left to their own timelines. 

Setting national policy for technological development involves making trade-offs and grappling with unknown future issues. How does a government account for technological uncertainty? What will the nature of its interaction with the private sector be? And does it make more sense to focus on boosting competitiveness in the near term or to place big bets on potential breakthroughs? 

The CHIPS and Science Act designated $39 billion for bringing chip factories, or “fabs,” and their key suppliers back to the United States, with an additional $11 billion committed to microelectronics R&D. At the center of the R&D program would be the National Semiconductor Technology Center, or NSTC—envisioned as a national “center of excellence” that would bring the best of the innovation ecosystem together to invent the next generation of microelectronics.  

In the year and a half since, CHIPS programs and offices have been stood up, and chip fabrication facilities in Arizona, Texas, and Ohio have broken ground. But it is the CHIPS R&D program that has an opportunity to shape the future of the field. Ultimately, there is a choice to make in terms of national R&D goals: the US can adopt a conservative strategy that aims to preserve its lead for the next five years, or it can orient itself toward genuine computing moonshots. The way the NSTC is organized, and the technology programs it chooses to pursue, will determine whether the United States plays it safe or goes “all in.” 

Welcome to the day of reckoning

In 1965, the late Intel founder Gordon Moore famously predicted that the path forward for computing involved cramming more transistors, or tiny switches, onto flat silicon wafers. Extrapolating from the birth of the integrated circuit seven years earlier, Moore forecast that transistor count would double regularly while the cost per transistor fell. But Moore was not merely making a prediction. He was also prescribing a technological strategy (sometimes called “transistor scaling”): shrink transistors and pack them closer and closer together, and chips become faster and cheaper. This approach not only led to the rise of a $600 billion semiconductor industry but ushered the world into the digital age. 

Ever insightful, Moore did not expect that transistor scaling would last forever. He referred to the point when this miniaturization process would reach its physical limits as the “day of reckoning.” The chip industry is now very close to reaching that day, if it is not there already. Costs are skyrocketing and technical challenges are mounting. Industry road maps suggest that we may have only about 10 to 15 years before transistor scaling reaches its physical limits—and it may stop being profitable even before that. 

To keep chips advancing in the near term, the semiconductor industry has adopted a two-part strategy. On the one hand, it is building “accelerator” chips tailored for specific applications (such as AI inference and training) to speed computation. On the other, firms are building hardware from smaller functional components—called “chiplets”—to reduce costs and improve customizability. These chiplets can be arranged side by side or stacked on top of one another. The 3D approach could be an especially powerful means of improving speeds. 

This two-part strategy will help over the next 10 years or so, but it has long-term limits. For one thing, it continues to rely on the same transistor-building method that is currently reaching the end of the line. And even with 3D integration, we will continue to grapple with energy-hungry communication bottlenecks. It is unclear how long this approach will enable chipmakers to produce cheaper and more capable computers.  

Building an institutional home for moonshots

The clear alternative is to develop alternatives to conventional computing. There is no shortage of candidates, including quantum computing; neuromorphic computing, which mimics the operation of the brain in hardware; and reversible computing, which has the potential to push the energy efficiency of computing to its physical limits. And there are plenty of novel materials and devices that could be used to build future computers, such as silicon photonics, magnetic materials,and superconductor electronics. These possibilities could even be combined to form hybrid computing systems.

None of these potential technologies are new: researchers have been working on them for many years, and quantum computing is certainly making progress in the private sector. But only Washington brings the convening power and R&D dollars to help these novel systems achieve scale. Traditionally, breakthroughs in microelectronics have emerged piecemeal, but realizing new approaches to computation requires building an entirely new computing “stack”—from the hardware level up to the algorithms and software. This requires an approach that can rally the entire innovation ecosystem around clear objectives to tackle multiple technical problems in tandem and provide the kind of support needed to “de-risk” otherwise risky ventures.

Does it make more sense to focus on boosting competitiveness in the near term or to place big bets on potential breakthroughs?

The NSTC can drive these efforts. To be successful, it would do well to follow DARPA’s lead by focusing on moonshot programs. Its research program will need to be insulated from outside pressures. It also needs to foster visionaries, including program managers from industry and academia, and back them with a large in-house technical staff. 

The center’s investment fund also needs to be thoughtfully managed, drawing on best practices from existing blue-chip deep-tech investment funds, such as ensuring transparency through due-diligence practices and offering entrepreneurs access to tools, facilities, and training. 

It is still early days for the NSTC: the road to success may be long and winding. But this is a crucial moment for US leadership in computing and microelectronics. As we chart the path forward for the NSTC and other R&D priorities, we’ll need to think critically about what kinds of institutions we’ll need to get us there. We may not get another chance to get it right.

Brady Helwig is an associate director for economy and PJ Maykish is a senior advisor at the Special Competitive Studies Project, a private foundation focused on making recommendations to strengthen long-term US competitiveness.

Ecommerce MGMT 0 Comments
Dec 22 2023
Recapturing early-internet whimsy with HTML

Websites weren’t always slick digital experiences. 

There was a time when surfing the web involved opening tabs that played music against your will and sifting through walls of Times New Roman text on a colored background. In the 2000s, before Squarespace and social media, websites were manifestations of individuality—built entirely from scratch using HTML, by users who had some knowledge of code and a desire to be on the internet. 

Scattered across the web are communities of programmers working to revive this seemingly outdated approach. Anchored in the concept of “HTML Energy,” a term coined by artists Laurel Schwulst and Elliott Cost, the movement is anything but a superficial appeal to retro aesthetics. It focuses on the tactile process of coding in HTML, exploring how the language invites self-expression and empowers individuals to claim their share of the web. Taking shape in small Discord channels and digital magazines, among other spaces, the HTML Energy movement is about celebrating the human touch in digital experiences. 

Today, the majority of the internet is optimized for social engagement, e-commerce, and streaming. Most internet traffic is concentrated in a small number of sites, all of which are owned by the same handful of companies. From lengthy ads to aggressive cookie settings, minor obstacles and nuisances are baked in. Users are constantly reminded that their access to the internet is conditional on the monetary interests of a few. The situation with X (formerly known as Twitter) perfectly encapsulates this state of internet ownership: it only took one executive to spark a mass exodus from the platform and to fragment its long-lived communities.

However, despite the monopolistic landscape of Big Tech, one fundamental reality continues to justify the internet’s democratic reputation: anyone can publish a website for free with HTML. With an abundance of real estate, the web technically has space for everyone. It’s just a matter of traffic. 

When I spoke to different members of the HTML Energy community, all consistently returned to one basic message: Everything on the web boils down to HTML. HTML is the backbone of any website. It’s the only thing needed for a website to run. While popular web development languages today use abridged commands that hide technical complexity through what’s known as data abstraction, HTML is granular, and previous coding knowledge is not a prerequisite. 

As Cost explains, it is precisely how forgiving HTML is that gives eager individuals an opportunity to self-publish on the web. With HTML, a site will still load even if a line of code is missing. The HTML Energy movement embraces these possibilities: learning via trial and error is welcomed, and creative experimentation is encouraged. 

Since the rise of site-building tools like Wix, the intricate and sometimes clunky experience of hard-coding fonts or pixel spacing into a site has been replaced by templates and conventions of user experience design. As mainstream digital experiences trend toward a homogeneous visual language, the human touch gets lost in the many layers of abstraction. Site creators grow more distant from their sites, and the web becomes more transactional.

But the HTML Energy movement calls on people to reexamine our relationship with technology. Crafting a site using HTML allows programmers to explore what a website can be. Unlike their corporate counterparts, people creating sites on their own don’t answer to shareholders. They don’t have the pressure to create profitable experiences, so their creations can take an endless variety of forms. 

Common types of HTML Energy sites include digital gardens, where elements change with the seasons; interactive poetry generators, where inputs from the user give rise to new meaning; and personal sites that share intimate details about their creators’ lives. In an internet that is increasingly consumerist, HTML Energy sites offer a gentle reminder that websites can be meditative experiences. 

The HTML Energy community advocates understanding HTML for what it quite literally is: a language. And it celebrates the way the rudimentary character of that language demands intention from the user. As an amalgamation of minuscule and intricate creative decisions, a site constructed using only HTML is a form of self-­expression. Viewing a site’s source code is as important as navigating its interface. There are often Easter eggs hidden in that code, such as cheeky messages or citations taken from other HTML sites. In many ways, an HTML site captures something of the creator’s identity: what did that individual choose to build, and how?

This fascination with different applications of HTML is also seen in physical community gatherings sometimes called “freewrites,” where members of the movement get together to write code. Sunday Sites and Fruitful School are among the websites that organize these gatherings, often integrating educational elements into their sessions to empower more people to join the movement. Meanwhile, sites like HTML Review showcase some of its products in the format of a literary magazine. 

PROJECT 1

Terrarium of Many Sceneries

Ji Kim’s Terrarium of Many Sceneries collages snippets of footage from an old iPhone. As visitors scroll through the site, images overlap and embedded audio clips play. When users click any image, a small description of when and where it was taken appears, alongside more accompanying media. 
Kim’s site is designed to mimic the sporadic, layered nature of memory. It is a digital experience that is intentionally fragmented and overwhelming—like trying to remember a family trip taken years ago. 

PROJECT 2

A Room with a Window

Shelby Wilson’s A Room with a Window is a site that allows for only one interaction: opening and closing a set of window shades. The site intentionally conflates physical and digital spaces: Wilson plays with the idea of a browser as a portal to a place with physical boundaries and edges, but also maintains surrealist components (the room doesn’t get darker when the blinds close) and randomized elements (the color of the room changes on each visit) to highlight the digital form.

PROJECT 3

HTML Garden

Spencer Chang’s site imagines what a garden might look like on the internet. Several “plants” made of native HTML elements grow, and the passage of time is acknowledged and noticeable upon each visit—seasons change, plants sprout and bloom. There’s no explicit action called for—just observation.

PROJECT 4

Prose Play

Katherine Yang’s Prose Play is an interactive poem that encourages users to input different words into a pre-set sentence structure. Framing words as variables, the site explores the interactivity of the internet. It puts the literary theory of the “Death of the Author”—the idea that the meaning of a text is not determined by the author’s intention but by the reader’s interpretation—in the context of code. 

PROJECT 5

Erich Friedman

Erich Friedman’s site is a personal encyclopedia of his life, with archives of everything from movie ratings to reviews of mini-golf courses across central Florida. Organized into the categories of Math Stuff, Puzzle Stuff, Personal Stuff, and Professional Stuff, the site is simple in structure. It uses basic HTML to showcase Friedman’s eclectic interests over the past decade, including a list of fun facts for every number from 0 to 9,999 and collections of math and trivia problems. The site does not drive any specific action. It merely stands as an exhaustive, candid portrait of Erich Friedman, occupying a small piece of the web. 

PROJECT 6

Museum of Screens

Toulou TouMou’s Museum of Screens is a site that houses browser games created by game enthusiasts. In order to interact with the games on display, users have to navigate the digital space like a physical museum visualized in ASCII graphics. There are actual visiting hours, with a “rest day” chosen at random. 
Created to give due credit to amateur developers during the era of Flash games, TouMou’s museum aims to highlight the importance of acknowledging authorship and the rich history of independent games. 

There is no centralized source for HTML Energy sites: serendipity makes them finding them feel special, like happening upon a piece of street art behind a parking lot. They’re not designed for discovery, nor are they optimized for any particular action. They simply engage with a visitor on the visitor’s terms, offering a portrait of their creator. If sites like Google or Facebook are the supermarkets and shopping malls where you buy your necessities, HTML Energy sites are like the hidden gardens you happen upon, unmarked on any map. 

Tiffany Ng is a freelance writer exploring the relationship between art, tech, and culture.

Ecommerce MGMT 0 Comments
Dec 12 2023
Human brain cells hooked up to a chip can do speech recognition

Brain organoids, clumps of human brain cells grown in a dish, can be hooked up to an electronic chip and carry out simple computational tasks, a new study shows. 

Feng Guo and his team at Indiana University Bloomington generated a brain organoid from stem cells, attached it to a computer chip, and connected their setup, known as Brainoware, to an AI tool. They found that this hybrid system could process, learn, and remember information. It was even able to carry out some rudimentary speech recognition. The work, published today in Nature Electronics, could one day lead to new kinds of bio-computers that are more efficient than conventional computers.

Scientists have been trying to build computers based on advanced biological systems for decades. Guo says that such computers could overcome some challenges of silicon-based computers, such as bottlenecks in data processing. 

Conventional computers are much better than brains in dealing with numbers, but human brains are better at processing complex information while using relatively little energy. “This is a first demonstration of using brain organoids [for computing],” says Guo. “It’s exciting to see the possibilities of organoids for biocomputing in the future.”     

With Brainoware, Guo aimed to use actual brain cells to send and receive information. When the researchers applied electrical stimulation to the hybrid system they’d built, Brainoware responded to those signals, and changes occurred in its neural networks. According to the researchers, this result suggests that the hybrid system did process information, and could perhaps even perform computing tasks without supervision.

Guo and his colleagues then attempted to see if Brainoware could perform any useful tasks. In one test, they used Brainoware to try to solve mathematical equations. They also gave it a benchmark test for speech recognition, using 240 audio clips of eight people pronouncing Japanese vowels. The clips were converted into electrical signals and applied to the Brainoware system. This generated signals in the neural networks of the brain organoid, which were then fed into an AI tool for decoding.

The researchers found that the brain organoid–AI system could decode the signals from the audio recordings, which is a form of speech recognition, says Guo. “But the accuracy was low,” he says. Although the system improved with training, reaching an accuracy of about 78%, it was still less accurate than artificial neural networks, according to the study. 

Lena Smirnova, an assistant professor of public health at Johns Hopkins University,  points out that brain organoids do not have the ability to truly hear speech but simply exhibit “a reaction” to pulses of electrical stimulation from the audio clips. And the study did not demonstrate whether Brainoware can process and store information over the long term or learn multiple tasks. Generating brain cell cultures in a lab and maintaining them long enough to perform computations is also a huge undertaking.

Still, she adds, “it’s a really good demonstration that shows the capabilities of brain organoids.”

Ecommerce MGMT 0 Comments
Dec 1 2023
Sustainability starts with the data center

When asked why he targeted banks, notorious criminal Willie Sutton reportedly answered, “Because that’s where the money is.” Similarly, when thoughtful organizations target sustainability, they look to their data centers—because that’s where the carbon emissions are.

The International Energy Agency (IEA) attributes about 1.5% of total global electricity use to data centers and data transmission networks. This figure is much higher, however, in countries with booming data storage sectors: in Ireland, 18% of electricity consumption was attributable to data centers in 2022, and in Denmark, it is projected to reach 15% by 2030. And while there have been encouraging shifts toward green-energy sources and increased deployment of energy-efficient hardware and software, organizations need to accelerate their data center sustainability efforts to meet ambitious net-zero targets.

For data center operators, options for boosting sustainability include shifting energy sources, upgrading physical infrastructure and hardware, improving and automating workflows, and updating the software that manages data center storage. Hitachi Vantara estimates that emissions attributable to data storage infrastructure can be reduced as much as 96% by using a combination of these approaches.

Critics might counter that, though data center decarbonization is a worthy social goal, it also imposes expenses that a company focused on its bottom line can ill afford. This, however, is a shortsighted view.

Data center decarbonization initiatives can provide an impetus that enables organizations to modernize, optimize, and automate their data centers. This leads directly to improved performance of mission-critical applications, as well as a smaller, denser, more efficient data center footprint—which then creates savings via reduced energy costs. And modern data storage and management solutions, beyond supporting sustainability, also create a unified platform for innovation and new business models through advanced data analytics, machine learning, and AI.

Dave Pearson, research vice president at IDC, says, “Decarbonization and the more efficient energy utilization of the data center are supported by the same technologies that support data center modernization. Modernization has sustainability goals, but obviously it provides all kinds of business benefits, including enabling data analytics and better business processes.”

Download the full report.

This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.

Ecommerce MGMT 0 Comments
Nov 19 2023
2023 Global Cloud Ecosystem

The cloud, fundamentally a tool for cost and resource efficiency, has long enabled companies and countries to organize around digital-first principles. It is an established capability that improves the bottom line for enterprises. However, maturity lags, and global standards are sorely needed.

Cloud capabilities play a crucial role in accelerating the global economy’s next stage of digital transformation. Results from our 2023 Global Cloud Ecosystem survey of executives indicate there are two stages of cloud maturity globally: one where firms adopt cloud to achieve essential opex and capex cost reduction, and a second where firms link cloud investments to a positive business value. Respondents indicate the two are converging quickly.

DOWNLOAD THE REPORT

The key findings are as follows:

  • Cloud helps the top and bottom line globally. Cloud computing infrastructure investment will be more than 60% of all IT infrastructure spend worldwide in 2023, according to analyst firm IDC, as flexible cloud resources continue to define efficiency and productivity for technology decision-makers. More than eight out of 10 survey respondents report more cost efficiency due to cloud deployments. While establishing a link between cloud capabilities and top-line profitability is challenging, 82% say they are currently tracking cloud ROI, and 66% report positive ROI from cloud investments.
  • Cloud-centric organizations expect strong data governance (but don’t always get it). Strong data privacy protection and governance is essential to accelerate cloud adoption. Perceptions of national data sovereignty and privacy frameworks vary, underscoring the lack of global standards. Most respondents decline to say their countries are leaders in the space, but more than two-thirds say they keep pace.
  • All in for zero-trust. Public and hybrid cloud assets raise cybersecurity concerns. But cloud is required to grow AI and automation, which help secure digital assets with data cataloging, access, and visibility. Because of the risk associated with AI, the broad surface of the data it draws on, and the way AI generates change, the zero-trust user paradigm has gained wide acceptance across industries. Some 86%of the survey respondents use zero-trust architecture. However, one-third do not routinely identify and classify cloud assets.
  • Sustainability in the cloud. The cloud’s primary function—scaling up computing resources—is a key enabler that mitigates compliance issues such as security; privacy; and environment, social, and governance (ESG). More than half (54%) of respondents say they use cloud tools for ESG reporting and compliance, and a large number (51%) use cloud to enhance diversity, equity, and inclusion (DEI) compliance.

Download the full report.

This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.

Ecommerce MGMT 0 Comments
Oct 29 2023
AI-powered 6G networks will reshape digital interactions

Sixth-generation (6G) mobile networks, underpinned by artificial intelligence (AI), are poised to combine communication and computing in a hyperconnected world of digital and physical experiences that will transform daily lives, experts predict.

“In the past, we talked about internet of things, but with 6G, we talk about intelligent or smart internet of things,” says Qin Fei, president of communications research institute at Vivo, a Chinese mobile phone maker that has stepped up R&D efforts into 6G since 2020.

Communication and tech companies are already planning for 6G wireless networks, even though 5G has yet to be fully rolled out globally. With improved data latency, security, reliability, and the ability to process massive volumes of global data in real time, experts like Qin believe 6G is set to transform our leisure and work. Among the new use cases for 6G networks envisioned by Vivo are mixed reality, holographic and multi-sensory communication, interactive 3D virtual digital humans, collaborative robots, and automated driving.

AI boost for next-gen networks

There are expectations for 6G to be deployed by 2030. The UN’s telecoms agency, International Telecommunication Union (ITU), has stated it plans to finish the initial 6G standardization process no later than the year 2030.

Optimized by AI technologies, experts expect 6G to have a bigger impact than 5G for two reasons. One, because it will enable the convergence of computing and mobile communications. Two, because it will integrate digital and physical realms and introduce new sensory experiences for users.

Qin says that “6G will provide super communication and ubiquitous information, and converge computing services, thus being the base for an interconnected and converged physical and digital world.” Capgemini agrees—predicting that 6G networks will enable immersive, ubiquitous, and sensory digital experiences on a massive scale. This will make it possible for 6G applications to “sense” their surroundings, and thereby turn the network into “our sixth sense”, according to a report by the consultancy.

All about convergence: AI and communication

As each generation of wireless networks becomes increasingly complex, they rely on other technologies to harness their power and make them easier to run. 6G is expected to be one of the first AI-native networks, where AI is embedded in the networking equipment. This will enable the network to learn and manage itself, be more autonomous, and make it cheaper to run.

“When we are designing the 6G network, we’re going to use AI technology in designing the air interface and also in managing the 6G network,” says Qin. Machine learning and AI-based network automation will be crucial to simplify network management and optimization. “The 6G network with AI inside is like a very good student,” he adds. “The 6G network will self-train, self-learn, and it will actually grow as a student to become more and more powerful.”

The 6G disruption

Although 6G standards and specifications are still under development, experts agree that it will be a leapfrog technology, thanks to its higher speed (estimates vary, but 6G could be between 10 times, 50 times, to 100 times faster than 5G) and significantly reduced latency; improved connectivity, security, and reliability; and an ability to integrate digital and physical versions of the world.

“For 5G, it’s mainly a communication technology—that’s its core,” says Qin. “But for 6G, besides enhanced communications technology, it also includes computing, as well as other relevant services.” Another benefit is wider geographical coverage than 5G— 6G will cover the whole planet and connect all kinds of machines, he adds.

Qin says 6G networks will also popularize the use of digital twins—virtual replicas of products or processes used to predict how the physical entities will perform in the real world. This will be possible due to 6G networks’ enhanced connectivity, stronger sensing capability, and capacity to collect massive amounts of data.

According to Qin: “We could have more powerful connectivity and sensing capability, so we could install more sensors in the physical world and collect a massive amount of data about this world. With this data we could build models to rebuild the world in the digital arena.”

Vivo vision

Vivo believes 6G will support dozens or hundreds of new services in a wide range of industries. The company is developing prototype 6G mobile technologies based on three trends—communication plus sensing; communication plus computing; and communication plus AI.

For example, Vivo is developing a prototype that can collect users’ biometric data to monitor their health while they are asleep. According to this technological vision, a person’s bedside phone could become a medical monitoring device. “If there is any health issue or abnormal behavior happening with respiration, then [the phone] could send an alert to the hospital,” says Qin.

Vivo also sees virtual and mixed reality glasses as another potential application for 6G that could revolutionize video streaming by making it a more compelling and immersive experience. Current AR glasses have limited computing power, says Qin. “Therefore, it needs to connect as a kind of edge device to the cloud so it could provide better experiences for the users.”

6G will also support self-driving or autonomous cars. “I believe autonomous driving will be very popular after 2030 and be supported by 6G,” says Qin. “Driverless cars need to really gather all kinds of data, for example, about the ambient environment, about road conditions and even the adjacent cars in order to make informed decisions [such as] whether it should speed up or break. 6G can provide the computing power and network.”

Rollout challenges

Although 5G mobile networks have yet to live up to initial expectations, most experts agree that 6G has the potential to deliver major advances in connectivity and computing power. However, like any complex and powerful new technology, 6G also faces challenges, including network capacity and energy consumption.

Getting to the 6G era requires an increase in network capacity. Finding the right telecommunications spectrum to support its rollout is crucial. It has not been finalized but 6.4 to 15 gigahertz is under consideration for 6G. “We think that the spectrum for 6G should be on the lower spectrum, like 6.4 to 7.1 gigahertz, because the lower band electromagnetic wave physically has much better coverage and penetration characteristics,” says Qin.

Minimizing 6G’s energy consumption and carbon emissions is another major task. 6G networks will have vastly more computing demands than 5G. Suppliers and users will need to cooperate to minimize energy use. According to a report by GSMA, which represents global telecom operators, energy-saving techniques (such as AI-driven sleep states and lithium-ion batteries) may help to make 6G more energy efficient.

Ultimately, 6G will only succeed if it delivers great experiences and services for consumers and businesses, says Qin. “We should avoid overdesign of 6G network, and we should really collaborate with different verticals.” Problems faced by 5G networks—for example, bottlenecks in other technologies needed to support new terminals for 5G, such as material sciences for augmented reality equipment—should be lessons for 6G development, he adds. “We hope that we can avoid these problems. That means we need the whole ecosystem to collaborate, to jointly develop 6G infrastructure, mobile terminals, and applications.”

This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.

Ecommerce MGMT 0 Comments
Oct 24 2023
The power of green computing

When performing radiation therapy treatment, accuracy is key. Typically, the process of targeting cancer-affected areas for treatment is painstakingly done by hand. However, integrating a sustainably optimized AI tool into this process can improve accuracy in targeting cancerous regions, save health care workers time, and consume 20% less power to achieve these improved results. This is just one application of sustainable-forward computing that can offer immense improvements to operations across industries while also lowering carbon footprints.

Investments now in green computing can offer innovative outcomes for the future, says chief product sustainability officer and vice president and general manager for Future Platform Strategy and Sustainability at Intel, Jen Huffstetler. But transitioning to sustainable practices can be a formidable challenge for many enterprises. The key, Huffstetler says, is to start small and conduct an audit to understand your energy consumption and identify which areas require the greatest attention. Achieving sustainable computing requires company-wide focus from CIOs to product and manufacturing departments to IT teams.

“It really is going to take every single part of an enterprise to achieve sustainable computing for the future,” says Huffstetler.

Emerging AI tools are on the cutting edge of innovation but often require significant computing power and energy. “As AI technology matures, we’re seeing a clear view of some of its limitations,” says Huffstetler. “These gains have near limitless potential to solve large-scale problems, but they come at a very high price.”

Mitigating this energy consumption while still enabling the potential of AI means carefully optimizing the models, software, and hardware of these AI tools. This optimization comes down to focusing on data quality over quantity when training models, using evolved programming languages, and turning to carbon-aware software.

As AI applications arise in unpredictable real-world environments with energy, cost, and time constraints, new approaches to computing are necessary.

This episode of Business Lab is produced in partnership with Intel.

Full Transcript

Laurel Ruma: From MIT Technology Review, I’m Laurel Ruma and this is Business Lab, the show that helps business leaders make sense of new technologies coming out of the lab and into the marketplace.

Our topic today is building an AI strategy that’s sustainable, from supercomputers, to supply chain, to silicon chips. The choices made now for green computing and innovation to make a difference for today and the future.

Two words for you: sustainable computing.

My guest is Jen Huffstetler. Jen is the chief product sustainability officer and vice president and general manager for Future Platform Strategy and Sustainability at Intel.

This podcast is produced in partnership with Intel.

Welcome, Jen.

Jen Huffstetler: Thanks for having me, Laurel.

Laurel: Well, Jen, a little bit of a welcome back. You studied chemical engineering at MIT and continue to be involved in the community. So, as an engineer, what led you to Intel and how has that experience helped you see the world as it is now?

Jen: Well, as I was studying chemical engineering, we had lab class requirements, and it so happened that my third lab class was microelectronics processing. That really interested me, both the intricacy and the integration of engineering challenges in building computer chips. It led to an internship at Intel. And I’ve been here ever since.

And what I really love about it is we are always working on the future of compute. This has shaped how I see the world, because it really brings to life how engineers, the technology that is invented can help to advance society, bringing access to education globally, improving healthcare outcomes, as well as helping to shape work overall. As we were able to move to this pandemic world, that was all technology infrastructure that helped to enable the world to continue moving forward while we were facing this pandemic.

Laurel: That’s really great context, Jen. So, energy consumption from data infrastructure is outpacing the overall global energy demand. As a result, IT infrastructure needs to become more energy efficient. So, what are the major challenges that large-scale enterprises are facing when developing sustainability strategies?

Jen: Yeah, when we survey IT leaders[1] , we find that 76% believe that there is a challenge in meeting their energy efficiency goals while increasing performance to meet the needs of the business. In fact, 70% state that sustainability and compute performance are in direct conflict.

So, we don’t believe they have to be in conflict if you’re really truly utilizing the right software, the right hardware, and the right infrastructure design. Making operations more sustainable, it can seem daunting, but what we advise enterprises as they’re embarking on this journey is to really do an audit to survey where the biggest area of impact could be and start there. Not trying to solve everything at once, but really looking at the measurement of energy consumption, for an example in a data center today, and then identifying what’s contributing the most to that so that you can build projects and work to reduce in one area at a time.

And what we like to say is that sustainability, it’s not the domain of any one group at a company. It really is going to take every single part of an enterprise to achieve sustainable computing for the future. That includes of course, the CIOs with these projects to focus on reducing the footprint of their computing profile, but also in design for product and manufacturing companies, making sure that they’re designing and architecting for sustainability, and throughout the overall operations to ensure that everyone is reducing consumption of materials, whether it’s in the factory, the number of flights that a marketing or sales team is taking, and beyond.

Laurel: That’s definitely helpful context. So technologies like AI require significant computing power and energy. So, there’s a couple questions around that. What strategies can be deployed to mitigate AI’s energy consumption while also enabling its potential? And then how can smart investment in hardware help with this?

Jen: This is a great question. Technologies like you mentioned, like AI, they can consume so much energy. It’s estimated that the ChatGPT-3 model consumes 1.28 gigawatt hours of electricity, and that’s the same as the consumption for 120 US homes for a year. So, this is mind-boggling.

But one of the things that we think about for AI is there’s the training component and the inference component. You think about a self-driving car, and you train the model once and then it’s running on up to a hundred million cars, and that’s the inference. And so what we actually are seeing is that 70 to 80% of the energy consumption, or two to three x the amount of power is going to be used running inference as it can be to train the model. So, when we think about what strategies can be employed for reducing the energy consumption, we think about model optimization, software optimization, and hardware optimization, and you can even extend it to data center design.

They’re all important, but starting with model optimization, the first thing that we encourage folks to think about is the data quality versus data quantity. And using smaller data sets to train the model will use significantly less energy. In fact, some studies show that many parameters within a trained neural network can be pruned by as much as 99% to yield a much smaller, a sparser network, and that will lower your energy consumption.

Another thing to consider is tuning the models for a lower accuracy of intake. And an example of this is something we call quantization, and this is a technique to reduce your computational and your memory costs of running inference, and that’s by representing the weights and the activations with lower precision data types, like an 8-bit integer instead of a 32-bit floating point.

So, those are some of the ways that you can improve the model, but you can also improve them and lower their energy costs by looking at domain-specific models. Instead of reinventing the wheel and running these large language models again and again, if you, for example, have already trained a large model to understand language semantics, you can build a smaller one that taps into that larger model’s knowledge base and it will result in similar outputs with much greater energy efficiency. We think about this as orchestrating an ensemble of models. Those are just a couple of the examples. We can get more into the software and hardware optimization as well.

Laurel: Yeah, actually maybe we should stay on that a bit, especially considering how energy intensive AI is. Is there also a significant opportunity for digital optimization with software, as you mentioned? And then you work specifically with product sustainability, so then how can that AI be optimized across product lines for efficiency for software and hardware? Because you’re going to have to think about the entire ecosystem, correct?

Jen: Yeah, that’s right. This is really an area where I think in the beginning of computing technology, you think about the very limited resources that were available and how tightly integrated the coding had to be to the hardware. You think about the older programming languages, assembly languages, they really focused on using the limited resources available in both memory and compute.

Today we’ve evolved to these programming languages that are much more abstracted and less tightly coupled, and so what leaves is a lot of opportunity to improve the software optimization to get better use out of the hardware that you already have that you’re deploying today. This can provide tremendous energy savings, and sometimes it can be just through a single line of code. One example is Modin, an open source library which accelerates Pandas applications, which is a tool that data scientists and engineers utilize in their work. This can accelerate the application by up to 90x and has near infinite scaling from a PC to a cloud. And all of that is just through a single line of code change.

There’s many more optimizations within open source code for Python, Pandas, PyTorch, TensorFlow, and Scikit. This is really important that the data scientists and engineers are ensuring that they’re utilizing the most tightly coupled solution. Another example for machine learning on Scikit is through a patch, or through an Anaconda distribution, you can achieve up to an 8x acceleration in the compute time while consuming eight and a half times less energy and 7x less energy for the memory portions. So, all of this really works together in one system. Computing is a system of hardware and software.

There’s other use cases where when running inference on a CPU, there are accelerators inside that help to accelerate AI workloads directly. We estimate that 65% to 70% of inference is run today on CPUs, so it’s critical to make sure that they’re matching that hardware workload, or the hardware to the workload that you want to run, and make sure that you’re making the most energy-efficient choice in the processor.

The last area around software that we think about is carbon-aware computing or carbon-aware software, and this is a notion that you can run your workload where the grid is the least carbon-intensive. To help enable that, we’ve been partnering with the Green Software Foundation to build something called the Carbon Aware SDK, and this helps you to use the greenest energy solutions and run your workload at the greenest time, or in the greenest locations, or both. So, that’s for example, it’s choosing to run when the wind is blowing or when the sun is shining, and having tools so that you are providing the insights to these software innovators to make greener software decisions. All of these examples are ways to help reduce the carbon emissions of computing when running AI.

Laurel: That’s certainly helpful considering AI has emerged across industries and supply chains as this extremely powerful tool for large-scale business operations. So, you can see why you would need to consider all aspects of this. Could you explain though how AI is being used to improve those kind of business and manufacturing productivity investments for a large-scale enterprise like Intel?

Jen: Yeah. I think Intel is probably not alone in utilizing AI across the entirety of our enterprise. We’re almost two companies. We have a very large global manufacturing operations that is both for the Intel products, which is sort of that second business, but also a foundry for the world’s semiconductor designers to build on our solutions.

When we think of chip design, our teams use AI to do things like IP block placement. So, they are looking at grouping the logic, the different types of IP. And when you place those cells closer together, you’re not only lowering cost and the area of silicon manufacturing that lowers your embodied carbon for a chip, but it also enables a 50% to 30% decrease in the timing or the latency between the communication of those logic blocks, and that accelerates processing. That’ll lower your energy costs as well.

We also utilize AI in our chip testing. We’ve built AI models to help us to optimize what used to be thousands of tests and reducing them by up to 70%. It saves time, cost, and compute resources, which as we’ve talked about, that will also save energy.

In our manufacturing world we use AI and image processing to help us test a 100% of the wafer, detect up to 90% of the failures or more. And we’re doing this in a way that scales across our global network and it helps you to detect patterns that might become future issues. All of this work was previously done with manual methods and it was slow and less precise. So, we’re able to improve our factory output by employing AI and image processing techniques, decreasing defects, lowering the waste, and improving overall factory output.

We as well as many partners that we work with are also employing AI in sales techniques where you can train models to significantly scale your sales activity. We’re able to collect and interpret customer and ecosystem data and translate that into meaningful and actionable insights. One example is autonomous sales motions where we’re able to offer a customer or partner the access to information, and serving that up as they’re considering their next decisions through digital techniques, no human interventions needed. And this can have significant business savings and deliver business value to both Intel and our customers. So, we expect even more use at Intel, touching almost every aspect of our business through the deployment of AI technologies.

Laurel: As you mentioned, there’s lots of opportunities here for efficiencies. So, with AI and emerging technologies, we can see these efficiencies from large data centers to the edge, to where people are using this data for real-time decision making. So, how are you seeing these efficiencies actually in play?

Jen: Yeah, when I look at the many use cases from the edge, to an on-prem enterprise data center, as well as to the hyperscale cloud, you’re going to employ different techniques, right? You’ve got different constraints at the edge, both with latency, often power, and space constraints. Within an enterprise you might be limited by rack power. And the hyperscale, they’re managing a lot of workloads all at once.

So, starting first with the AI workload itself, we talked about some of those solutions to really make sure that you’re optimizing the model for the use case. There’s a lot of talk about these very large language models, over a hundred billion parameters. Every enterprise use case isn’t going to need models of that size. In fact, we expect a large number of enterprise models to be 7 billion parameters, but using those techniques we talked about to focus it on answering the questions that your enterprise needs. When you bring those domain specific models in play, they can run on even a single CPU versus this very large-scale dedicated accelerator clusters. So, that’s something to think about when you’re looking at, what’s the size of the problem I’m trying to solve, where do I need to train it, how do I need to run the inference, and what’s the exact use case? So, that’s the first thing I would take into account.

The second thing is, as energy becomes ever more a constraint across all of those domains, we are looking at new techniques and tools in order to get the most out of the energy that’s available to that data center, to that edge location. Something that we are seeing an increasing growth and expecting it to grow ever more over time is something called liquid cooling. Liquid cooling is useful at edge use cases because it is able to provide a contained solution, where sometimes you’ve got more dust, debris, particles, you think about telco or base stations that are out in very remote locations. So, how can you protect the compute and make it more efficient with the energy that’s available there?

We see the scaling both through enterprise data centers all the way up to large hyperscale deployments because you can reduce the energy consumption by up to 30%, and that’s important when today up to 40% of the energy in a data center is used to keep it cool. So, it’s kind of mind boggling the amount of energy or inefficiency that’s going into driving the compute. And what we’d love to see is a greater ratio of energy to compute, actually delivering compute output versus cooling it. And that’s where liquid cooling comes in.

There’s a couple of techniques there, and they have different applications, as I mentioned. Immersion’s actually one that would be really useful in those environments where it’s very dusty or there’s a lot of pollution at the edge where you’ve got a contained system. We’re also seeing cold plate or direct to chip. It’s already been in use for well over a decade in high performance computing applications, but we’re seeing that scale more significantly in these AI cluster buildouts because many data centers are running into a challenge with the amount of energy they’re able to get from their local utilities. So, to be able to utilize what they have and more efficiently, everyone is considering how am I going to deploy liquid cooling?

Laurel: That’s really interesting. It certainly shows the type of innovation that people are thinking about constantly. So, one of those other parts of innovation is how do you think about this from a leadership perspective? So, what are some of those best practices that can help an enterprise accelerate sustainability with AI?

Jen: Yeah, I think just to summarize what we’ve covered, it’s emphasizing that data quality over quantity, right? The smaller dataset will require less energy. Considering the level of accuracy that you really need for your use case. And again, where can you utilize that INT8 versus those compute intensive FP32 calculations. Leveraging domain-specific models so that you’re really right sizing the model for the task. Balancing your hardware and software from edge to cloud, and within a more heterogeneous AI infrastructure. Making sure that you’re using the computing chip set that’s necessary to meet your specific application needs. And utilizing hardware accelerators where you can to save energy both in the CPU as well. Utilizing open source solutions where there’s these libraries that we’ve talked about, and toolkits, and frameworks that have optimizations to ensure you’re getting the greatest performance from your hardware. And integrating those concepts of carbon-aware software.

Laurel: So, when we think about how to actually do this, Intel is actually a really great example, right? So, Intel’s committed to reaching net zero emissions in its global operations by 2040. And the company’s cumulative emissions over the last decade are nearly 75% lower than what they would’ve been without interim sustainability investments. So, then how can Intel’s tools and products help other enterprises then meet their own sustainability goals? I’m sure you have some use case examples.

Jen: Yeah, this is really the mission I’m on, is how can we help our customers lower their footprint? One of the first things I’ll just touch upon is, because you mentioned our 2040 goals, is that our data center processors are built with 93% renewable electricity. That immediately helps a customer lower their Scope 3 emissions. And that’s part of our journey to get to sustainable compute.

There’s also embedded accelerators within the Xeon processors that can deliver up to 14x better energy efficiency. That’s going to lower your energy consumption in data center no matter where you’ve deployed that compute. And of course, we have newer AI accelerators like Intel Gaudi, and they really are built to maximize the training and inference throughput and efficiency up to 2x over competing solutions. Our oneAPI software helps customers to take advantage of those built-in accelerators with solutions like an analytics toolkit and deep learning neural network software with optimized code.

We take all those assets, and just to give you a couple of customer examples, the first would be SK Telecom. This is the largest mobile operator in South Korea, 27 million subscribers. They were looking to analyze the massive amount of data that they have and really to optimize their end-to-end network AI pipeline. So, we partnered with them, utilizing the hardware and software solutions that we’ve talked about. And by utilizing these techniques, they were able to optimize their legacy GPU based implementation by up to four times, and six times for the deep learning training and inference. And they moved it to just a processor-based cluster. So, this really, it’s just an example where when you start to employ the hardware and the software techniques, and you utilize everything that’s inside the solution in the entire pipeline, how you can tightly couple the solution. And it doesn’t need to be this scaled out dedicated accelerator cluster. So, anyway, that’s one example. We have case studies.

Another one that I really love is with Siemens Healthineers. So, this is a healthcare use case. And you can envision for radiation therapy, you need to really be targeted where you’re going to put the radiation in the body, that it’s just hitting the organs that are being affected by the cancer. This contouring of the organs to target the solution was previously done by hand. And when you bring AI into the workflow, you’re not only saving healthcare workers’ time, of which we know that’s at a premium since there’s labor shortages throughout this industry, that they were able to improve the accuracy, improve the image generation 35 times faster, utilizing 20% less power, and enabling those healthcare workers to attend to the patients.

The last example is an intercom global telecommunication system provider with KDDI, which is Japan’s number one telecom provider. They did a proof of concept on their 5G network using AI to predict the network traffic. By looking at their solutions, they were able to scale back the frequency of the CPUs that were used and even idling them when not needed. And they were able to achieve significant power savings by employing those solutions. These are just ways where you can look at your own use cases, making sure that you’re meeting your customer SLAs or service level agreements, as is very critical in any mobile network, as all of us being consumers of that mobile network agree. We don’t like it when that network’s down. And these customers of ours were able to deploy AI, lower their energy consumption of their compute, while meeting their end use case needs.

Laurel: So Jen, this has been a great conversation, but looking forward, what are some product and technology innovations you’re excited to see emerge in the next three to five years?

Jen: Yeah, outside of the greater adoption of liquid cooling, which we think is foundational for the future of compute. In the field of AI, I’m thinking about new architectures that are being pioneered. There’s some at MIT, as I was talking to some of the professors there, but we also have some in our own labs and pathfinding organizations.

One example is around neuromorphic computing. As AI technology matures, we’re seeing a clear view of some of its limitations. These gains have near limitless potential to solve large-scale problems, but they come at a very high price, as we talked about with the computational power, the amount of data that gets pre-collected, pre-processed, et cetera.

So, some of these emerging AI applications arise in that unpredictable real world environment, and as you talked about some of those edge use cases. There could be power latency or data constraints, and that requires fundamentally new approaches. Neuromorphic computing is one of those, and it represents a fundamental rethinking of computer architecture down to the transistor level. And this is inspired by the form and the function of our human biological neural networks in our brains. It departs from those familiar algorithms and programming abstractions of conventional computing to unlock orders of magnitude gains in efficiency and performance. It can be up to 1,000x. I’ve even seen use cases of 2,500x energy efficiency over traditional compute architectures.

We have the Loihi research processor that incorporates these self-learning capabilities, novel neuro models, and asynchronous spike-based communication. And there is a software community that is working to evolve the use cases together on this processor. It consumes less than a watt of power for a variety of applications. So, it’s that type of innovation that really gets me excited for the future.

Laurel: That’s fantastic, Jen. Thank you so much for joining us on the Business Lab.

Jen: Thank you for having me. It was an honor to be here and share a little bit about what we’re seeing in the world of AI and sustainability.

Laurel: Thank you.

That was Jen Huffstetler, the chief product sustainability officer and vice president and general manager for Future Platform Strategy and Sustainability at Intel, whom I spoke with from Cambridge, Massachusetts, the home of MIT and MIT Technology Review.

That’s it for this episode of Business Lab. I’m your host, Laurel Ruma. I’m the Global Director of Insights, the custom publishing division of MIT Technology Review. We were founded in 1899 at the Massachusetts Institute of Technology. And you can find us in print on the web and at events each year around the world. For more information about us and the show, please check out our website at technologyreview.com.

This show is available wherever you get your podcasts. If you enjoyed this episode, we hope you’ll take a moment to rate and review us. Business Lab is a production of MIT Technology Review. This episode was produced by Giro Studios. Thanks for listening.

This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.

Ecommerce MGMT 0 Comments