Elon Musk famously said any company relying on lidar is “doomed.” Tesla instead believes automated driving functions are built on visual recognition and is even working to remove the radar. China’s Xpeng begs to differ.
Founded in 2014, Xpeng is one of China’s most celebrated electric vehicle startups and went public when it was just six years old. Like Tesla, Xpeng sees automation as an integral part of its strategy; unlike the American giant, Xpeng uses a combination of radar, cameras, high-precision maps powered by Alibaba, localization systems developed in-house, and most recently, lidar to detect and predict road conditions.
“Lidar will provide the 3D drivable space and precise depth estimation to small moving obstacles even like kids and pets, and obviously, other pedestrians and the motorbikes which are a nightmare for anybody who’s working on driving,” Xinzhou Wu, who oversees Xpeng’s autonomous driving R&D center, said in an interview with TechCrunch.
“On top of that, we have the usual radar which gives you location and speed. Then you have the camera which has very rich, basic semantic information.”
Xpeng is adding lidar to its mass-produced EV model P5, which will begin delivering in the second half of this year. The car, a family sedan, will later be able to drive from point A to B based on a navigation route set by the driver on highways and certain urban roads in China that are covered by Alibaba’s maps. An older model without lidar already enables assisted driving on highways.
The system, called Navigation Guided Pilot, is benchmarked against Tesla’s Navigate On Autopilot, said Wu. It can, for example, automatically change lanes, enter or exit ramps, overtake other vehicles, and maneuver another car’s sudden cut-in, a common sight in China’s complex road conditions.
“The city is super hard compared to the highway but with lidar and precise perception capability, we will have essentially three layers of redundancy for sensing,” said Wu.
By definition, NGP is an advanced driver-assistance system (ADAS) as drivers still need to keep their hands on the wheel and take control at any time (Chinese laws don’t allow drivers to be hands-off on the road). The carmaker’s ambition is to remove the driver, that is, reach Level 4 autonomy two to four years from now, but real-life implementation will hinge on regulations, said Wu.
“But I’m not worried about that too much. I understand the Chinese government is actually the most flexible in terms of technology regulation.”
Musk’s disdain for lidar stems from the high costs of the remote sensing method that uses lasers. In the early days, a lidar unit spinning on top of a robotaxi could cost as much as $100,000, said Wu.
“Right now, [the cost] is at least two orders low,” said Wu. After 13 years with Qualcomm in the U.S., Wu joined Xpeng in late 2018 to work on automating the company’s electric cars. He currently leads a core autonomous driving R&D team of 500 staff and said the force will double in headcount by the end of this year.
“Our next vehicle is targeting the economy class. I would say it’s mid-range in terms of price,” he said, referring to the firm’s new lidar-powered sedan.
The lidar sensors powering Xpeng come from Livox, a firm touting more affordable lidar and an affiliate of DJI, the Shenzhen-based drone giant. Xpeng’s headquarters is in the adjacent city of Guangzhou about 1.5 hours’ drive away.
Xpeng isn’t the only one embracing lidar. Nio, a Chinese rival to Xpeng targeting a more premium market, unveiled a lidar-powered car in January but the model won’t start production until 2022. Arcfox, a new EV brand of Chinese state-owned carmaker BAIC, recently said it would be launching an electric car equipped with Huawei’s lidar.
Musk recently hinted that Tesla may remove radar from production outright as it inches closer to pure vision based on camera and machine learning. The billionaire founder isn’t particularly a fan of Xpeng, which he alleged owned a copy of Tesla’s old source code.
In 2019, Tesla filed a lawsuit against Cao Guangzhi alleging that the former Tesla engineer stole trade secrets and brought them to Xpeng. XPeng has repeatedly denied any wrongdoing. Cao no longer works at Xpeng.
While Livox claims to be an independent entity “incubated” by DJI, a source told TechCrunch previously that it is just a “team within DJI” positioned as a separate company. The intention to distance from DJI comes as no one’s surprise as the drone maker is on the U.S. government’s Entity List, which has cut key suppliers off from a multitude of Chinese tech firms including Huawei.
Other critical parts that Xpeng uses include NVIDIA’s Xavier system-on-the-chip computing platform and Bosch’s iBooster brake system. Globally, the ongoing semiconductor shortage is pushing auto executives to ponder over future scenarios where self-driving cars become even more dependent on chips.
Xpeng is well aware of supply chain risks. “Basically, safety is very important,” said Wu. “It’s more than the tension between countries around the world right now. Covid-19 is also creating a lot of issues for some of the suppliers, so having redundancy in the suppliers is some strategy we are looking very closely at.”
Xpeng could have easily tapped the flurry of autonomous driving solution providers in China, including Pony.ai and WeRide in its backyard Guangzhou. Instead, Xpeng becomes their competitor, working on automation in-house and pledges to outrival the artificial intelligence startups.
“The availability of massive computing for cars at affordable costs and the fast dropping price of lidar is making the two camps really the same,” Wu said of the dynamics between EV makers and robotaxi startups.
“[The robotaxi companies] have to work very hard to find a path to a mass-production vehicle. If they don’t do that, two years from now, they will find the technology is already available in mass production and their value become will become much less than today’s,” he added.
“We know how to mass-produce a technology up to the safety requirement and the quarantine required of the auto industry. This is a super high bar for anybody wanting to survive.”
Xpeng has no plans of going visual-only. Options of automotive technologies like lidar are becoming cheaper and more abundant, so “why do we have to bind our hands right now and say camera only?” Wu asked.
“We have a lot of respect for Elon and his company. We wish them all the best. But we will, as Xiaopeng [founder of Xpeng] said in one of his famous speeches, compete in China and hopefully in the rest of the world as well with different technologies.”
5G, coupled with cloud computing and cabin intelligence, will accelerate Xpeng’s path to achieve full automation, though Wu couldn’t share much detail on how 5G is used. When unmanned driving is viable, Xpeng will explore “a lot of exciting features” that go into a car when the driver’s hands are freed. Xpeng’s electric SUV is already available in Norway, and the company is looking to further expand globally.
IonQ, the trapped ion quantum computing company that recently went public via a SPAC, today announced that it is integrating its quantum computing platform with the open-source Qiskit software development kit. This means Qiskit users can now bring their programs to IonQ’s platform without any major modifications to their code.
At first glance, that seems relatively unremarkable, but it’s worth noting that Qiskit was founded by IBM Research and is IBM’s default tool for working with its quantum computers. There is a healthy bit of competition between IBM and IonQ (and, to be fair, many others in this space), in part because both are betting on very different technologies at the core of their platforms. While IonQ is betting on trapped ions, which allows its machines able to run at room temperature, IBM’s technique requires its machine to be supercooled.
“IonQ is excited to make our quantum computers and APIs easily accessible to the Qiskit community,” said IonQ CEO & President Peter Chapman. “Open source has already revolutionized traditional software development. With this integration, we’re bringing the world one step closer to the first generation of widely-applicable quantum applications.”
On the one hand, it’s hard not to look at this as IonQ needling IBM a bit, but it’s also an acknowledgment that Qiskit has become somewhat of a standard for developers who want to work with quantum computers. But putting these rivalries aside, we’re also in the early days of quantum computing and with no clear leader yet, anything that makes these various platforms more interoperable is a win for developers who want to dip their feet into writing for them.
Motional will integrate its driverless technology into Hyundai’s new all-electric SUV to create the company’s first robotaxi. At the start of 2023, customers in certain markets will be able to book the fully electric, fully autonomous taxi through the Lyft app.
The Hyundai IONIQ 5, which was revealed in February with a consumer release date expected later this year, will be fully integrated with Motional’s driverless system. The vehicles will be equipped with the hardware and software needed for Level 4 autonomous driving capabilities, including LiDAR, radar and cameras to provide the vehicle’s sensing system with 360 degrees of vision, and the ability to see up to 300 meters away. This level of driverless technology means a human will not be required to take over driving.
The interior living space will be similar to the consumer model, but additionally equipped with features needed for robotaxi operation, according to a Motional spokesperson. Motional did not reveal whether or not the vehicle would still have a steering wheel, and images of the robotaxi aren’t yet available.
Motional’s IONIQ 5 robotaxis have already begun testing on public roads and closed courses, and they’ll be put through more months of testing and real-world experience before being deployed on Lyft’s platform. The company says it’ll complete testing only once it’s confident that the taxis are safer than a human driver.
Motional, the Aptiv-Hyundai $4 billion joint venture aimed at commercializing driverless cars, announced its partnership with Lyft in December, signaling the ride-hailing company’s primary involvement in Motional’s plans. The company recently announced that it began testing its driverless tech on public roads in Las Vegas. Hyundai’s IONIQ 5 is Motional’s second platform to go driverless on public roads.
IBM has installed a couple of its own Quantum System One machines across the world in recent years, but today it announced its first private-sector U.S. deployment thanks to a new ten-year partnership with the Cleveland Clinic. This not only marks IBM’s first U.S. install of one of its quantum computers outside of its own facilities, but also the first time a healthcare institute purchases and houses a quantum computer. And thanks to this deal, Cleveland will also get access to IBM’s upcoming next-gen 1,000+ qubit quantum system.
We’re still in the very early days of commercializing quantum computing and for most current users, having access to a system over the cloud is sufficient for the experiments they are running. But increasingly, we are seeing research institutes and even some commercial users who are looking to install on-premises quantum computers to have full access to a dedicated machine.
This new deal is part of a larger partnership between IBM and the Cleveland Clinic, which also involves IBM’s hybrid cloud portfolio for high-performance computing and its AI tools. The partnership also forms the foundation of Cleveland Clinic’s new Center for Pathogen Research & Human Health, which is supported by $500 million in investments from the State of Ohio, Jobs Ohio and Cleveland Clinic.
“What we’re announcing here is the first — I’m going to call them private sector or nonprofit — but still, it’s the first sort of non-government organization that is going to have not only fully dedicated systems, but what is really, really remarkable is our commitment for the decades,” Dario Gil, IBM’s SVP and Director of IBM Research, told me. “In a way, they are partnering with us for the entire roadmap. So it’s not only taking receipt and getting access to a fleet of quantum computers and the next-generation quantum computer for next year. They’re also the first ones who are signing up and says, ‘I want the first 1,000+ qubit system.”
He noted that it takes a very forward-looking organization to invest heavily in quantum computing today. It’s one thing for a nation-state to start working with this nascent technology, given the potential it has in a wide variety of fields, but it’s another for a non-profit to make a similar bet. “The level of ambition is really, really high on their end because they’re thinking about the future,” Gil said of the Cleveland Clinic’s leadership.
Gil noted that as part of the overall deal, Cleveland Clinic’s researchers will also get access to IBM’s entire quantum portfolio in the cloud. IBM will maintain and support the on-premises quantum computer and they will remain IBM-owned machines, similar to its deals with government research labs in Japan and Germany, he explained.
“Maintaining it and supporting it is really critical,” Gil said about why that’s the case. “And they need us and our expertise to be able to do that. And also, you know, we do it because it’s like one of the most sensitive technologies that we have in IBM. So we are exquisitely focused on maintaining the security and safety for the machines.”
As part of the overall deal, IBM and Cleveland Clinic will also work on building skills among Cleveland Clinic’s researchers in quantum computing, but also AI and high-performance computing.
“Through this innovative collaboration, we have a unique opportunity to bring the future to life,” said Tom Mihaljevic, M.D., President and CEO of Cleveland Clinic. “These new computing technologies will revolutionize discovery in the life sciences and ultimately improve people’s lives. The Discovery Accelerator will enable our renowned teams to build a forward-looking digital infrastructure and transform medicine, while training the workforce of the future and growing our economy.”
While quantum computing may still be in its infancy, most pundits in the industry will tell you that now is the time to learn the basic concepts. And while there is little that’s immediately intuitive on the hardware side of quantum computing, the actual software tools that most players in the industry are developing today should feel somewhat familiar to virtually any developer.
Unsurprisingly, the “IBM Quantum Developer Certification,” as it’s officially called, focuses on IBM’s own software tools and especially Qiskit, its SDK for working with quantum computers. Qiskit has already proven quite popular, with more than 600,000 installs, and when IBM Quantum and the Qiskit team hosted a quantum summer school last year, almost 5,000 developers participated.
But on top of knowing their way around the basics of Qiskit (think defining and executing quantum circuits) developers also need to learn some of the basics of quantum computing itself. Once you know your way around Bloch spheres, Pauli matrices and Bell states, you’ll probably be in good shape for taking the certification exam, which will be administered on the Pearson VUE platform.
Abe Asfaw, the global lead for Quantum Education and Open Science at IBM, told me that this is just the first of a series of planned quantum certifications.
“What we’ve built is a multi-tiered developer certification,” he told me. “The first tier is what we’re releasing in this announcement and that tier gets developers introduced to how to work with quantum circuits. How do you use Qiskit […] to build out a quantum circuit and how do you run it on a quantum computer? And once you run it on a quantum computer, how do you look at the results and how do you interpret the results? This sets the stage for the next series of certifications that we’re developing, which are then going to be attached to use cases that are being explored in optimization, chemistry and finance. All of these can now be sort of integrated into the developer workflow once we have enabled someone to show that they can work with quantum circuits.”
Asfaw stressed that IBM has focused on educating developers about quantum computing for quite a while now, in part because it takes some time to develop the skills and intuition to build quantum circuits. He also noted that the open-source Qiskit SDK has integrated a lot of the tools that developers need to work at both the circuit level — which is a bit closer to writing in C or maybe even assembly in the classical computing world — and at the application level, where a lot of that is abstracted away.
“The idea is to make it easy for someone who is currently developing, whether it’s in the cloud, whether it’s using Python, to be able to run these tools and integrate quantum computing into their workflow,” Asfaw said. “I think the hardest part, to be very honest, is just giving someone the comfort to know that quantum computing is real today and that you can work with quantum computers. It’s as easy as opening up a Jupyter notebook and writing some code in Python.”
DeepSee.ai, a startup that helps enterprises use AI to automate line-of-business problems, today announced that it has raised a $22.6 million Series A funding round led by led by ForgePoint Capital. Previous investors AllegisCyber Capital and Signal Peak Ventures also participated in this round, which brings the Salt Lake City-based company’s total funding to date to $30.7 million.
The company argues that it offers enterprises a different take on process automation. The industry buzzword these days is ‘robotic process automation,’ but DeepSee.ai argues that what it does is different. I describe its system as ‘knowledge process automation’ (KPA). The company itself defines this as a system that “mines unstructured data, operationalizes AI-powered insights, and automates results into real-time action for the enterprise.” But the company also argues that today’s bots focus on basic task automation that doesn’t offer the kind of deeper insights that sophisticated machine learning models can bring to the table. The company also stresses that it doesn’t aim to replace knowledge workers but help them leverage AI to turn the plethora of data that businesses now collect into actionable insights.
“Executives are telling me they need business outcomes and not science projects,” writes DeepSee.ai CEO Steve Shillingford. “And today, the burgeoning frustration with most AI-centric deployments in large-scale enterprises is they look great in theory but largely fail in production. We think that’s because right now the current ‘AI approach’ lacks a holistic business context relevance. It’s unthinking, rigid, and without the contextual input of subject-matter experts on the ground. We founded DeepSee to bridge the gap between powerful technology and line-of-business, with adaptable solutions that empower our customers to operationalize AI-powered automation – delivering faster, better, and cheaper results for our users.”
To help businesses get started with the platform, DeepSee.ai offers three core tools. There’s DeepSee Assembler, which ingests unstructured data and gets it ready for labeling, model review and analysis. Then, DeepSee Atlas can use this data to train AI models that can understand a company’s business processes and help subject-matter experts define templates, rules and logic for automating a company’s internal processes. The third tool, DeepSee Advisor, meanwhile focuses on using text analysis to help companies better understand and evaluate their business processes.
Currently, the company’s focus is on providing these tools for insurance companies, the public sector and capital markets. In the insurance space, use cases include fraud detection, claims prediction and processing, and using large amounts of unstructured data to identify patterns in agent audits, for example.
That’s a relatively limited number of industries for a startup to operate in, but the company says it will use its new funding to accelerate product development and expand to new verticals.
“Using KPA, line-of-business executives can bridge data science and enterprise outcomes, operationalize AI/ML-powered automation at scale, and use predictive insights in real time to grow revenue, reduce cost, and mitigate risk,” said Sean Cunningham, Managing Director of ForgePoint Capital. “As a leading cybersecurity investor, ForgePoint sees the daily security challenges around insider threat, data visibility, and compliance. This investment in DeepSee accelerates the ability to reduce risk with business automation and delivers much-needed AI transparency required by customers for implementation.”
Bay Area-based construction startup TraceAir today announced a $3.5 million Series A. Led by London-based XTX Ventures, this round brings the company’s total funding up to $7 million. The raise includes existing investor Metropolis VC, along with new additions Liquid 2 Ventures, GEM Capital, GPS Ventures and Andrew Filev.
We first noted the company back in 2016, when it pitched a method for using drones to spot construction errors before they become too expense. It’s a pretty massive field that various technology companies are attempting to solve through a variety of different means, ranging from quadrupedal robots to site-scanning hard hats.
Last February, TraceAir announced a new drone management tool. “Haul Router provides the best mathematically objective hauls for each given drone scan,” the company noted at the time. “Any employee can use the tool to design a haul road and export the results to feed into grading equipment.”
The pandemic has thrown the construction industry for a loop (along with countless others). But unlike other sectors, demand still remains high in many places. TraceAir is hoping its solution will prove beneficial as many outfits seek a way to continue the process in spite of uncertainty.
“The Covid-19 pandemic created new challenges for the U.S. and worldwide construction industries, resulting in delayed projects and growing unemployment rates,” CEO Dmitry Korolev said in a release tied to the news. “Our platform allows industry leaders to manage projects more efficiently and collaborate with their teams remotely, minimizing the need for a physical presence on-site.”
TraceAir says the additional funding will go toward its sales and marketing, along with future product developments, including an unnamed product set for release this quarter.
LatticeFlow, an AI startup that was spun out of ETH Zurich in 2020, today announced that it has raised a $2.8 million seed funding round led by Swiss deep-tech fund btov and Global Founders Capital, which previously backed the likes of Revolut, Slack and Zalando.
The general idea behind LatticeFlow is to build tools that help AI teams build and deploy AI models that are safe, reliable and trustworthy. The problem today, the team argues, is that models get very good at finding the right statistical patterns to hit a given benchmark. That makes them inflexible, though, since these models were optimized for accuracy in a lab setting, not for robustness in the real world.
“One of the most commonly used paradigms for evaluating machine learning models is just aggregate metrics, like accuracy. And, of course, this is a super coarse representation of how good a model really is,” Pavol Bielik, the company’s CTO explained. “What we want to do is, we provide systematic ways of monitoring models, assessing their reliability across different relevant data slices and then also provide tools for improving these models.”
Building these kinds of models that are more flexible yet still provide robust results will take a new arsenal of tools, though, as well as the right team with deep expertise in these areas. Clearly, though, this is a founding team with the right background. In addition to CTO Bielik, the founding team includes Petar Tsankov, the company’s CEO and former senior researcher and lecturer at ETH Zurich, as well as ETH professors Martin Vechev, who leads the Secure, Reliable and Intelligence Systems lab at ETH, and Andreas Krause, who leads ETH’s Learning & Adaptive Systems lab. Tsankov’s last startup, DeepCode, was acquired by cybersecurity firm Snyk in 2020.
It’s also worth noting that Vechev, who previously co-founded ETH spin-off ChainSecurity, and his group at ETH previously developed ERAN, a verifier for large deep learning models with millions of parameters, that last year won the first competition for certifying deep neural networks. While the team was already looking at creating a company before winning this competition, Vechev noted that gave the team the confirmation that it was on the right path.
“We want to solve the main AI problem, which is making AI usable. This is the overarching goal,” Vechev told me. “[…] I don’t think you can actually found the company just purely based on the certification work. I think the kinds of skills that people have in the company, my group, Andreas [Krause]’s group, they all complement each other and cover a huge space, which I think is very, very unique. I don’t know of other companies who have covered this range of skills in these pressing points and have done groundbreaking work before.”
LatticeWorks already has a set of pilot customers who are trialing its tools. These include Swiss railways (SBB), which is using it to build a tool for automatic rail inspections, Germany’s Federal Cyber Security Bureau and the U.S. Army. The team is also working with other large enterprises that are using its tools to improve their computer vision models.
“Machine Learning (ML) is one of the core topics at SBB, as we see a huge potential in its application for an improved, intelligent and automated monitoring of our railway infrastructure,” said Dr. Ilir Fetai and Andre Roger, the leads of SBB’s AI team. “The project on robust and reliable AI with LatticeFlow, ETH, and Siemens has a crucial role in enabling us to fully exploit the advantages of using ML.”
For now, LatticeFlow remains in early access. The team plans to use the funding to accelerate its product development and bring on new customers. The team also plans to build out a presence in the U.S. in the near future.