High-quality data is the fuel that powers AI algorithms. Without a continual flow of labeled data, bottlenecks can occur and the algorithm will slowly get worse and add risk to the system.
It’s why labeled data is so critical for companies like Zoox, Cruise and Waymo, which use it to train machine learning models to develop and deploy autonomous vehicles. That need is what led to the creation of Scale AI, a startup that uses software and people to process and label image, lidar and map data for companies building machine learning algorithms. Companies working on autonomous vehicle technology make up a large swath of Scale’s customer base, although its platform is also used by Airbnb, Pinterest and OpenAI, among others.
The COVID-19 pandemic has slowed, or even halted, that flow of data as AV companies suspended testing on public roads — the means of collecting billions of images. Scale is hoping to turn the tap back on, and for free.
The company, in collaboration with lidar manufacturer Hesai, launched this week an open-source data set called PandaSet that can be used for training machine learning models for autonomous driving. The data set, which is free and licensed for academic and commercial use, includes data collected using Hesai’s forward-facing PandarGT lidar with image-like resolution, as well as its mechanical spinning lidar known as Pandar64. The data was collected while driving urban areas in San Francisco and Silicon Valley before officials issued stay-at-home orders in the area, according to the company.
“AI and machine learning are incredible technologies with an incredible potential for impact, but also a huge pain in the ass,” Scale CEO and co-founder Alexandr Wang told TechCrunch in a recent interview. “Machine learning is definitely a garbage in, garbage out kind of framework — you really need high-quality data to be able to power these algorithms. It’s why we built Scale and it’s also why we’re using this data set today to help drive forward the industry with an open-source perspective.”
The goal with this lidar data set was to give free access to a dense and content-rich data set, which Wang said was achieved by using two kinds of lidars in complex urban environments filled with cars, bikes, traffic lights and pedestrians.
“The Zoox and the Cruises of the world will often talk about how battle-tested their systems are in these dense urban environments,” Wang said. “We wanted to really expose that to the whole community.”
The data set includes more than 48,000 camera images and 16,000 lidar sweeps — more than 100 scenes of 8s each, according to the company. It also includes 28 annotation classes for each scene and 37 semantic segmentation labels for most scenes. Traditional cuboid labeling, those little boxes placed around a bike or car, for instance, can’t adequately identify all of the lidar data. So, Scale uses a point cloud segmentation tool to precisely annotate complex objects like rain.
Open sourcing AV data isn’t entirely new. Last year, Aptiv and Scale released nuScenes, a large-scale data set from an autonomous vehicle sensor suite. Argo AI, Cruise and Waymo were among a number of AV companies that have also released data to researchers. Argo AI released curated data along with high-definition maps, while Cruise shared a data visualization tool it created called Webviz that takes raw data collected from all the sensors on a robot and turns that binary code into visuals.
Scale’s efforts are a bit different; for instance, Wang said the license to use this data set doesn’t have any restrictions.
“There’s a big need right now and a continual need for high-quality labeled data,” Wang said. “That’s one of the biggest hurdles overcome when building self-driving systems. We want to democratize access to this data, especially at a time when a lot of the self-driving companies can’t collect it.”
That doesn’t mean Scale is going to suddenly give away all of its data. It is, after all a for-profit enterprise. But it’s already considering collecting and open sourcing fresher data later this year.
Palantir, the data analytics company co-founded by Peter Thiel, is already an active tech player in the scrum for federal contracts, but it’s playing a new and increasingly prominent role in providing the government with software tools to address the COVID-19 crisis.
This month, the Department of Veterans Affairs awarded a new $5 million contract to Palantir through veteran-owned software reseller i3 Federal LLC, which lists Palantir as a partner. The contract, set to run through November, will provide the VA with Palantir’s Gotham software to “track and analyze COVID-19 outbreak areas and make timely decisions with insight into supply chain capacity, hospital inventory, social service utilization and lab diagnostics.”
Palantir’s Gotham tool, best known for its use by law enforcement agencies, is one of the company’s main products and collects disparate data streams onto one platform. Twitter user Jack Poulson first spotted the contract, which was later reported by FedScoop.
A smaller contract also listed this month awards Palantir $2 million to provide the National Institutes of Health within the U.S. Department of Health and Human Services (HHS) a”COVID-19 dataset aggregation proof of concept.”
Palantir moved early to provide its services to governments grappling with the threat of the coronavirus and by mid-March was already working with the CDC to model infection patterns and anticipate hospital equipment needs.
In April, HHS awarded two contracts worth nearly $25 million to the company for a software platform called HHS Protect Now, intended to inform public health decisions made by the White House’s Coronavirus Task Force. HHS didn’t solicit competition for those contracts, noting that the COVID-19 crisis created a situation of “unusual and compelling urgency.”
As the pharmaceuticals industry turns its attention to precision medicine — the search for ever more tailored treatments for specific diseases using genetic engineering — Octant is using the same technologies to engage in drug discovery and diagnostics on a mass scale.
The company’s technology genetically engineers DNA to act as an identifier for the most common drug receptors inside the human genome. Basically, it’s creating QR codes that can flag and identify how different protein receptors in cells respond to chemicals. These are the biological sensors which help control everything from immune responses to the senses of sight and smell, the firing of neurons; even the release of hormones and communications between cells in the body are regulated.
“Our discovery platform was designed to map and measure the interconnected relationships between chemicals, multiple drug receptor pathways and diseases, enabling us to engineer multi-targeted drugs in a more rational way, across a wide spectrum of targets,” said Sri Kosuri, Octant’s co-founder and chief executive officer, in a statement.
Octant’s work is based on a technology first developed at the University of California Los Angeles by Kosuri and a team of researchers, which slashed the cost of making genetic sequences to $2 per gene from $50 to $100 per gene.
“Our method gives any lab that wants the power to build its own DNA sequences,” Kosuri said in a 2018 statement. “This is the first time that, without a million dollars, an average lab can make 10,000 genes from scratch.”
Joining Kosuri in launching Octant is Ramsey Homsany, a longtime friend of Kosuri’s, and a former executive at Google and Dropbox . Homsany happened to have a background in molecular biology from school, and when Kosuri would talk about the implications of the technology he developed, the two men knew they needed to for a company.
“We use these new tools to know which bar code is going with which construct or genetic variant or pathway that we’re working with [and] all of that fits into a single well,” said Kosuri. “What you can do on top of that is small molecule screening… we can do that with thousands of different wells at a time. So we can build these maps between chemicals and targets and pathways that are essential to drug development.”
Before coming to UCLA, Kosuri had a long history with companies developing products based on synthetic biology on both the coasts. Through some initial work that he’d done in the early days of the biofuel boom in 2007, Kosuri was connected with Flagship Ventures, and the imminent Harvard-based synthetic biologist George Church . He also served as a scientific advisor to Gen9, a company acquired by the multi-billion dollar synthetic biology powerhouse, Ginkgo Bioworks.
“Some of the most valuable drugs in history work on complex sets of drug targets, which is why Octant’s focus on polypharmacology is so compelling,” said Jason Kelly, the co-founder and CEO of Gingko Bioworks, and a member of the Octant board, in a statement. “Octant is engineering a lot of luck and cost out of the drug discovery equation with its novel platform and unique big data biology insights, which will drive the company’s internal development programs as well as potential partnerships.”
The new technology arrives at a unique moment in the industry where pharmaceutical companies are moving to target treatments for diseases that are tied to specific mutations, rather than look at treatments for more common disease problems, said Homsany.
“People are dropping common disease problems,” he said. “The biggest players are dropping these cases and it seems like that just didn’t make sense to us. So we thought about how would a company take these new technologies and apply them in a way that could solve some of this.”
One reason for the industry’s turn away from the big diseases that affect large swaths of the population is that new therapies are emerging to treat these conditions which don’t rely on drugs. While they wouldn’t get into specifics, Octant co-founders are pursuing treatments for what Kosuri said were conditions “in the metabolic space” and in the “neuropsychiatric space”.
Helping them pursue those targets, since Octant is very much a drug development company, is $20 million in financing from investors led by Andreessen Horowitz .
“Drug discovery remains a process of trial and error. Using its deep expertise in synthetic biology, the Octant team has engineered human cells that provide real-time, precise and complete readouts of the complex interactions and effects that drug molecules have within living cells,” said Jorge Conde, general partner at Andreessen Horowitz, and member of the Octant board of directors. “By querying biology at this unprecedented scale, Octant has the potential to systematically create exhaustive maps of drug targets and corresponding, novel treatments for our most intractable diseases.”
Microsoft today announced that Azure Quantum, its partner-centric quantum computing platform for developers who want to get started with quantum computing, is now in limited preview. First announced at Microsoft Ignite 2019, Azure Quantum brings together the hardware from IonQ, Honeywell, QCI and Microsoft, services from the likes of 1QBit, and the classical computing capabilities of the Azure cloud. With this move to being in limited preview, Microsoft is now opening the service up to a small number of select partners and customers.
At its current stage, quantum computing isn’t exactly a mission-critical capability for any business, but given how fast things are moving and how powerful the technology will be once it’s matured a bit over the next few years, many experts argue that now is the time to get started — especially because of how different quantum computing is from classical computing and how it will take developers a while to develop.
At Ignite, Microsoft also open-sourced its Quantum Development Kit, compilers and simulators.
With all of this, the company is taking a different approach from some of its competitors. In addition, Microsoft also currently has to partner with quantum hardware companies simply because its own quantum hardware efforts haven’t quite reached the point where they are viable. The company is taking a very different approach from the likes of IBM or Rigetti by betting on a different kind of qubit at the core of its machine. And while it has made some breakthroughs in recent months, it doesn’t yet have a working qubit – or if it does, it hasn’t publicly talked about it.
Taiwanese Semiconductor Manufacturing Co., the world’s largest contract semiconductor maker, has stopped taking new orders from Huawei Technologies, one of its largest customers, according to the Nikkei Asian Review. The report said the decision was made to comply with new United States export controls, announced last Friday, that are meant to make it more difficult for Huawei to obtain chips produced using U.S. technology, including manufacturing equipment.
Orders taken before the ban or already in production will not be affected, if they can ship before September 14. Huawei, the world’s largest telecom equipment maker, is TSMC’s second-biggest customer after Apple. TSMC makes many of the advanced chips used by Huawei, including in its smartphones.
The U.S. Commerce Department released its new orders on Friday, which specifically target Huawei by making it harder for the company to create chips using U.S. software and technology, even in foundries located abroad.
On the same day as the Commerce Department’s announcement, TSMC said that it is opening a new $12 billion advanced chip foundry in Arizona with support from the state and the U.S. federal government. Once opened, the plant will allow more of TSMC’s American clients to fabricate their chips domestically.
TSMC’s announcement came after the Wall Street Journal reported that White House officials were in discussions with TSMC and Intel to build foundries in the U.S. in order to reduce reliance on factories in Asia and the international supply chain.
In an email, a TSMC representative told TechCrunch that the company does not disclose customers’ order details. She added that TSMC complies with laws and applicable regulations, and is “following the U.S. export rule change closely” and “working closely with outside counsels to conduct legal analysis and ensure a comprehensive examination and interpretation of these rules.”
This is the latest restriction the U.S. government has leveled against Huawei citing national security concerns. Along with ZTE, Huawei was identified as a potential threat to security by the House Intelligence Committee in 2012.
The two companies have denied the charges, but under the Trump administration, the U.S government’s efforts to stop both from doing business with U.S. companies has intensified. According to the Nikkei Asian Review report, Huawei anticipated the Commerce Department’s new orders and has been building a year’s worth inventory of chips needed for its telecom equipment.
TechCrunch has contacted Huawei for comment.
Divergent, the Los Angeles-based startup aiming to revolutionize vehicle manufacturing, has cut about one-third of its staff amid the COVID-19 pandemic that has upended startups and major corporations alike.
The company, which employed about 160 people, laid off 57 workers, according to documents filed with the California Employment Development Department. Founder and CEO Kevin Czinger didn’t provide specific numbers. However, he did confirm to TechCrunch that he had to reduce staff due to the COVID-19 pandemic. A core team remains, he said.
“Whenever you’re doing something that’s affecting people’s jobs — and especially in a company where I basically recruited everyone and knew everyone by face and name — it’s obviously super painful to do that under any circumstance,” Czinger said in an interview this week.
The company’s No. 1 priority was to ensure long-term financial stability and secure the core team, technology development and customer programs no matter what the scenario, Czinger said, adding that there is still enormous uncertainty surrounding the real impact and duration of the COVID-19 pandemic.
“This was about making the company as totally weatherproof as possible,” Czinger said.
Divergent 3D is essentially a Tier 1 supplier for the automotive and aerospace industry. But it can hardly be considered a traditional supplier. After resigning as CEO of the now-defunct EV startup Coda Automotive in 2010, Czinger began to focus on how the vehicle manufacturing process could become more efficient and less wasteful.
Divergent 3D was born out of that initial exploration. The company developed an additive manufacturing platform designed to make it easier and faster to design and build new cars at a fraction of the cost — all while reducing the environmental impact that traditional factories have.
The platform is an end-to-end digital production system that uses high-speed 3D printers to make complex parts out of metal alloys. This system produces the structures of vehicles, such as the full frame, subframes and suspension structures that are part of the crash-performance structure of the vehicle.
In its early years as a company, Divergent 3D was perhaps best known for Blade, the first automobile to use 3D printing to form the body and chassis. Divergent 3D made Blade — which was on the auto show circuit in 2016 — to demonstrate the technology platform.
It was enough to get the attention of investors and at least two global OEMs as customers. Divergent can’t name the customers because of non-disclosure agreements.
The company has raised about $150 million from investors that include venture capital fund Horizons Ventures, automotive and aerospace engineering services company Altran Technologies and Chinese backers O Luxe Holdings, an investment conglomerate backed by the Hong Kong-based real estate investment magnate Li Ka-shing and Shanghai Alliance Investment Limited, an investment arm of the Shanghai Municipal Government.
The latest example of Divergent’s technology is the 21C, a hypercar unveiled in March that was built using the additive manufacturing platform. The high-performance 3D-printed vehicle was produced by Czinger Vehicles. Divergent 3D and Czinger Vehicles are wholly owned subsidiaries under Divergent Technologies.
Czinger said the company is poised to navigate the pandemic and ultimately survive. Divergent 3D has two global OEMs as customers. Structures such as chassis components and subframes, for which Divergent has supply contracts, are going through various testing and validation stages, depending on the program. Those programs, which are for serial production vehicles, are moving forward, Czinger said.
There will be delays as automakers have slowed or stopped operations. Czinger is hopeful that by 2021 the company will be able to announce that its 3D-printed structures will be production vehicles.
Voyage has cleared a regulatory hurdle that will allow the company to expand its self-driving service from the private roads of a retirement community in San Jose, Calif. to public roads throughout the rest of the state.
The California Public Utilities Commission issued a permit Monday that gives Voyage permission to transport passengers in its self-driving vehicles on the state’s public roads. The permit, which is part of the state’s Autonomous Vehicle Passenger Service pilot, puts Voyage in a new and growing group of companies seeking to expand beyond traditional AV testing. Aurora, AutoX, Cruise, Pony.ai, Zoox and Waymo have all received permits to participate in the CPUC’s Drivered Autonomous Vehicle Passenger Service Pilot program.
The permit also puts Voyage on a path toward broader commercialization.
The company was operating six autonomous vehicles — always with a human safety driver behind the wheel — in The Villages, a community of more than 4,000 residents in San Jose, Calif. (Those activities have been suspended temporarily under a statewide stay-at-home order prompted by the COVID-19 pandemic.) Voyage also operates in a 40-square-mile, 125,000-resident retirement city in central Florida.
Voyage didn’t need a CPUC permit because the community is made up of private roads, although CEO Oliver Cameron said the company wanted to adhere to state rules regardless of any technicalities. Voyage was also motivated by a grander ambition to transport residents of The Villages to destinations outside of the community.
“We want to bring people to all the things that live outside The Villages, facilities like hospitals and grocery stores,” Voyage CEO Oliver Cameron told TechCrunch in an interview Monday.
Voyage’s strategy was to start with retirement communities — places with specific customer demand and a simpler surrounding environment. The demographic that Voyage serves has an average age of 70. The aim isn’t to change its customer base. Instead, Cameron wants to expand the company’s current operational design domain to give Voyage a bigger reach.
The end goal is for Voyage’s core customers — people Cameron dubs power users — to be able to use the service for everything from heading to a neighbor’s house for dinner to shopping, doctor’s visits and even the airport.
Announcement time! We recently received a CPUC permit granting permission to move CA residents in driverless cars.
— Voyage (@voyage) April 20, 2020
The CPUC authorized in May 2018 two pilot programs for transporting passengers in autonomous vehicles. The first one, called the Drivered Autonomous Vehicle Passenger Service Pilot program, allows companies to operate a ride-hailing service using autonomous vehicles as long as they follow specific rules. Companies are not allowed to charge for rides, a human safety driver must be behind the wheel and certain data must be reported quarterly.
The second CPUC pilot would allow driverless passenger service — although no company has yet to obtain that permit.
Under the permit, Voyage can’t charge for rides. However, there might be some legal wiggle room. Voyage can technically charge for rides within The Villages; in fact, prior to the COVID-19 pandemic-related shutdown, the company had started charging for a ride-hailing service.
Rides outside of The Villages would have to be free, although it’s unclear if the company could charge for mileage or time until the vehicle left the community.
Voyage has aspirations to take this further. The company is also applying for a traditional Transportation Charter Permit, which is required for limousine, bus and other third-party charter services. Cameron said the company had to go through the stringent application process for the CPUC’s Drivered AV permit first.
The CPUC programs shouldn’t be confused with the California Department of Motor Vehicles, which regulates and issues permits for testing autonomous vehicles on public roads — always with a safety driver. There are 65 companies that hold autonomous vehicle testing permits issued by the DMV. Companies that want to participate in the CPUC program must have a testing permit with the DMV.
Hi and welcome back to The Station, a weekly newsletter dedicated to the future (and present) of transportation. I’m your host Kirsten Korosec, senior transportation reporter at TechCrunch.
What you’re reading here is an abbreviated version of The Station. To get the complete newsletter, which comes out every weekend, go here and click The Station.
There wasn’t a ton of news in micromobility this week, but I came across an interesting read over at City Lab about whether or not cities should financially support micromobility services. Shared bikes and scooters provide transportation options to city-dwellers during a time when some cities are deciding to scale back public transportation operations in order to keep its employees and residents safe.
In Portland, City Lab pointed to how the city agreed to temporarily waive e-scooter fees as long as Spin passed those savings onto riders. Now, Spin rides cost about 50% less in Portland.
But, as the authors write, “While we believe that waiving e-scooter fees and offering public funding may be necessary, we harbor no illusions that it would be easy to do so in the current fiscal environment.”
— Megan Rose Dickey
We hear things. But we’re not selfish. Let’s share.
Layoffs are nothing new in this COVID-19 world. More than 260 startups have laid off 25,010 workers, according Layoffs.fyi, a website that is attempting to track cuts in the startup ecosystem amid the COVID-19 pandemic.
Not all of these layoffs are directly related to the COVID-19 pandemic. In many cases, the pandemic has merely augmented pre-existing problems. One such example is Kodiak Robotics, an autonomous trucking startup, that laid off 20% of its staff on Wednesday (about 15 of its 85-person staff). The Information was the first to report the layoffs and TechCrunch has since confirmed those numbers. The official line is that Kodiak reduced its headcount due to the dramatic impact COVID-19 has had on the economy. The move was couched as the best way to position Kodiak for the future.
We’ve learned from several people that the company was already facing considerable headwinds on the fundraising front.
Kodiak Robotics came out of stealth in August 2018 with $40 million in a Series A funding round led by Battery Ventures. CRV, Lightspeed Venture Partners and Tusk Ventures also participated in the round. The company likely attracted interest and investment because of its founders. CEO Don Burnette was part of the Google self-driving project before leaving and co-founding Otto in early 2016, along with Anthony Levandowski, Lior Ron and Claire Delaunay. Uber then acquired Otto (and its co-founders). Burnette left Uber to launch Kodiak in April 2018 with Paz Eshel, a former venture capitalist and now the startup’s COO.
The pair scaled up quickly. The company, headquartered in Mountain View, Calif., went on a hiring spree in 2019 and opened a new facility in North Texas to support commercial deliveries using its fleet of eight trucks. Autonomous vehicle technology startups are already capitally intensive. But Kodiak was also trying to launch a carrier service — not just developing the self-driving truck stack.
Fundraising efforts started late last year and Kodiak was hoping to raise a $100 million round on a $300 million pre-money valuation, according to two sources. It was suggested that Kodiak already had a lead. However, the company has had trouble closing a Series B round with attractive terms, according to several sources who spoke to TechCrunch on condition of anonymity. When COVID-19 erupted it put more pressure on the startup.
Kodiak is hardly alone. Autonomous vehicle technology startups have had a more tepid reception from investors since spring 2019. It’s still possible to raise funds. But it’s harder now — particularly those seeking larger raises — and the terms are less desirable.
Pony .ai is the latest autonomous vehicle startup to turn its efforts to delivery — at least temporarily. The company announced this week it will partner with e-commerce platform Yamibuy to provide autonomous last-mile delivery service to customers in Irvine, Calif.
The new delivery service was launched to provide additional capacity to address the surge of online orders triggered by the COVID-19 pandemic, Pony.ai said.
Pony.ai, which recently raised $400 million from Toyota Motor Corporation, has focused on shuttling people, not packages. The company has launched ride-sharing and commuter pilots in Fremont and Irvine, California and Guangzhou, China.
Pony.ai now said it will use its Irvine robotaxi fleet of 10 electric Hyundai Kona vehicles for delivery through at least mid-summer. It’s not clear how, or if, Pony.ai can generate revenue with this new delivery service. The company is in talks with the California Department of Motor Vehicles, the agency that issues AV testing permits, about this issue. The DMV doesn’t allow AV testing fleets to charge money by delivering goods or rides. However, a deployment permit, which Pony.ai has for its Irvine service, does allow for commercial use, just not a delivery fee.
Pronto.ai, a startup co-founded by controversial star engineer Anthony Levandowski, is not pursuing Level 4 autonomous vehicle technology, Instead, the company is developing an advanced driver assistance system product for trucks called Copilot. Pronto AI was originally called Kache.ai, according to paperwork discovered at the time by TechCrunch, and was registered as a corporation with the California Secretary of State.
The startup has maintained a low profile since August 2019 when Levandowski was indicted by a federal grand jury on theft of trade secrets, forcing him to step down as CEO. Levandowski has since reached a plea deal. Now, it seems that the company is making some moves.
Pronto.ai recently applied for a five-year exemption from the federal government that would let drivers in trucks with Pronto’s CoPilot technology to stay on the road longer than current rules allow. The request to the Federal Motor Carrier Safety Administration, which was first reported by Freight Waves, would let drivers to drive up to 13 hours within a 15-consecutive hour driving window after coming on duty, following 10 consecutive hours off duty.
Drivers are typically allowed to drive up to 11 hours in a 14-hour window, after being off duty for 10 or more consecutive hours.
Boosted, startup behind the Boosted Boards and, more recently, the Boosted Rev electric scooter, would typically fall into micromobbin’. But it deserves it’s own segment this week.
Five weeks ago, Boosted laid off “a significant portion” of its team and began actively seeking a buyer. It seems that a sale never materialized and Lime swooped in and bought up Boosted’s core patents, according to a report from The Verge. Lime was apparently working on acquiring Boosted’s intellectual property since the end of 2019. The shared scooter company snapped up the IP after a proposed acquisition from Yamaha fell through for Boosted.
Boosted cofounder and former CEO Sanjay Dastoor, who left the board 18 months ago, posted a message to the Boosted subreddit shortly after The Verge story published that suggests Lime’s acquisition was broader than originally thought.
Dastoor wrote that the company is closed and will likely enter into some form of bankruptcy protection. He also wrote that Lime had purchased all the assets and IP of the company and appears to be in possession of everything at Boosted’s headquarters in Mountain View, including access to the building. Here’s one important nugget:
“As far as I can tell, this includes design files, software and code, diagnostics, parts, and test equipment I’m not sure if this includes the responsibility for warranty coverage for boards and scooters sold before. I do know that a handful of former engineers at Boosted, most senior is Michael Hillman who joined as VP Engineering last year, are now at Lime and may be able to help. Regardless of how this is structured, if we want our products to continue being supported, including parts for boards or any software diagnostic tests and debugging, their cooperation and help will be needed.”
He added that some Boosted employees have been trying unsuccessfully to service and send boards back to customers for weeks.
“I’m not a lawyer, but I suspect that those boards should rightfully get back to their owners and should be safe to ride, and I’m trying to find a way to help with this,” Dastoor wrote. “In the meantime, I’d recommend folks who are looking to get in touch more urgently should reach out to Lime directly.”
Starship Technologies has launched a robot food delivery service in Tempe, Ariz., as part of the autonomous delivery startup’s expansion plans following a $40 million funding round announced last August.
Starship Technologies, which was launched in 2014 by Skype co-founders Ahti Heinla and Janus Friis, has been ramping up commercial services in the past year, including a plan to expand to 100 universities by late summer 2021.
Now, with the COVID-19 pandemic forcing traditional restaurants to close and placing more pressure on gig economy workers, Starship Technologies has an opportunity to accelerate that growth.
Tempe isn’t the only new areas added amid the COVID-19 pandemic. Starship added a grocery delivery service in Washington, D.C in late March and expanded to Irvine, Calif. It also expanded its service area in Milton Keynes, U.K., where it has been operating since 2018. The company said it plans to add more cities in the coming weeks.
“The demand for contactless delivery has expanded exponentially in recent weeks,” Ryan Tuohy, who heads up business development at Starship Technologies, said in a statement. “We’re looking forward to serving the Tempe community as more people are looking for ways to support local businesses while spending more time at home. Our robots are doing autonomous deliveries in five countries and we’re grateful that our robots can make life a little bit easier for everyone.”
The autonomous robots, which can carry up to 20 pounds, could find a new customer base as people seek ways to get groceries and food without having to visit in person. Users place their order via the Starship Deliveries app and drop a pin where they want the delivery sent. The robot’s progress can be watched via an interactive map. Once the robot arrives, users receive an alert, and can then meet and unlock it through the app. The robots, which can cross streets, climb curbs, travel at night and operate in both rain and snow, are monitored remotely by Starship. Human operators can take control of the robots if needed.
In Tempe, the delivery service will initially employ more than 30 autonomous, on-demand robots between 10:30 a.m. and 8:30 p.m. daily in a geofenced area that includes several restaurants and a residential area. The service area is located about two miles from Arizona State University. Local residents are able to use the app to order from three restaurants, including Fate Brewing Company, Tempe City Tacos and Venezia’s Pizza of “Breaking Bad” fame.
Starship Technologies said it will expand the Tempe service area and add more restaurants and grocery stores soon.
And while COVID-19 has caused universities to close, Starship said it is continuing delivery services on multiple college campuses across the U.S. where international and grad students are residing.
Seeqc, a startup that is part of a relatively new class of quantum computing companies that is looking at how to best use classical computing to manage quantum processors, today announced that it has raised $5 million from M Ventures, the strategic corporate venture capital arm of Merck, the German pharmaceutical giant. Merck will be a strategic partner for Seeqc and will help it to develop its R&D efforts to develop useful application-specific quantum computers.
With this, New York State-based Seeqc has now raised a total of $11 million, including a recent $6.8 million seed round that included BlueYard Capital, Cambium, NewLab and the Partnership Fund for New York City.
Since developing new pharmaceuticals is an obvious use case for quantum computing, it makes sense that large pharmaceutical companies are trying to get ahead of their competitors by making strategic investments in companies like Seeqc.
The company is a spin-out of Hypres, a company that specializes in building superconductor integrated circuits. Hypres itself had raised about $100 million in total and notes that much of the work it did on building its solutions are now part of Seeqc.
As a company spokesperson told me, the idea behind Seeqc is to bring today’s room-sized quantum computers down to a more manageable scale. It’s doing so by combining its (and Hypres’) expertise in building superconductors with a hybrid approach to combines analog and digital. This includes digital qubit control and readout, together with the company’s own proprietary chip technology that integrates classical and quantum circuits into a hybrid system (and by default, quantum computers are hybrid systems that need a classical computer to control them).
The company argues that co-locating the classical compute with the quantum processor is critical to achieving the best performance. And since it owns and operates its own fab to build these chips, Seeqc also believes that it is one of the few companies that has the right infrastructure and expertise in place to design, test and build these superconductors.
“The ‘brute force’ or labware approach to quantum computing contemplates building machines with thousands or even millions of qubits requiring multiple analog cables and, in some cases, complex CMOS readout/control for each qubit, but that doesn’t scale effectively as the industry strives to deliver business-applicable solutions,” said John Levy, co-chief executive officer at Seeqc. “With Seeqc’s hybrid approach, we utilize the power of quantum computers in a digital system-on-a-chip environment, offering greater control, cost reduction and with a massive reduction in energy, introducing a more viable path to commercial scalability.”
The company believes that its approach can cut the cost of today’s large-scale quantum computers to 1/400th. All of this, of course, is still a while out and for now, the company will use the new funding to build a small-scale version of its system.
“We’re excited to be working with a world leading team and fab on one of the most pressing issues in modern quantum computing,” says Owen Lozman, Vice President at M Ventures. “We recognize that scaling the current generations of superconducting quantum computers beyond the noisy intermediate-scale quantum era will require fundamental changes in qubit control and wiring. Building on deep expertise in single flux quantum technologies, Seeqc has a clear, and importantly cost-efficient, pathway towards addressing existing challenges and disrupting analog, microwave-controlled architectures.”
Seeqc is, of course, not the only startup working on more efficient quantum control schemes. Quantum Machines, for example, also recently raised quite a bit of venture capital for its hardware/software quantum orchestration platform that also includes a custom processor, though that company’s overall approach is quite different from Seeqc’s.
ClimateView, a Swedish software development company working on monitoring and visualization tools for greenhouse gas emissions, said it has raised $2.5 million in its latest round of financing.
While the world is gripped by the material and economic toll of the COVID-19 epidemic, the problems society faces from longterm global climate change have not gone away.
It’s against this backdrop that investors including the Norrsken Foundation, an impact investment firm established by Klarna co-founder Niklas Adalberth; and Nordic Makers, an angel syndicate composed of founders from Zendesk, Sitecore, and Unity Technologies, decided to invest in ClimateView. Nordic Makers, Max Ventures, and GGV Capital also participated in the funding, the company said.
Using ClimateView’s software, cities around the world have a window into their climate data — including emissions and other sustainability and resilience information — so that they can plan accordingly for how best to proceed with decarbonization efforts and climate change mitigation plans.
So far, around 1,348 municipalities, townships, and villages in 26 countries have declared a climate emergency, but there’s no real effort to understand from a systems perspective what steps need to be taken to mitigate the worst impacts of the changing global climate, the company said.
“It’s an exciting time for ClimateView as we work to reinvent the way in which society works with the climate challenge,” said founder and chief executive Tomer Shalit, in a statement. “Our solution-focused approach to climate action is already gaining traction in a number of cities across the globe and we hope that, with this investment, we can continue to lay the groundwork for decision making so that, together, the world’s cities and nations can forge a common path towards global carbon neutrality.”
Historically, environmental policy and planning has been limited by a lengthy decision-making, planning-intensive process that hasn’t been able to access the latest data visualization tools and projections to make decisions based on current developments, the company said.
ClimateView’s software provides a central hub of all development, emissions, and projected urban planning data to accelerate the planning process.
The company’s premier project has been its work with the Swedish Climate Policy Council, which used the ClimateView software and suite of services to release a publicly available digital roadmap using the company’s Panorama software.
“Norrsken invests in startups that make the world better, so ClimateView is an ideal fit for us,”said Tove Larssen, a general partner with Norrsken. “We are really intrigued by their ambition to provide a global platform that makes it possible to fight climate change faster and more efficiently, and are delighted to be on board to help them achieve this goal.”
In the time of COVID-19, much of what transpires from the science world to the general public relates to the virus, and understandably so. But other domains, even within medical research, are still active — and as usual, there are tons of interesting (and heartening) stories out there that shouldn’t be lost in the furious activity of coronavirus coverage. This last week brought good news for several medical conditions as well as some innovations that could improve weather reporting and maybe save a few lives in Cambodia.
Arrhythmia is a relatively common condition in which the heart beats at an abnormal rate, causing a variety of effects, including, potentially, death. Detecting it is done using an electrocardiogram, and while the technique is sound and widely used, it has its limitations: first, it relies heavily on an expert interpreting the signal, and second, even an expert’s diagnosis doesn’t give a good idea of what the issue looks like in that particular heart. Knowing exactly where the flaw is makes treatment much easier.
Ultrasound is used for internal imaging in lots of ways, but two recent studies establish it as perhaps the next major step in arrhythmia treatment. Researchers at Columbia University used a form of ultrasound monitoring called Electromechanical Wave Imaging to create 3D animations of the patient’s heart as it beat, which helped specialists predict 96% of arrhythmia locations compared with 71% when using the ECG. The two could be used together to provide a more accurate picture of the heart’s condition before undergoing treatment.
Another approach from Stanford applies deep learning techniques to ultrasound imagery and shows that an AI agent can recognize the parts of the heart and record the efficiency with which it is moving blood with accuracy comparable to experts. As with other medical imagery AIs, this isn’t about replacing a doctor but augmenting them; an automated system can help triage and prioritize effectively, suggest things the doctor might have missed or provide an impartial concurrence with their opinion. The code and data set of EchoNet are available for download and inspection.
We have airplanes and drones in our airspace and satellites in space, but what about the space in between: the stratosphere?
There are platforms, such as blimps, balloons and high-altitude long endurance (HALE) fixed-wing platforms that can duplicate functions now performed by drones or satellites in a more technically and commercially viable manner.
Commercial drones operate in our airspace below 400 feet. Commercial aircraft fly between 9-12km (30,000-39,000 feet). Satellites operate in low Earth orbit (LEO, 500-1200km), mid Earth orbit (MEO, 2000-36,000km) and geostationary Earth orbit (GEO, 36,000km).
But what about the vast space in between our air space and LEO? The approximately 488km of space known as the stratosphere, is, at present, largely uninhabited and underutilized.
Imagine if a platform wants to loiter over a single point on the Earth for an extended period of time, either to maintain situational awareness and consistent surveillance over an area of interest or maintain communications. For example, after a natural disaster, it would be invaluable and life saving to have eyes, ears and a voice in the sky monitoring and helping the afflicted. Or what if the platform were able to monitor a natural disaster before it made landfall to collect better data on the storm’s size, location and path?
Other reasons why it might be advantageous to have persistent real-time video from the sky is surveillance of vast maritime regions and borders, identification of objects of interest and monitoring events, including storms, fires and environmental disasters, on behalf of first responders and enforcement agencies.
Another example could be global internet connectivity. If platforms mesh together and talk to one another, they could connect the world below in a much more effective and efficient manner than ground-based fiber optic cables. It could monitor our oceans or protect vulnerable people from exploitation. And the potential military, intelligence and governmental applications are obvious and substantial.
In short, the applications are abundant and the potential market for this type of platform massive.
Right now, the prevalent existing airborne platforms are drones and planes, and the prevalent existing space-based platforms are satellites. Each platform has various benefits, but none are optimized for many of the missions described above and, thus, do not necessarily accomplish those missions in the most efficient and effective manner.
Image Credits: Impossible Aerospace
Image Credits: alxpin (opens in a new window) /Getty Images
Constellation of LEO satellites
Image Credits: Spire Global
The solutions above are optimized for other types of critical missions. For example, drones are great for monitoring crops or inspecting infrastructure (as Drone Deploy software enables) or delivering emergency medical supplies (which Zipline and Google Wing are doing). Remotely-operated planes like General Atomics MQ-1 Predator have offensive military applications.
Constellations of LEO satellites in space, like Spire Global, can provide maritime, aviation and weather monitoring and prediction, or take photos of the world, as Planet Labs does. Lastly, GEO satellites can also be used for monitoring weather, communication and surveillance, but at a high level, not localized.
There are a handful of companies working on solutions specifically optimized for the mission of loitering over a single point. These solutions include balloons, blimps and HALE (high-altitude long endurance) platforms in the stratosphere.
Image Credits: WorldView
Companies like Loon, WorldView and WindBorne use air currents in the stratosphere to loiter over a single point. Their platforms have no propulsion on board and the structure consists of two balloons, a lift and a ballast balloon. The lift balloon contains either helium or hydrogen and is sealed with special UV-coated material. They use a compressor to add or remove air from the ballast balloon so that it becomes lighter or heavier to make the balloon go up or down depending on wind speed and direction and which air current they would like to ride.
Image Credits: MR1805 (opens in a new window) / Getty Images
Companies like Zenith and Skydweller are working on high-altitude long endurance (HALE) fixed-wing platforms. These high-aspect-ratio aircraft (which means long but slender wings) are powered by sunlight hitting the solar panels on the wings. The power that is generated can either power the plane and payload or be stored in the batteries. Therefore, if enough power is generated and stored during the day to last throughout the night, the plane can fly indefinitely.
*TRL: technology readiness level
For all of these platforms, there will be additional challenges in the areas of manufacturing and mission management. The platforms need to be manufactured and launched cheaply, quickly and reliably. This takes time and money. Additionally, there are issues relating to who will monitor the platforms once they are in the stratosphere — the company that built the platform or the customers whose payload the platform is holding?
Another issue that platforms that operate in the stratosphere will face relates to who regulates the stratosphere. Obviously, putting and operating platforms in the stratosphere raises a number of regulatory and legal questions that will have to be resolved.
I believe there is enough room in this market (and certainly in the stratosphere) for all of these platforms to be successful. They complement existing platforms such as drones and satellites and, for certain critical missions, can be more effective and efficient than their counterparts that operate in the airspace or in LEO/GEO.
Anthony Levandowski, the star self-driving car engineer who was at the center of a trade secrets lawsuit, has filed a motion to compel Uber into arbitration in the hopes that his former employee will have to shoulder the cost of at least $179 million judgment against him.
The motion to compel arbitration filed this week is part of Levandowski’s bankruptcy proceedings. It’s the latest chapter in a long and winding legal saga that has entangled Uber and Waymo, the former Google self-driving project that is now a business under Alphabet.
The motion represents the first legal step to force Uber to stand by an indemnity agreement with Levandowski. Uber signed an indemnity agreement in 2016 when it acquired Levandowski’s self-driving truck startup Otto . Under the agreement, Uber said it would indemnify — or compensate — Levandowski against claims brought by his former employer Google.
In Uber’s view the stakes are at least $64 million, according to the ride-hailing company’s annual report filed with the U.S. Securities and Exchange Commission . Although Levandowski, who was ordered in March 2020 to pay Google $179 million, is clearly shooting for more.
“For much of the past three years, Anthony ceded control of his personal defense to Uber because Uber insisted on controlling his defense as part of its duty to indemnify him. Then, when Uber didn’t like the outcome, it suddenly changed its mind and said it would not indemnify him. What Uber did is wrong, and Anthony has to protect his rights as a result,” Levandowksi’s lawyer Neel Chatterjee of Goodwin Procter said in an emailed statement to TechCrunch.
Levandowski was an engineer and one of the founding members in 2009 of the Google self-driving project, which was internally called Project Chauffeur. The Google self-driving project later spun out to become Waymo, a business under Alphabet. Levandowski was paid about $127 million by Google for his work on Project Chauffeur, according to the court document filed this week.
Levandowski left Google in January 2016 and started Otto, a self-driving trucking company, with three other Google veterans Lior Ron, Claire Delaunay and Don Burnette. Uber acquired Otto less than eight months later.
Before the acquisition closed, Uber conducted due diligence including hiring outside forensic investigation firm Stroz Friedberg to review the electronic devices of Levandowski and other Otto employees, according to the recent court filing. The investigation discovered that Levandowski had files belonging to Google on his devices, as well as indications that evidence may have been destroyed.
Uber agreed to a broad indemnification agreement in spite of the forensic evidence, which would protect Levandowski against claims brought by Google relating to his previous employment. Levandowski was worried that Google would attempt to get back any or all of the $127 million in compensation he had received.
That forecast didn’t take long to come true. Two months after the acquisition, Google made two arbitration demands against Levandowski and Ron. Uber wasn’t a party to either arbitration. However, it was on the hook under the indemnification agreement, to defend Levandowski.
Uber accepted those obligations and defended Levandowski. While the arbitrations played out, Waymo separately filed a lawsuit in February 2017 against Uber, for trade secret theft. Waymo alleged in the suit, which went to trial and ended in a settlement, that Levandowski stole trade secrets, which were then used by Uber. Under the settlement, Uber agreed to not incorporate Waymo’s confidential information into their hardware and software. Uber also agreed to pay a financial settlement that included 0.34% of Uber equity, per its Series G-1 round $72 billion valuation. That calculated at the time to about $244.8 million in Uber equity.
Meanwhile, the arbitration panel issued an interim award in March 2019 against each of Google’s former employees, including a $127 million judgment against Levandowski. The judgment also included another $1 million that Levandowski and Ron were jointly liable for. Google submitted a request for interest, attorney fees and other costs. A final award was issued in December.
Ron settled in February with Google for $9.7 million. However, Levandowski, disputed the ruling. The San Francisco County Superior Court denied his petition in March, granting Google’s petition to hold Levandowski to the arbitration agreement under which he was liable.
As the legal wrangling between Google and Levandowski and Uber played out, the engineer faced criminal charges. In August 2019, he was indicted by a federal grand jury with 33 counts of theft and attempted theft of trade secrets while working at Google. Last month, Levandowski reached a plea agreement with the U.S. District Attorney and pleaded guilty to one count of stealing trade secrets.
Levandowski’s lawyers argue that when the final judgment was entered against him, Uber reneged on its indemnification agreement. Levandowski said he was forced to file for Chapter 11 bankruptcy because Uber has refused to pay.
“While Uber and Levandowski are parties to an indemnification agreement, whether Uber is ultimately responsible for such indemnification is subject to a dispute between the Company and Levandowski,” Uber said, using similar language found in its annual report filed with the SEC.
Even if Levandowski’s legal team is able to convince a judge to compel Uber into arbitration, that doesn’t mean the outcome will be positive. Arbitration could take months to play out. In the end, Levandowski could still lose. But the filing allows Levandowski to speak out — albeit using legalese — and share details of his employment at Google and Uber. Among those are details about what Uber knew (and when) about Levandowski’s activities in recruiting Google employees as well as information he had downloaded onto his laptop, and discovered during the forensic investigation.
The first cracks between Uber and Levandowski appeared in April 2018, based on a timeline in the court document. It was then that Uber told Levandowski it intended to seek reimbursement for expenses used to defend him in the arbitration, according to claims laid out in the motion. Uber told Levandowski at the time, that one reason it was seeking reimbursement is because Levandowski “refused to testify at his deposition through an unjustifiably broad invocation of the Fifth Amendment.” Levandowski had used the Fifth Amendment in the deposition during the arbitration with Google.
Uber never requested Levandowski waive his Fifth Amendment rights and testify during the arbitration, according to the court document. Levandowski said that he immediately alerted Google and the arbitration panel that he was willing to testify and offered to make himself available for deposition before the arbitration hearing.
Development of artificial intelligence agents tends to frequently be measured by their performance in games, but there’s a good reason for that: Games tend to offer a wide proficiency curve, in terms of being relatively simple to grasp the basics, but difficult to master, and they almost always have a built-in scoring system to evaluate performance. DeepMind’s agents have tackled board game Go, as well as real-time strategy video game StarCraft – but the Alphabet company’s most recent feat is Agent57, a learning agent that can beat the average human on each of 57 Atari games with a wide range of difficulty, characteristics and gameplay styles.
Being better than humans at 57 Atari games may seem like an odd benchmark against which to measure the performance of a deep learning agent, but it’s actually a standard that goes all the way back to 2012, with a selection of Atari classics including Pitfall, Solaris, Montezuma’s Revenge and many others. Taken together, these games represent a broad range of difficulty levels, as well as requiring a range of different strategies in order to achieve success.
That’s a great type of challenge for creating a deep learning agent because the goal is not to build something that can determine one effective strategy that maximizes your chances of success every time you play a game – instead, the reason researchers build these agents and set them to these tasks at all is to develop something that can learn across multiple and shifting scenarios and conditions, with the long-term aim of building a learning agent that approaches general AI – or AI that is more human in terms of being able to apply its intelligence to any problem put before it, including challenges it’s never encountered before.
DeepMind’s Agent57 is remarkable because it performs better than human players on each of the 57 games in the Atari57 set – previous agents have been able to be better than human players on average – but that’s because they were extremely good at some of the simpler games that basically just worked via a simple action-reward loop, but terrible at games that required more advanced play, including long-term exploration and memory, like Montezuma’s Revenge.
The DeepMind team addressed this by building a distributed agent with different computers tackling different aspects of the problem, with some tuned to focus on novelty rewards (encountering things they haven’t encountered before), with both short- and long-term time horizons for when the novelty value resets. Others sought out more simple exploits, figuring out which repeated pattern provided the biggest reward, and then all the results are combined and managed by an agent equipped with a meta-controller that allows it to weight the costs and benefits of different approaches based on which game it encounters.
In the end, Agent57 is an accomplishment, but the team says it can stand to be improved in a few different ways. First, it’s incredibly computationally expensive to run, so they will seek to streamline that. Second, it’s actually not as good at some of the simpler games as some simpler agents – even though it excels at the the top 5 games in terms of challenge to previous intelligent agents. The team says it has ideas for how to make it even better at the simpler games that other, less sophisticated agents, are even better at.
New York University is among the many academic, private and public institutions doing what it can to address the need for personal protective equipment (PPE) among healthcare workers across the world. The school worked quickly to develop an open source face shield design, and is now offering that design freely to any and all in order to help scale manufacturing to meet needs.
Face shields are a key piece of equipment for frontline healthcare workers operating in close contact with COVID-19 patients. They’re essentially plastic, transparent masks that extend fully to cover a wearer’s face. These are to be used in tandem with N95 and surgical masks, and can protect a healthcare professional from exposure to droplets containing the virus expelled by patients when they cough or sneeze.
The NYU project is one of many attempts to scale production of face masks, but many others rely on 3D-printing. This has the advantage of allowing even very small commercial 3D print operations and individuals to contribute, but 3D printing takes a lot of time – roughly 30 minutes to an hour per print. NYU’s design requires only basic materials, including two pieces of clear, flexible plastic and an elastic band, and it can be manufactured in less than a minute by essentially any production facility that includes equipment for producing flat products (whole punches, laser cutters, etc.).
This was designed in collaboration with clinicians, and over 100 of them have already been distributed to emergency rooms. NYU’s team plans to ramp to scale production of up to 300,000 of these once they have materials in hand at the factories of production partners they’re working with, which include Daedalus Design and Production, PRG Scenic Technologies and Showman Fabricators.
Now, the team is putting the design out there for pubic use, including a downloadable tool kit so that other organizations can hopefully replicate what they’ve done and get more into circulation. They’re also welcoming inbound contact from manufacturers who can help scale additional production capacity.
Other initiatives are working on different aspects of the PPE shortage, including efforts to build ventilators and extend their use to as many patients as possible. It’s a great example of what’s possible when smart people and organizations collaborate and make their efforts available to the community, and there are bound to be plenty more examples like this as the COVID-19 crisis deepens.
Self-driving truck startup TuSimple is partnering with automotive supplier ZF to develop and produce autonomous vehicle technology, such as sensors, on a commercial scale.
The partnership, slated to begin in April, will cover China, Europe and North America. The two companies will co-develop sensors needed in autonomous vehicle technology such as cameras, lidar, radar and a central compute. As part of the partnership, ZF will contribute engineering support to validate and integrate TuSimple’s autonomous system into the vehicle.
TuSimple launched in 2015 and has operations in China, San Diego and Tucson, Ariz. The company has been working on a “full-stack solution,” an industry term that means developing and bringing together all of the technological pieces required for autonomous driving. TuSimple is developing a Level 4 system, a designation by the SAE that means the vehicle takes over all of the driving in certain conditions.
TuSimple has managed to scale up its operations and attract investors even as other companies in the nascent autonomous vehicle technology industry have faltered. The company has raised nearly $300 million to date from investors such as Sina, UPS and Tier 1 supplier Mando Corporation. It’s now making about 20 autonomous trips between Arizona and Texas each week with a fleet of more than 40 autonomous trucks. All of the trucks have a human safety operator behind the wheel.
The partnership is an important milestone for TuSimple as the startup prepares to bring autonomous-ready trucks to market, TuSimple chief product officer Chuck Price said in a statement. The plan is for TuSimple to combine its self-driving software with ZF’s ability to build automotive grade products.
The partnership doesn’t remove every barrier for TuSimple. Moving from development to deployment takes millions of dollars of investment. If a company can move from testing to commercial deployment, it must still navigate daily operations efficiently in the aim of becoming profitable.
The impending shortage of ventilators for U.S. hospitals is likely already a crisis, but will become even more dire as the number of patients with COVID-19 that are suffering from severe symptoms and require hospitalization grows. That’s why a simple piece of hardware newly approved by the FDA for emergency use – and available free via source code and 3D-printing for hospitals – might be a key ingredient in helping minimize the strain on frontline response efforts.
The Prisma Health VESper is a deceptively simple-looking three-way connector that expands use of one ventilator to treat up to four patients simultaneously. The device is made for use with ventilators that comply to existing ISO standard ventilator hardware and tubing, and allows use of filtering equipment to block any possible transmission of viruses and bacteria.
VESper works in device pairs, with one attached to the intake of the ventilator, and another attached to the return. They can also be stacked to allow for treatment of up to four patients at once – provided that the patients require the same clinical treatment in terms of oxygenation, including the oxygen mix as well as the air pressure and other factors.
This was devised by Dr. Sarah Farris, an emergency room doctor, who shared the concept with her husband Ryan Farris, a software engineer who developed the initial prototype design for 3D printing. Prisma Health is making the VESper available upon request via its printing specifications, but it should be noted that the emergency use authorization under which the FDA approved its use means that this is only intended effectively as a last resort measure – for institutions where ventilators approved under established FDA rules have already been exhausted, and no other supply or alternative is available in order to preserveve the life of patients.
Devices cleared under FDA Emergency Use Authorization (EUA) like this one are fully understood to be prototypes, and the conditions of their use includes a duty to report the results of how they perform in practice. This data contributes to the ongoing investigation of their effectiveness, and to further development and refinement of their design in order to maximize their safety and efficacy.
In addition to offering the plans for in-house 3D printing, Prisma Health has sourced donations to help print units for healthcare facilities that don’t have access to their own 3D printers. The first batch of these will be funded by a donation from the Sargent Foundation of South Carolina, but Prisma Health is seeking additional donations to fund continued research as well as additional production.
Four years ago, mathematician Vlad Voroninski saw an opportunity to remove some of the bottlenecks in the development of autonomous vehicle technology thanks to breakthroughs in deep learning.
Now, Helm.ai, the startup he co-founded in 2016 with Tudor Achim, is coming out of stealth with an announcement that it has raised $13 million in a seed round that includes investment from A.Capital Ventures, Amplo, Binnacle Partners, Sound Ventures, Fontinalis Partners and SV Angel. More than a dozen angel investors also participated, including Berggruen Holdings founder Nicolas Berggruen, Quora co-founders Charlie Cheever and Adam D’Angelo, professional NBA player Kevin Durant, Gen. David Petraeus, Matician co-founder and CEO Navneet Dalal, Quiet Capital managing partner Lee Linden and Robinhood co-founder Vladimir Tenev, among others.
Helm.ai will put the $13 million in seed funding toward advanced engineering and R&D and hiring more employees, as well as locking in and fulfilling deals with customers.
Helm.ai is focused solely on the software. It isn’t building the compute platform or sensors that are also required in a self-driving vehicle. Instead, it is agnostic to those variables. In the most basic terms, Helm.ai is creating software that tries to understand sensor data as well as a human would, in order to be able to drive, Voroninski said.
That aim doesn’t sound different from other companies. It’s Helm.ai’s approach to software that is noteworthy. Autonomous vehicle developers often rely on a combination of simulation and on-road testing, along with reams of data sets that have been annotated by humans, to train and improve the so-called “brain” of the self-driving vehicle.
Helm.ai says it has developed software that can skip those steps, which expedites the timeline and reduces costs. The startup uses an unsupervised learning approach to develop software that can train neural networks without the need for large-scale fleet data, simulation or annotation.
“There’s this very long tail end and an endless sea of corner cases to go through when developing AI software for autonomous vehicles, Voroninski explained. “What really matters is the unit of efficiency of how much does it cost to solve any given corner case, and how quickly can you do it? And so that’s the part that we really innovated on.”
Voroninski first became interested in autonomous driving at UCLA, where he learned about the technology from his undergrad adviser who had participated in the DARPA Grand Challenge, a driverless car competition in the U.S. funded by the Defense Advanced Research Projects Agency. And while Voroninski turned his attention to applied mathematics for the next decade — earning a PhD in math at UC Berkeley and then joining the faculty in the MIT mathematics department — he knew he’d eventually come back to autonomous vehicles.
By 2016, Voroninski said breakthroughs in deep learning created opportunities to jump in. Voroninski left MIT and Sift Security, a cybersecurity startup later acquired by Netskope, to start Helm.ai with Achim in November 2016.
“We identified some key challenges that we felt like weren’t being addressed with the traditional approaches,” Voroninski said. “We built some prototypes early on that made us believe that we can actually take this all the way.”
Helm.ai is still a small team of about 15 people. Its business aim is to license its software for two use cases — Level 2 (and a newer term called Level 2+) advanced driver assistance systems found in passenger vehicles and Level 4 autonomous vehicle fleets.
Helm.ai does have customers, some of which have gone beyond the pilot phase, Voroninski said, adding that he couldn’t name them.
Humio, a startup that has built a modern unlimited logging solution, announced a $20 million Series B investment today.
Dell Technologies Capital led the round with participation from previous investor Accel. Today’s investment brings the total raised to $32 million, according to the company.
Humio co-founder and CEO Geeta Schmidt says the startup wanted to build a solution that would allow companies to log everything, while reducing the overall cost associated with doing that, a tough problem due to the resource and data volume involved. The company deals with customers who are processing multiple terabytes of data per day.
“We really wanted to build an infrastructure where it’s easy to log everything and answer anything in real time. So we built an index-free logging solution which allows you to ask […] ad hoc questions over large volumes of data,” Schmidt told TechCrunch.
They are able to ingest so much data by using streaming technology, says company EVP of sales Morten Gram. “We have this real time streaming engine that makes it possible for customers to monitor whatever they know they want to be looking at. So they can build dashboards and alerts for these [metrics] that will be running in real time,” Gram explained.
What’s more, because the solution enables companies to log everything, rather than pick and choose what to log, they can ask questions about things they might not know, such as an on-going security incident or a major outage, and trace the answer from the data in the logs as the incident is happening.
Perhaps more importantly, the company has come up with technology to reduce the cost associated with processing and storing such high volumes of data. “We have thought a lot about trying to do a lot more with a lot less resources. And so, for example, one of our customers, who moved from a competitor, has gone from 80 servers to 14 doing the same volumes of data,” she said.
Deepak Jeevankumer, managing director and lead investor at Dell Technologies Capital, says that his firm recognized that Humio was solving these issues in a creative and modern way.
“Humio’s team has created a new log analysis architecture for the microservices age. This can support real-time analysis at full-speed ingest, while decreasing cost of storage and analysis by at least an order of magnitude,” he explained. “In a short-period of time, Humio has won the confidence of many Fortune 500 customers who have shifted their log platforms to Humio from legacy, decade-old architectures that do not scale for the cloud world.”
The company’s customers include Netlify, Bloomberg, HP Aruba and Michigan State University. It offers on-prem, cloud and hosted SaaS products. Today, the company also announced it was introducing an unlimited ingest plan for hosted SaaS customers.