Monthly Archives: May 2021

News: Light is the key to long-range, fully autonomous EVs

• May 25, 2021
• by:   
• in:  Blog 
• note:  no comments 

Photonic computing technology is on the brink of being commercially available and has the potential to supercharge the current roadmap of autonomous driving while also reducing its carbon footprint.

Advanced driver assistance systems (ADAS) hold immense promise. At times, the headlines about the autonomous vehicle (AV) industry seem ominous, with a focus on accidents, regulation or company valuations that some find undeserving. None of this is unreasonable, but it makes the amazing possibilities of a world of AVs seem opaque.

One of the universally accepted upsides of AVs is the potential positive impact on the environment, as most AVs will also be electric vehicles (EVs).

Industry analyst reports project that by 2023, 7.3 million vehicles (7% of the total market) will have autonomous driving capabilities requiring $1.5 billion of autonomous-driving-dedicated processors. This is expected to grow to $14 billion in 2030, when upward of 50% of all vehicles sold will be classified as SAE Level 3 or higher, as defined by the National Highway Traffic Safety Administration (NHTSA).

While photonic chips are faster and more energy efficient, fewer chips will be needed to reach SAE Level 3; however, we can expect this increased compute performance to accelerate the development and availability of fully SAE Level 5 autonomous vehicles. In that case, the market for autonomous driving photonic processors will likely far surpass the projection of $14 billion by 2030.

When you consider all of the broad-based potential uses of autonomous electric vehicles (AEVs) — including taxis and service vehicles in major cities, or the clean transport of goods on our highways — we begin to see how this technology can rapidly begin to significantly impact our environment: by helping to bring clean air to some of the most populated and polluted cities.

The problem is that AEVs currently have a sustainability problem.

To operate efficiently and safely, AEVs must leverage a dizzying array of sensors: cameras, lidar, radar and ultrasonic sensors, to name just a few. These work together, gathering data to detect, react and predict in real time, essentially becoming the “eyes” for the vehicle.

While there’s some debate surrounding the specific numbers of sensors required to ensure effective and safe AV, one thing is unanimously agreed upon: These cars will create massive amounts of data.

Reacting to the data generated by these sensors, even in a simplistic way, requires tremendous computational power — not to mention the battery power required to operate the sensors themselves. Processing and analyzing the data involves deep learning algorithms, a branch of AI notorious for its outsized carbon footprint.

To be a viable alternative, both in energy efficiency and economics, AEVs need to get close to matching gas-powered vehicles in range. However, the more sensors and algorithms an AEV has running over the course of a journey, the lower the battery range — and the driving range — of the vehicle.

Today, EVs are barely capable of reaching 300 miles before they need to be recharged, while a traditional combustion engine averages 412 miles on a single tank of gas, according to the U.S. Department of Energy. Adding autonomous driving into the mix widens this gap even further and potentially accelerates battery degradation.

Recent work published in the journal Nature Energy claims that the range of an automated electric vehicle is reduced by 10%-15% during city driving.

At the 2019 Tesla Autonomy Day event, it was revealed that driving range could be reduced by up to 25% when Tesla’s driver-assist system is enabled during city driving. This reduces the typical range for EVs from 300 miles to 225 — crossing a perceived threshold of attractiveness for consumers.

A first-principle analysis takes this a step further. NVIDIA’s AI compute solution for robotaxis, DRIVE, has a power consumption of 800 watts, while a Tesla Model 3 has an energy consumption rate of about 11.9 kWh/100 km. At the typical city speed limit of 50 km/hour (about 30 mph), the Model 3 is consuming approximately 6 kW — meaning power solely dedicated to AI compute is consuming approximately 13% of total battery power intended for driving.

This illustrates how the power-hungry compute engines used for automated EVs pose a significant problem for battery life, vehicle range and consumer adoption.

This problem is further compounded by the power overhead associated with cooling the current generation of the power-hungry computer chips that are currently used for advanced AI algorithms. When processing heavy AI workloads, these semiconductor chip architectures generate massive amounts of heat.

As these chips process AI workloads, they generate heat, which increases their temperature and, as a consequence, performance declines. More effort is then needed and energy wasted on heat sinks, fans and other cooling methods to dissipate this heat, further reducing battery power and ultimately EV range. As the AV industry continues to evolve, new solutions to eliminate this AI compute chip heat problem are urgently needed.

The chip architecture problem

For decades, we have relied on Moore’s law, and its lesser-known cousin Dennard scaling, to deliver more compute power per footprint repeatedly year after year. Today, it’s well known that electronic computers are no longer significantly improving in performance per watt, resulting in overheating data centers all over the world.

The largest gains to be had in computing are at the chip architecture level, specifically in custom chips, each for specific applications. However, architectural breakthroughs are a one-off trick — they can only be made at singular points in time in computing history.

Currently, the compute power required to train artificial intelligence algorithms and perform inference with the resulting models is growing exponentially — five times faster than the rate of progress under Moore’s law. One consequence of that is a huge gap between the amount of computing needed to deliver on the massive economic promise of autonomous vehicles and the current state of computing.

Autonomous EVs find themselves in a tug of war between maintaining battery range and the real-time compute power required to deliver autonomy.

Photonic computers give AEVs a more sustainable future

Fundamental innovation in computing and battery technology may be required to fully deliver on the promise of AEVs with the range, safety and performance demanded by consumers. While quantum computers are an unlikely short- or even medium-term solution to this AEV conundrum, there’s another, more available solution making a breakthrough right now: photonic computing.

Photonic computers use laser light, instead of electrical signals, to compute and transport data. This results in a dramatic reduction in power consumption and an improvement in critical, performance-related processor parameters, including clock speed and latency.

Photonic computers also enable inputs from a multitude of sensors to run inference tasks concurrently on a single processor core (each input encoded in a unique color), while a traditional processor can only accommodate one job at a time.

The advantage that hybrid photonic semiconductors have over conventional architectures lies within the special properties of light itself. Each data input is encoded in a different wavelength, i.e., color, while each runs on the same neural network model. This means that photonic processors not only produce more throughput compared to their electronic counterparts, but are significantly more energy efficient.

Photonic computers excel in applications that require extreme throughput with low latency and relatively low power consumption — applications like cloud computing and, potentially, autonomous driving, where the real-time processing of vast amounts of data is required.

Photonic computing technology is on the brink of becoming commercially available and has the potential to supercharge the current roadmap of autonomous driving while also reducing its carbon footprint. It’s clear that interest in the benefits of self-driving vehicles is increasing and consumer demand is imminent.

So it is crucial for us to not only consider the industries it will transform and the safety it can bring to our roads, but also ensure the sustainability of its impact on our planet. In other words, it’s time to shine a little light on autonomous EVs.

News: Hyundai is launching in-car payments in the all-electric Ioniq 5

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

Hyundai developed an in-car payment system that will debut in its upcoming all-electric Ioniq 5 crossover that will offer drivers the ability to find and pay for EV charging, food and parking — the latest example of automakers finding new ways to generate revenue and offer customers features that are typically associated with smartphones. When

Hyundai developed an in-car payment system that will debut in its upcoming all-electric Ioniq 5 crossover that will offer drivers the ability to find and pay for EV charging, food and parking — the latest example of automakers finding new ways to generate revenue and offer customers features that are typically associated with smartphones.

When the vehicle comes to North America in fall 2021, the payments system will launch with Dominoes, ParkWhiz and Chargehub, the company said Monday. The in-car payments system was just one of several new details released during the Ioniq 5’s North American debut.

The payments feature works through Bluelink, Hyundai’s branded connected car system that gives users control over various vehicle functions and services. Bluelink, which requires a subscription, is offered in three different packages that cover areas such as vehicle maintenance and alerts, remote climate control and unlocking and locking as well as destination search. Bluelink also can be linked to a user’s Google Assistant feature on their smartphone to send information to their Hyundai vehicle.

The in-car payments system will eventually expand to include other companies that fall into the charging, food and coffee on-the-go and parking categories. A company spokesperson said Hyundai will continue to add new merchants regularly via the Xevo Marketplace platform.

The Ioniq 5 is the company’s first dedicated battery-electric vehicle built on the new Electric-Global Modular Platform, or E-GMP platform. This platform is shared with Kia and is the underlying foundation of the new EV6.

If the Ioniq name sounds familiar, it’s because it already exists. In 2016, Hyundai introduced the Ioniq, a hatchback that came in hybrid, plug-in hybrid and electric versions. The Korean automaker is using that vehicle as the jumping off point for its new EV brand.

All of the vehicles under the Ioniq brand will have the E-GMP platform. The Ioniq 5 is based on Hyundai’s Concept 45, a monocoque-style body crossover that the company unveiled in 2019 at the International Motor Show in Frankfurt. Designers of the Concept 45 leaned on some of the lines and characteristics from Hyundai’s first concept, the 1974 Pony Coupe. The “45” name comes, in part, from the 45-degree angles at the front and rear of the vehicle.

Hyundai has yet to release pricing for the Ioniq 5.

News: Florida’s ban on bans will test First Amendment rights of social media companies

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

Florida governor Ron DeSantis has signed into law a restriction on social media companies’ ability to ban candidates for state offices and news outlets, and in doing so offered a direct challenge to those companies’ perceived free speech rights. The law is almost certain to be challenged in court as both unconstitutional and in direct

Florida governor Ron DeSantis has signed into law a restriction on social media companies’ ability to ban candidates for state offices and news outlets, and in doing so offered a direct challenge to those companies’ perceived free speech rights. The law is almost certain to be challenged in court as both unconstitutional and in direct conflict with federal rules.

The law, Florida Senate Bill 7072, provides several new checks on tech and social media companies. Among other things:

• Platforms cannot ban or deprioritize candidates for state office
• Platforms cannot ban or deprioritize any news outlet meeting certain size requirements
• Platforms must be transparent about moderation processes and give users notice of moderation actions
• Users and the state will have the right to sue companies that violate the law

The law establishes rules affecting these companies’ moderation practices; that much is clear. But whether doing so amounts to censorship — actual government censorship, not the general concept of limitation frequently associated with the word — is an open question, if a somewhat obvious one, that will likely be forced by legal action against SB 7072.

While there is a great deal of circumstantial precedent and analysis, the problem of “are moderation practices of social media companies protected by the First Amendment” is as yet unsettled. Legal scholars and existing cases fall strongly on the side of “yes,” but there is no single definitive precedent that Facebook or Twitter can point to.

The First Amendment argument starts with the idea that although social media are very unlike newspapers or book publishers, they are protected in much the same way by the Constitution from government interference. “Free speech” is a term that is interpreted extremely liberally, but if a company spending money is considered a protected expression of ideas, it’s not a stretch to suggest that same company applying a policy of hosting or not hosting content should be as well. If it is, then the government is prohibited from interfering with it beyond very narrow definitions of unprotected speech (think shouting “fire” in a crowded theater). That would sink Florida’s law on constitutional grounds.

The other conflict is with federal law, specifically the much-discussed Section 230, which protects companies from being liable for content they publish (i.e. the creator is responsible instead), and also for the choice to take down content via rules of their own choice. As the law’s co-author Senator Ron Wyden (D-OR) has put it, this gives those companies both a shield and a sword with which to do battle against risky speech on their platforms.

But SB 7072 removes both sword and shield: it would limit who can be moderated, and also creates a novel cause for legal action against the companies for their remaining moderation practices.

Federal and state law are often in disagreement, and there is no handbook for how to reconcile them. On one hand, witness raids of state-legalized marijuana shops and farms by federal authorities. On the other, observe how strong consumer protection laws at the state level aren’t preempted by weaker federal ones because to do so would put people at risk.

On the matter of Section 230 it’s not straightforward who is protecting whom. Florida’s current state government claims that it is protecting “real Floridians” against the “Silicon Valley elites.” But no doubt those elites (and let us be candid — that is exactly what they are) will point out that in fact this is a clear-cut case of government overreach, censorship in the literal sense.

These strong legal objections will inform the inevitable lawsuits by the companies affected, which will probably be filed ahead of the law taking effect and aim to have it overturned.

Interestingly, two companies that will not be affected by the law are two of the biggest, most uncompromising corporations in the world: Disney and Comcast. Why, you ask? Because the law has a special exemption for any company “that owns and operates a theme park or entertainment complex” of a certain size.

That’s right, there’s a Mouse-shaped hole in this law — and Comcast, which owns Universal Studios, just happens to fit through as well. Notably this was added in an amendment, suggesting two of the largest employers in the state were unhappy at the idea of new liabilities for any of their digital properties.

This naked pandering to local corporate donors puts proponents of this law at something of an ethical disadvantage in their righteous battle against the elites, but favor may be moot in a few months’ time when the legal challenges, probably being drafted at this moment, call for an injunction against SB 7072.

News: Zocdoc says ‘programming errors’ exposed access to patients’ data

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

Zocdoc says it has fixed a bug that allowed current and former staff at doctor’s offices and dental practices to access patient data because their user accounts weren’t properly decommissioned. The New York-based company revealed the issue in a letter to the California attorney general’s office, which requires companies with more than 500 residents of

Zocdoc says it has fixed a bug that allowed current and former staff at doctor’s offices and dental practices to access patient data because their user accounts weren’t properly decommissioned.

The New York-based company revealed the issue in a letter to the California attorney general’s office, which requires companies with more than 500 residents of the state affected by a security lapse or breach to disclose the incident.

Zocdoc, which lets prospective patients book appointments with doctors and dentists, said that it gives each medical or dental practice usernames and passwords for its staff to access appointments made through Zocdoc, but that “programming errors” — essentially a software bug in Zocdoc’s own systems — “allowed some past or current practice staff members to access the provider portal after their usernames and passwords were intended to be removed, deleted or otherwise limited.”

The letter confirmed that patient data stored in Zocdoc’s portal could have been accessed, including a patient’s name, email address, phone number, and the times and dates of their appointments, but also other data that may have been shared with the practice — such as insurance details, Social Security numbers and details of the patient’s medical history.

But Zocdoc said payment card numbers, radiological or diagnostic reports, and medical records were not taken, since it does not store this data.

In an email, Zocdoc spokesperson Sandra Glading said that the company discovered the bug in August 2020, but “due to the complexity of the code, it took a significant amount of investigation to determine which, if any, practices and users were affected and how.” The company said it provided notice to the California’s attorney general’s office “as soon as was practicable.”

Zocdoc said it has “detailed logs that can detect exploitation of any data, including any potential exploitation of this vulnerability,” and that after a review of those logs and other investigative work, “we have no indication, at this time, that any personal information was misused in any way.”

Around 6 million users access Zocdoc a month, the company said.

If this incident sounds vaguely familiar, it’s because this was a near-identical security issue to one Zocdoc reported in 2016. A letter filed at the time cited similar “programming errors” that allowed staff at medical providers to improperly access patient data.

News: Daily Crunch: Police search 2 Twitter offices in India after politician receives warning label

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

Hello friends and welcome to Daily Crunch, bringing you the most important startup, tech and venture capital news in a single package.

To get a roundup of TechCrunch’s biggest and most important stories delivered to your inbox every day at 3 p.m. PDT, subscribe here.

Welcome back to Daily Crunch. It’s Monday, May 24, and all I can think about is how much I want a Surface Duo now that it can do two-screen gaming. And one of those new iMacs. I don’t need either, of course, but that doesn’t stop my coveting both of the gadgets. Alas.

Regardless, it was a super busy start to the week, with lots of startup funding rounds, more in the long-running saga of governments trying to control social media platforms, even more IPO news and the latest troubles with Tesla. Let’s cut the chatter and dive in. — Alex

The TechCrunch Top 3

• Governments vs. Tech: Indian police tied to the central government showed up at two different Twitter offices today, in what appeared to be an intimidation effort following Twitter’s decision to not unlabel a tweet from a member of the current ruling party as manipulated media.

A few things here. First, India is not alone in trying to force social media companies to behave as local governments want them to. That said, what the current Indian government is doing is particularly egregious and doesn’t bode well for the country’s tech ecosystem as a whole.

• Tesla owes Norway: American electric car darling Tesla appears to be in hot water with Norway after a “Norwegian conciliation council” ordered the company to pay $16,000 each to thousands of Model S owners after “it found that a software update led to longer charging times.” Ouch. Tesla will have to sell lots of American regulatory credits to cover that loss. Norway is a key market for EVs.
• U.S. cities buy abuse-linked tech: From the “you should read this” files, the latest report from our own Zack Whittaker and IPVM states that “at least a hundred U.S. counties, towns and cities have bought Chinese-made surveillance systems that the U.S. government has linked to human rights abuses.” Not good.

Startups and VC

As always we’re picking and choosing the best rounds from the day, so feel free to scrounge around the blog if you need even more!

Solidus Labs raises $20M for crypto-snooping: As the value of cryptocurrencies rose in the past year, so too did business at Solidus Labs, which detects “volume and price manipulation” among bitcoin and its brethren. Per its CEO, Asaf Meir, his company saw a “400% increase in inbound demand over 2020.” That sounds about right. Also, Solidus should drop a monthly report on the level of manipulation on every exchange and crypto. That would rule.

Fireflies.ai raises $14M to record, transcribe and connect your meetings: Former Acceleprise company Fireflies is building software that will record and transcribe your meetings, and then connect the text — and perhaps the embedded tasks — to other bits of software. It’s interesting, and growing like a weed. So Khosla helped put $14 million into it.

Mono raises $2M to power African fintech: From building the Plaid for Africa to “power the internet economy in Africa,” Mono is not short on vision. And now it has had its accounts refreshed to pursue its plans to “[streamline] various financial data in a single API for companies and third-party developers.” APIs are cool. Fintech is cool. Fintech APIs are extra cool. That’s our take.

Flat6Labs raises $13.2M to fund Egyptian startups: This is fund news, but it’s small enough that it fits inside the startup section today. In short, since 2011 Flat6Labs has been an accelerator in Egypt and Tunisia. And now it has a new checkbook to play with.

Inside Zeta Global’s IPO: Finally from the startup world today, Zeta Global is going public. It’s an offering that could set the tone for the martech world for some time to come. So, we dug into its numbers a bit tardily to figure out just what Zeta has that public investors might want.

When to walk away from a VC who wants to invest in your startup

Ofri Ben-Porat flew from London to NYC to meet potential investors, but at the last minute, one canceled, claiming illness. Moments later, he received a DocSend push notification informing him that said VC had just opened the pitch deck he’d sent days before.

Undeterred, he showed up anyway and pretended he hadn’t received their email. The discussion went well; after he flew home, the VCs offered pre-terms and due diligence, “but ultimately, I didn’t feel right taking money from them,” says Ben-Porat.

Securing the right amount of funding at the right moment can make or break a startup, but founders who can’t identify red flags — or worse, ignore them — will live to regret it.

(Extra Crunch is our membership program, which helps founders and startup teams get ahead. You can sign up here.)

Big Tech Inc.

What has a zillion hands and likes to copy its friends? Facebook! This time, however, Facebook could be doing something interesting. TechCrunch dug through some creator-friendly feature work that Big Blue is putting into its TikTok clone. So it’s still running the copy machine at full tilt, just with a few upgrades.

Turning back to Apple, not everything is M1 chips and purple iPhones. Some things are less good at the Cupertino-based technology leviathan. Today TechCrunch reported on a few different macOS vulnerabilities that, frankly, don’t sound good. What’s the old adage? Buy a PC; they just work?

In happier Apple news, the company has a pile of new software updates for your enjoyment, especially if you are an iPhone or iPad user.

Don’t fret, Microsoft fans, we have something for you as well. Namely a review of the new Surface Laptop 4. It’s pretty darn good, keeping all its predecessor’s weaknesses and strengths, with new guts.

Wrapping, ByteDance has another chart-topping app; Airbnb is doubling down on guest flexibility, though we have questions; SiriusXM is partnering with TikTok on a new channel; and SensorTower is making sure that there is at least some M&A to report on.

TechCrunch Experts: Email Marketing

Image Credits: Getty Images

TechCrunch Experts is still collecting survey responses to help us identify the top email marketers in tech!

At this time, we’re not looking for self-nominations — we’re only seeking nominations from clients. We want to hear all about your experience and how you found the right expert for your needs. Fill out the survey here.

We’re excited to move this project forward. Visit techcrunch.com/experts to find out more!

TC Eventful

Last year, we held our first dedicated space startup event, TC Sessions: Space, featuring some of the industry’s top founders and leaders, including Rocket Lab’s Peter Beck, Lockheed Martin’s Lisa Callahan, Amazon’s Dave Limp, NASA’s Kathy Lueders and many more. This year, we’re excited to announce we’re doing it again with TC Sessions: Space 2021, happening virtually December 14-15.

TC Sessions: Mobility 2021 is right around the corner of your calendar (June 9). If you want to place your ground-breaking, edge-cutting, envelope-pushing (no extra charge for clichés) early-stage startup in front of the world’s leading mobility movers, shakers and makers you gotta hustle. You have just one week left to buy one of our remaining three Startup Exhibitor Packages.

News: Lordstown Motors slashes production forecast for its electric pickup

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

Lordstown Motors’ cash-rich SPAC dreams have turned out to be nothin’ more than wishes. The automaker reported Monday a disappointing first-quarter earnings that was a pile-up of red ink-stained negativity. Lowlights include higher-than-expected forecasted expenses, a need to raise more capital, and lower-than-anticipated production of its Endurance vehicle this year – from around 2,200 vehicles

Lordstown Motors’ cash-rich SPAC dreams have turned out to be nothin’ more than wishes. The automaker reported Monday a disappointing first-quarter earnings that was a pile-up of red ink-stained negativity.

Lowlights include higher-than-expected forecasted expenses, a need to raise more capital, and lower-than-anticipated production of its Endurance vehicle this year – from around 2,200 vehicles to just 1,000. In short, the company is set to consume more cash than the street expected, and is further from mass production of its first vehicle than promised.

The value of the company, which went public via a SPAC last year, has fallen sharply from its post-combination highs. Today its shares are off another 7% after the close of trading, thanks to its Q1 2021 report.

Investors were not thrilled with the company that 11 months ago showed off a prototype of Endurance, the all-electric pickup truck that it has bet its future on.

Lordstown Motors is an offshoot of CEO Steve Burns’ other company, Workhorse Group, a battery-electric transportation technology company that is also a publicly traded company. Workhorse is a small company that was founded in 1998 and has struggled financially at various points. Its offshoot, Lordstown Motors has previously said it planned 20,000 electric trucks annually, starting in the second half of 2021, at the former GM Assembly Plant in Lordstown, Ohio. Lordstown Motors acquired in November the 6.2 million-square-foot factory from GM.

Production woes, capital concerns

Lordstown reported a $125 million net loss on zero revenue, along with capital expenditures of $53 million in the first quarter. And yet, Lordstown had little to show for its outsized spending.

The company said in a release that it would still begin production of its Endurance electric pickup truck this year but that its output “would be at best 50% of our prior expectations.” That fact on top of its massive cash drawdown was hardly investor catnip.

“Our research indicates a very robust demand for our vehicles,” Burns told investors during a call Monday. “However, capital may limit our ability to make as many vehicles as we would like, and as such, we are constantly evaluating our capital needs, and the various types of capital available to us, including strategic capital.”

The EV company anticipates ending 2021 with just $50 to $75 million in liquidity, despite its recent SPAC combination that helped capitalize its operations. Lordstown finished 2020 with $630 million in cash; it wrapped Q1 2021 with $587 million. The company anticipates “capital expenditures of between $250 [million] and $275 million,” in addition to its regular cash consumption from operating costs.

Burns said the company was in discussions with an unnamed financial entity for asset-backed financing.

“We have zero debt and we have a lot of assets, and we’re buying a lot of parts. So there’s folks that want to finance that,” he said. Lordstown is also still pursuing an Advanced Technology Vehicles Manufacturing loan from the U.S. Department of Energy. Executives said DOE has done several rounds on due diligence but declined to comment on the timing, though Burns said multiple times that Tesla wouldn’t exist had it not gotten an ATVM loan in January 2010.

For post-combination SPAC companies, Lordstown’s lackluster results and bearish trading are yet more indication that the boom in using blank-check agreements to take EV and other automotive-focused companies public was perhaps premature.

Lordstown announced its SPAC merger in September 2020 with a market value of $1.6 billion. Its shares soared to $31.80 apiece at their 52-week highs. Today they are worth $8.77.

Burns lauded the company’s purported competitive advantages, including its hub motor architecture and physical simplicity, which he said would translate into a lower cost of ownership. But the company has stiff competition from new EV entrants Rivian and Tesla (should the Cybertruck ever hit production) and legacy automakers like Ford, which debuted the electric model of its nameplate F-150 truck model earlier this month with a price point under $40,000.

But Burns reiterated his feeling that the company was on par with its competitors and that it wants to be “ready to pounce” in response to vehicle demand. The CEO also said he was confident that the truck would hit the 250-mile target range, though this is less than both the Rivian R1T and the Ford F-150 Lightning.

Lordstown also gave a brief update on pre-orders following its announcement in January that it hit a milestone of 100,000 preorders. Burns said around 30,000 of those had been converted to what it’s calling “vehicle purchase agreements,” but he demurred on exactly how many of those customers have paid anything, saying only that “many of those” agreements, including some kind of down payment.

The company also began work on its second vehicle, an electric van, with a prototype anticipated later this summer.

Financial results

Turning to Lordstown’s first quarter performance, we’re observing a pre-revenue company in the weeds of testing, and scaling production for an incredibly complex product. Which is an expensive endeavor.

Here’s the chart:

Image Credits: Lordstown

The company’s greater-than-before sales and administrative costs are whatever compared to its spiraling research and development spend. For investors holding onto Lordstown shares in hopes of its eventual early construction runs leading to mass production that is now further in the future, it’s a tough income statement to digest.

In the first quarter of 2021 the company spent around $91,000 in research and development expenses. “The higher than expected R&D spend is largely from higher part costs from a supply chain that remains under duress, from collocations, and which impacted our beta costs, higher costs of shipping included expedited shipping and greater use of temporary external engineering,” Lordstown CFO Julio Rodriguez said.

Company executives also briefly addressed accusations by short seller Hindenburg Research, who claimed the automaker was faking pre-orders of its vehicles. Hindenburg said that “extensive research reveals that the company’s orders appear largely fictitious and used as a prop to raise capital and confer legitimacy.”

Burns told investors that the company established a special independent committee to investigate the allegations in the report. This is in addition to a separate investigation from the U.S. Securities and Exchange Commission, which the company is cooperating with, he said.

In the wake of Lordstown’s results, however, shares of Tesla and Nikola were largely flat.

News: Flush with $42M, hot AI startup Faculty plans to hoover-up more PhDs… and steer clear of politics

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

In the wake of the news that UK-based AI startup Faculty has raised $42.5 million in a growth funding round, I teased out more from CEO and co-founder Marc Warner on what his plans are for the company. Faculty seems to have an uncanny knack of winning UK government contracts, after helping Boris Johnson win

In the wake of the news that UK-based AI startup Faculty has raised $42.5 million in a growth funding round, I teased out more from CEO and co-founder Marc Warner on what his plans are for the company.

Faculty seems to have an uncanny knack of winning UK government contracts, after helping Boris Johnson win his Vote Leave campaign and thus become Prime Minister. It’s even helping sort out the mess that Brexit has subsequently made of the fishing industry, problems with the NHS, and telling global corporates like Red Bull and Virgin Media what to suggest to their customers. Meanwhile, it continues to hoover up Ph.D. graduates at a rate of knots to work on its AI platform.

But, speaking to me over a call, Warner said the company no longer has plans to enter the political sphere again: “Never again. It’s very controversial. I don’t want to make out that I think politics is unethical. Trying to make the world better, in whatever dimension you can, is a good thing … But from our perspective, it was, you know, ‘noisy,’ and our goal as an organization is, despite current appearances to the contrary, is not to spend tonnes of time talking about this stuff. We do believe this is an important technology that should be out there and should be in a broader set of hands than just the tech giants, who are already very good at it.”

On the investment, he said: “Fundamentally, the money is about doubling down on the UK first and then international expansion. Over the last seven years or so we have learned what it takes to do important AI, impactful AI, at scale. And we just don’t think that there’s actually much of it out there. Customers are rightly sometimes a bit skeptical, as there’s been hype around this stuff for years and years. We figured out a bunch of the real-world applications that go into making this work so that it actually delivers the value. And so, ultimately, the money is really just about being able to build out all of the pieces to do that incredibly well for our customers.”

He said Faculty would be staying firmly HQ’d in the UK to take advantage of the UK’s talent pool: “The UK is a wonderful place to do AI. It’s got brilliant universities, a very dynamic startup scene. It’s actually more diverse than San Francisco. There’s government, there’s finance, there are corporates, there’s less competition from the tech giants. There’s a bit more of a heterogeneous ecosystem. There’s no sense in which we’re thinking, ‘Right, that’s it, we’re up and out!’. We love working here, we want to make things better. We’ve put an enormous amount of effort into trying to help organizations like the government and the NHS, but also a bunch of UK corporates in trying to embrace this technology, so that’s still going to be a terrifically important part of our business.”

That said, Faculty plans to expand abroad: “We’re going to start looking further afield as well, and take all of the lessons we’ve learned to the US, and then later Europe.”

But does he think this funding round will help it get ahead of other potential rivals in the space? “We tend not to think too much in terms of rivals,” he says. “The next 20 years are going to be about building intelligence into the software that already exists. If you look at the global market cap of the software businesses out there, that’s enormous. If you start adding intelligence to that, the scale of the market is so large that it’s much more important to us that we can take this incredibly important technology and deploy it safely in ways that actually improve people’s lives. It could be making products cheaper or helping organizations make their services more efficient.”

If that’s the case then does Faculty have any kind of ethics panel overseeing its work? “We have an internal ethics panel. We have a set of principles and if we think a project might violate those principles, it gets referred to that ethics panel. It’s randomly selected from across faculty. So we’re quite careful about the projects that we work on and don’t. But to be honest, the vast majority of stuff that’s going on is very vanilla. They are just clearly ‘good for the world’ projects. The vast majority of our work is doing good work for corporate clients to help them make their businesses that bit more efficient.”

I pressed him to expand on this issue of ethics and the potential for bias. He says Faculty “builds safety in from a start. Oddly enough, the reason I first got interested in AI was reading Nick Bostrom’s work about superintelligence and the importance of AI safety. And so from the very, very first fellowship [Faculty AI researchers are called Fellows] all the way back in 2014, we’ve taught the fellows about AI safety. Over time, as soon as we were able, we started contributing to the research field. So, we’ve published papers in all of the biggest computer science conferences Neurips, ICM, ICLR, on the topic of AI safety. How to make algorithms fair, private, robust and explainable. So these are a set of problems that we care a great deal about. And, I think, are generally ‘underdone’ in the wider ecosystem. Ultimately, there shouldn’t be a separation between performance and safety. There is a bit of a tendency in other companies to say, ‘Well, you can either have performance, or you can have safety.’ But of course, we know that’s not true. The cars today are faster and safer than the Model T Ford. So it’s a sort of a false dichotomy. We’ve invested a bunch of effort in both those capabilities, so we obviously want to be able to create a wonderful performance for the task at hand, but also to ensure that the algorithms are fair, private, robust and explainable wherever required.”

That also means, he says, that AI might not always be the ‘bogeyman’ the phrase implies: “In some cases, it’s probably not a huge deal if you’re deciding whether to put a red jumper or a blue jumper at the top of your website. There are probably not huge ethical implications in that. But in other circumstances, of course, it’s critically important that the algorithms are safe and are known to be safe and are trusted by both the users and anyone else who encounters them. In a medical context, obviously, they need to be trusted by the doctors and the patients need to make sure they actually work. So we’re really at the forefront of deploying that stuff.”

Last year the Guardian reported that Faculty had won seven government contracts in 18 months. To what does he attribute this success? “Well, I mean, we lost an enormous number more! We are a tiny supplier to government. We do our best to do work that is valuable to them. We’ve worked for many many years with people at the home office,” he tells me.

“Without wanting to go into too much detail, that 18 months stretches over multiple Prime Ministers. I was appointed to the AI Council under Theresa May. Any sort of insinuations on this are just obviously nonsense. But, at least historically, most of our work was in the private sector and that continues to be critically important for us as an organization. Over the last year, we’ve tried to step up and do our bit wherever we could for the public sector. It’s facing such a big, difficult situation around COVID, and we’re very proud of the things we’ve managed to accomplish with the NHS and the impact that we had on the decisions that senior people were able to undertake.”

Returning to the issue of politics I asked him if he thought – in the wake of events such as Brexit and the election of Donald Trump, which were both affected by AI-driven political campaigning – AI is too dangerous to be applied to that arena? He laughed: “It’s a funny old funny question… It’s a really odd way to phrase a question. AI is just a technology. Fundamentally, AI is just maths.”

I asked him if he thought the application of AI in politics had had an outsized or undue influence, on the way that political parties have operated in the last few years: “I’m afraid that is beyond my knowledge,” he says. But does Faculty have regrets about working in the political sphere?

“I think we’re just focused on our work. It’s not that we have strong feelings, either way, it’s just that from our perspective, it’s much, much more interesting to be able to do the things that we care about, which is deploying AI in the real world. It’s a bit of a boring answer! But it is truly how we feel. It’s much more about doing the things we think are important, rather than judging what everyone else is doing.”

Lastly, we touched on the data science capabilities of the UK and what the new fund-raising will allow the company to do.

He said: “We started an education program. We have roughly 10% of the UK’s PhDs in physics, maths, engineering, applying to the program. Roughly 400 or so people have been through that program and we plan to expand that further so that more and more people get the opportunity to start a career in data science. And then inside Faculty specifically, we think we’ll be able to create 400 new jobs in areas like software engineering, data science, product management. These are very exciting new possibilities for people to really become part of the technology revolution. I think there’s going to be a wonderful like new energy in Faculty, and hopefully a positive small part in increasing the UK tech ecosystem.”

Warner comes across as sincere in his thoughts about the future of AI and is clearly enthusiastic about where Faculty can take the whole field next, both philosophically and practically. Will Faculty soon be challenging that other AI leviathan, DeepMind, for access to all those Ph.D.s? There’s no doubt it will.

News: 2 CEOs are better than 1

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

The co-CEO model can work — with the right ingredients in place.

Netflix has two CEOs: Co-founder Reed Hastings oversees the streaming side of the company, while Ted Sarandos guides Netflix’s content.

Warby Parker has co-CEOs as well — its co-founders went to college together. Other companies like the tech giant Oracle and luggage maker Away have shifted from having co-CEOs in recent years, sparking a wave of headlines suggesting that the model is broken.

While there isn’t a lot of research on companies with multiple CEOs, the data is more promising than the headlines would suggest. One study on public companies with co-CEOs revealed that the average tenure for co-CEOs, about 4.5 years, was comparable to solitary CEOs, “suggesting that this arrangement is more stable than previously believed.”

The study’s authors also found that co-CEOs were spread across industry types and that splitting the role can “complement each other in terms of educational background or executive responsibilities.”

I serve as co-CEO of an organic meal delivery company with my sister Laureen. Having two CEOs has helped us take Fresh n’ Lean to new heights. We closed 2020 with $87 million in revenue, more than double from the year before, and project similar growth this year.

We complement each other well, and the results bear that out. During the decade that we’ve served as co-CEOs, the company has grown from a very small team to 475 full-time employees and 40 part-time employees. We’ve delivered more than 17 million meals, launched four different meal lines, expanded our retail offerings, partnered with some great names in sports and fitness, and saw our annual revenues climb exponentially.

The leadership structure isn’t for every company, but it’s been a great fit for Fresh n’ Lean. Here’s why.

Divide and conquer to shorten your learning curve by 50%

Laureen launched the company in 2010 out of her one-bedroom apartment.

“Those early years were especially tough,” she said. “I consistently worked 20-hour days as I performed just about every role — cooking dishes, preparing labels, making deliveries and performing customer service duties. I was devoting so much energy into product, packaging and logistics, but in order for the company to grow, I needed help with marketing, tech and finance.”

Those areas happened to be my strengths. There was too much for one person to oversee as CEO and not enough hours in the day. But given the equal challenges that both sides of the company presented and the trust we shared, it made sense for us to be side by side on the organizational chart.

News: Malware caught using a macOS zero-day to secretly take screenshots

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

Almost exactly a month ago, researchers revealed a notorious malware family was exploiting a never-before-seen vulnerability that let it bypass macOS security defenses and run unimpeded. Now, some of the same researchers say another malware can sneak onto macOS systems, thanks to another vulnerability. Jamf says it found evidence that the XCSSET malware was exploiting a

Almost exactly a month ago, researchers revealed a notorious malware family was exploiting a never-before-seen vulnerability that let it bypass macOS security defenses and run unimpeded. Now, some of the same researchers say another malware can sneak onto macOS systems, thanks to another vulnerability.

Jamf says it found evidence that the XCSSET malware was exploiting a vulnerability that allowed it access to parts of macOS that require permission — such as accessing the microphone, webcam or recording the screen — without ever getting consent.

XCSSET was first discovered by Trend Micro in 2020 targeting Apple developers, specifically their Xcode projects that they use to code and build apps. By infecting those app development projects, developers unwittingly distribute the malware to their users, in what Trend Micro researchers described as a “supply-chain-like attack.” The malware is under continued development, with more recent variants also targeting Macs running the newer M1 chip.

Once the malware is running on a victim’s computer, it uses two zero-days — one to steal cookies from the Safari browser to get access to a victim’s online accounts, and another to quietly install a development version of Safari, allowing the attackers to modify and snoop on virtually any website.

But Jamf says the malware was exploiting a previously undiscovered third zero-day in order to secretly take screenshots of the victim’s screen.

macOS is supposed to ask the user for permission before it allows any app — malicious or otherwise — to record the screen, access the microphone or webcam, or open the user’s storage. But the malware bypassed that permissions prompt by sneaking in under the radar by injecting malicious code into legitimate apps.

Jamf researchers Jaron Bradley, Ferdous Saljooki, and Stuart Ashenbrenner explained in a blog post, shared with TechCrunch, that the malware searches for other apps on the victim’s computer that are frequently granted screen-sharing permissions, like Zoom, WhatsApp and Slack, and injects malicious screen recording code into those apps. This allows the malicious code to “piggyback” the legitimate app and inherit its permissions across macOS. Then, the malware signs the new app bundle with a new certificate to avoid getting flagged by macOS’ built-in security defenses.

The researchers said that the malware used the permissions prompt bypass “specifically for the purpose of taking screenshots of the user’s desktop,” but warned that it was not limited to screen recording. In other words, the bug could have been used to access the victim’s microphone, webcam or capture their keystrokes, such as passwords or credit card numbers.

It’s not clear how many Macs the malware was able to infect using this technique. But Apple confirmed to TechCrunch that it fixed the bug in macOS 11.4, which was made available as an update today.

News: Deep Science: Robots, meet world

• May 24, 2021
• by:   
• in:  Blog 
• note:  no comments 

Research papers come out far too frequently for anyone to read them all. That’s especially true in the field of machine learning, which now affects (and produces papers in) practically every industry and company. This column aims to collect some of the most relevant recent discoveries and papers — particularly in, but not limited to,

Research papers come out far too frequently for anyone to read them all. That’s especially true in the field of machine learning, which now affects (and produces papers in) practically every industry and company. This column aims to collect some of the most relevant recent discoveries and papers — particularly in, but not limited to, artificial intelligence — and explain why they matter.

This edition, we have a lot of items concerned with the interface between AI or robotics and the real world. Of course most applications of this type of technology have real-world applications, but specifically this research is about the inevitable difficulties that occur due to limitations on either side of the real-virtual divide.

One issue that constantly comes up in robotics is how slow things actually go in the real world. Naturally some robots trained on certain tasks can do them with superhuman speed and agility, but for most that’s not the case. They need to check their observations against their virtual model of the world so frequently that tasks like picking up an item and putting it down can take minutes.

What’s especially frustrating about this is that the real world is the best place to train robots, since ultimately they’ll be operating in it. One approach to addressing this is by increasing the value of every hour of real-world testing you do, which is the goal of this project over at Google.

In a rather technical blog post the team describes the challenge of using and integrating data from multiple robots learning and performing multiple tasks. It’s complicated, but they talk about creating a unified process for assigning and evaluating tasks, and adjusting future assignments and evaluations based on that. More intuitively, they create a process by which success at task A improves the robots’ ability to do task B, even if they’re different.

Humans do it — knowing how to throw a ball well gives you a head start on throwing a dart, for instance. Making the most of valuable real-world training is important, and this shows there’s lots more optimization to do there.

Another approach is to improve the quality of simulations so they’re closer to what a robot will encounter when it takes its knowledge to the real world. That’s the goal of the Allen Institute for AI’s THOR training environment and its newest denizen, ManipulaTHOR.

Image Credits: Allen Institute

Simulators like THOR provide an analogue to the real world where an AI can learn basic knowledge like how to navigate a room to find a specific object — a surprisingly difficult task! Simulators balance the need for realism with the computational cost of providing it, and the result is a system where a robot agent can spend thousands of virtual “hours” trying things over and over with no need to plug them in, oil their joints and so on.