FreshRSS

🔒
❌ About FreshRSS
There are new available articles, click to refresh the page.
Today — April 22nd 2021Your RSS feeds

Lina Khan’s timely tech skepticism makes for a refreshingly friendly FTC confirmation hearing

By Devin Coldewey

One never knows how a confirmation hearing will go these days, especially one for a young outsider nominated to an important position despite challenging the status quo and big business. Lina Khan, just such a person up for the position of FTC Commissioner, had a surprisingly pleasant time of it during today’s Senate Commerce Committee confirmation hearing — possibly because her iconoclastic approach to antitrust makes for good politics these days.

Khan, an associate professor of law at Columbia, is best known in the tech community for her incisive essay “Amazon Antitrust’s Paradox,” which laid out the failings of regulatory doctrine that have allowed the retail giant to progressively dominate more and more markets. (She also recently contributed to a House report on tech policy.)

When it was published, in 2018, the feeling that Amazon had begun to abuse its position was, though commonplace in some circles, not really popular in the Capitol. But the growing sense that laissez-faire or insufficient regulations have created monsters in Amazon, Google, and Facebook (to start) has led to a rare bipartisan agreement that we must find some way, any way will do, of putting these upstart corporations back in their place.

This in turn led to a sense of shared purpose and camaraderie in the confirmation hearing, which was a triple header: Khan joined Bill Nelson, nominated to lead NASA, and Leslie Kiernan, who would join the Commerce Department as General Counsel, for a really nice little three-hour chat.

Khan is one of several in the Biden administration who signal a new approach to taking on Big Tech and other businesses that have gotten out of hand, and the questions posed to her by Senators from both sides of the aisle seemed genuine and got genuinely satisfactory answers from a confident Khan.

She deftly avoided a few attempts to bait her — including one involving Section 230; wrong Commission, Senator — and her answers primarily reaffirmed her professional opinion that the FTC should be better informed and more preemptive in its approach to regulating these secretive, powerful corporations.

Here are a few snippets representative of the questioning and indicative of her positions on a few major issues (answers lightly edited for clarity):

On the FTC getting involved in the fight between Google, Facebook, and news providers:

“Everything needs to be on the table. Obviously local journalism is in crisis, and i think the current COVID moment has really underscored the deep democratic emergency that is resulting when we don’t have reliable sources of local news.”

She also cited the increasing concentration of ad markets and the arbitrary nature of, for example, algorithm changes that can have wide-ranging effects on entire industries.

Lina Khan, commissioner of the Federal Trade Commission (FTC) nominee for U.S. President Joe Biden, speaks during a Senate Commerce, Science and Transportation Committee confirmation hearing in Washington, D.C.

Image Credits: Graeme Jennings/Washington Examiner/Bloomberg / Getty Images

On Clarence Thomas’s troubling suggestion that social media companies should be considered “common carriers”:

“I think it prompted a lot of interesting discussion,” she said, very diplomatically. “In the Amazon article, I identified two potential pathways forward when thinking about these dominant digital platforms. One is enforcing competition laws and ensuring that these markets are competitive.” (i.e. using antitrust rules)

“The other is, if we instead recognize that perhaps there are certain economies of scale, network externalities that will lead these markets to stay dominated by a very few number of companies, then we need to apply a different set of rules. We have a long legal tradition of thinking about what types of checks can be applied when there’s a lot of concentration and common carriage is one of those tools.”

“I should clarify that some of these firms are now integrated in so many markets that you may reach for a different set of tools depending on which specific market you’re looking at.”

 

(This was a very polite way of saying common carriage and existing antitrust rules are totally unsuitable for the job.)

On potentially reviewing past mergers the FTC approved:

“The resources of the commission have not really kept pace with the increasing size of the economy, as well as the increasing size and complexity of the deals the commission is reviewing.”

“There was an assumption that digital markets in particular are fast moving so we don’t need to be concerned about potential concentration in the markets, because any exercise of power will get disciplined by entry and new competition. Now of course we know that in the markets you actually have significant network externalities in ways that make them more sticky. In hindsight there’s a growing sense that those merger reviews were a missed opportunity.”

(Here Senator Blackburn (R-TN) in one of the few negative moments fretted about Khan’s “lack of experience in coming to that position” before asking about a spectrum plan — wrong Commission, Senator.)

On the difficulty of enforcing something like an order against Facebook:

“One of the challenges is the deep information asymmetry that exists between some of these firms and enforcers and regulators. I think it’s clear that in some instances the agencies have been a little slow to catch up to the underlying business realities and the empirical realities of how these markets work. So at the very least ensuring the agencies are doing everything they can to keep pace is gonna be important.”

“In social media we have these black box algorithms, proprietary algorithms that can sometimes make it difficult to know what’s really going on. The FTC needs to be using its information gathering capacities to mitigate some of these gaps.”

On extending protections for children and other vulnerable groups online:

Some of these dangers are heightened given some of the ways in which the pandemic has rendered families and children especially dependent on some of these [education] technologies. So I think we need to be especially vigilant here. The previous rules should be the floor, not the ceiling.


Overall there was little partisan bickering and a lot of feeling from both sides that Khan was, if not technically experienced at the job (not rare with a coveted position like FTC Commissioner), about as competent a nominee as anyone could ask for. Not only that but her highly considered and fairly assertive positions on matters of antitrust and competition could help put Amazon and Google, already in the regulatory doghouse, on the defensive for once.

Daily Crunch: Google Meet will get a new look and new features

By Anthony Ha

Google announces upgrades to Google Meet, Amazon is bringing its palm scanner to Whole Foods and Microsoft looks at the effect of video calls on our brains. This is your Daily Crunch for April 21, 2021.

The big story: Google Meet will get a new look and new features

Google Meet is getting a number of updates, including a new user interface that should make the controls more visible (rather than hiding them in menus), the ability to pin multiple video feeds, autozoom (which will automatically place you in the center of the frame) and background replacement, starting with just a few scenes.

It sounds like these changes aren’t happening all at once, but will roll out gradually over the next few months. Google said the goal is to make online meetings “more immersive, inclusive and productive.”

The tech giants

Foxconn’s Wisconsin factory plans scaled back dramatically — The Taiwanese manufacturing giant is scaling back its investment from $10 billion to $672 million.

Amazon is bringing its Amazon One palm scanner to select Whole Foods as a payment option — That means Whole Foods customers could choose to scan their palm over the reader to pay for their purchases.

Instagram launches tools to filter out abusive DMs based on keywords and emojis — It will also allow users to proactively block people, even if they try to make contact from a new account.

Startups, funding and venture capital

Remote hiring startup Deel raises $156M at a $1.25B valuation after 20x growth in 2020 — Deel aims to allow businesses “to hire anyone, anywhere, in a compliant manner.”

Discount grocery startup Misfits Market raises $200M — This round moves the startup known for selling “ugly” fruits and vegetables into unicorn territory.

AppOmni raises $40M for tools to secure enterprise SaaS apps — The startup has built a platform to help monitor SaaS apps and their activity, provide guidance to warn or block when things might go wrong and fix problems when they do occur.

Advice and analysis from Extra Crunch

Four ways martech will shift in 2021 — First and foremost, differentiation is going to be imperative.

Micromobility’s next big business is software, not vehicles — The days of the shared, dockless micromobility model are numbered, at least according to Puneeth Meruva of Trucks Venture Capital.

Dear Sophie: How can I get my startup off the ground and visit the US? — The latest edition of Dear Sophie, the advice column that answers immigration-related questions about working at technology companies.

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

Everything else

This is your brain on Zoom — Microsoft has done a little brain science and found out that yeah, constant video calls do increase your stress and brain noise.

New privacy bill would end law enforcement practice of buying data from brokers — A new bill known as the Fourth Amendment is Not for Sale Act would seal up a loophole that intelligence and law enforcement agencies use to obtain troves of sensitive and identifying information.

The Daily Crunch is TechCrunch’s roundup of our biggest and most important stories. If you’d like to get this delivered to your inbox every day at around 3pm Pacific, you can subscribe here.

Apple and Google pressed in antitrust hearing on whether app stores share data with product development teams

By Sarah Perez

In today’s antitrust hearing in the U.S. Senate, Apple and Google representatives were questioned on whether they have a “strict firewall” or other internal policies in place that prevent them from leveraging the data from third-party businesses operating on their app stores to inform the development of their own competitive products. Apple, in particular, was called out for the practice of copying other apps by Senator Richard Blumenthal (D-CT), who said the practice had become so common that it earned a nickname with Apple’s developer community: “sherlocking.”

Sherlock, which has its own Wikipedia entry under software, comes from Apple’s search tool in the early 2000s called Sherlock. A third-party developer, Karelia Software, created an alternative tool called Watson. Following the success of Karelia’s product, Apple added Watson’s same functionality into its own search tool, and Watson was effectively put out of business. The nickname “Sherlock” later became shorthand for any time Apple copies an idea from a third-party developer that threatens to or even destroys their business.

Over the years, developers claimed Apple has “sherlocked” a number of apps, including Konfabulator (desktop widgets), iPodderX (podcast manager), Sandvox (app for building websites) and Growl (a notification system for Mac OS X) and, in more recent years, F.lux (blue light reduction tool for screens) Duet and Luna (apps that makes iPad a secondary display), as well as various screen-time-management tools. Now Tile claims Apple has also unfairly entered its market with AirTag.

During his questioning, Blumenthal asked Apple and Google’s representatives at the hearing — Kyle Andeer, Apple’s chief compliance officer and Wilson White, Google’s senior director of Public Policy & Government Relations, respectively — if they employed any sort of “firewall” in between their app stores and their business strategy.

Andeer somewhat dodged the question, saying, “Senator, if I understand the question correctly, we have separate teams that manage the App Store and that are engaged in product development strategy here at Apple.”

Blumenthal then clarified what he meant by “firewall.” He explained that it doesn’t mean whether or not there are separate teams in place, but whether there’s an internal prohibition on sharing data between the App Store and the people who run Apple’s other businesses.

Andeer then answered, “Senator, we have controls in place.”

He went on to note that over the past 12 years, Apple has only introduced “a handful of applications and services,” and in every instance, there are “dozens of alternatives” on the App Store. And, sometimes, the alternatives are more popular than Apple’s own product, he noted.

“We don’t copy. We don’t kill. What we do is offer up a new choice and a new innovation,” Andeer stated.

His argument may hold true when there are strong rivalries, like Spotify versus Apple Music, or Netflix versus Apple TV+, or Kindle versus Apple Books. But it’s harder to stretch it to areas where Apple makes smaller enhancements — like when Apple introduced Sidecar, a feature that allowed users to make their iPad a secondary display. Sidecar ended the need for a third-party app, after apps like Duet and Luna first proved the market.

Another example was when Apple built screen-time controls into its iOS software, but didn’t provide the makers of third-party screen-time apps with an API so consumers could use their preferred apps to configure Apple’s Screen Time settings via the third-party’s specialized interface or take advantage of other unique features.

Blumenthal said he interpreted Andeer’s response as to whether Apple has a “data firewall” as a “no.”

Posed the same question, Google’s representative, White, said his understanding was that Google had “data access controls in place that govern how data from our third-party services are used.”

Blumenthal pressed him to clarify if this was a “firewall,” meaning, he clarified again, “do you have a prohibition against access?”

“We have a prohibition against using our third-party services to compete directly with our first-party services,” White said, adding that Google has “internal policies that govern that.”

The senator said he would follow up on this matter with written questions, as his time expired.

New privacy bill would end law enforcement practice of buying data from brokers

By Taylor Hatmaker

A new bill known as the Fourth Amendment is Not for Sale Act would seal up a loophole that intelligence and law enforcement agencies use to obtain troves of sensitive and identifying information to which they wouldn’t otherwise have legal access.

The new legislation, proposed by Senators Ron Wyden (D-OR) and Rand Paul (R-KY), would require government agencies to obtain a court order to access data from brokers. Court orders are already required when the government seeks analogous data from mobile providers and tech platforms.

“There’s no reason information scavenged by data brokers should be treated differently than the same data held by your phone company or email provider,” Wyden said. Wyden describes the loophole as a way that police and other agencies buy data to “end-run the Fourth Amendment.”

Paul criticized the government for using the current data broker loophole to circumvent Americans’ constitutional rights. “The Fourth Amendment’s protection against unreasonable search and seizure ensures that the liberty of every American cannot be violated on the whims, or financial transactions, of every government officer,” Paul said.

Critically, the bill would also ban law enforcement agencies from buying data on Americans when it was obtained through hacking, violations of terms of service or “from a user’s account or device.”

That bit highlights the questionable practices of Clearview AI, a deeply controversial tech company that sells access to a facial recognition search engine. Clearview’s platform collects pictures of faces scraped from across the web, including social media sites, and sells access to that data to police departments around the country and federal agencies like ICE.

In scraping their sites for data to sell, Clearview has run afoul of just about every major social media platform’s terms of service. Facebook, YouTube, Twitter, LinkedIn and Google have all denounced Clearview for using data culled from their services and some have even sent cease-and-desists ordering the data broker to stop.

The bill would also expand privacy laws to apply to infrastructure companies that own cell towers and data cables, seal up workarounds that allow intelligence agencies to obtain metadata from Americans’ international communications without review by a FISA court and ensure that agencies seek probable cause orders to obtain location and web browsing data.

The bill, embedded below, isn’t just some nascent proposal. It’s already attracted bipartisan support from a number of key co-sponsors, including Senate Majority Leader Chuck Schumer and Bernie Sanders on the Democratic side and Republicans Mike Lee and Steve Daines. A House version of the legislation was also introduced Wednesday.

 

Before yesterdayYour RSS feeds

Google misled consumers over location data settings, Australia court finds

By Natasha Lomas

Google’s historical collection of location data has got it into hot water in Australia where a case brought by the country’s Competition and Consumer Commission (ACCC) has led to a federal court ruling that the tech giant misled consumers by operating a confusing dual-layer of location settings in what the regulator describes as a “world-first enforcement action”.

The case relates to personal location data collected by Google through Android mobile devices between January 2017 and December 2018.

Per the ACCC, the court ruled that “when consumers created a new Google Account during the initial set-up process of their Android device, Google misrepresented that the ‘Location History’ setting was the only Google Account setting that affected whether Google collected, kept or used personally identifiable data about their location”.

“In fact, another Google Account setting titled ‘Web & App Activity’ also enabled Google to collect, store and use personally identifiable location data when it was turned on, and that setting was turned on by default,” it wrote.

The Court also ruled that Google misled consumers when they later accessed the ‘Location History’ setting on their Android device during the same time period to turn that setting off because it did not inform them that by leaving the ‘Web & App Activity’ setting switched on, Google would continue to collect, store and use their personally identifiable location data.

“Similarly, between 9 March 2017 and 29 November 2018, when consumers later accessed the ‘Web & App Activity’ setting on their Android device, they were misled because Google did not inform them that the setting was relevant to the collection of personal location data,” the ACCC added.

Similar complaints about Google’s location data processing being deceptive — and allegations that it uses manipulative tactics in order to keep tracking web users’ locations for ad-targeting purposes — have been raised by consumer agencies in Europe for years. And in February 2020 the company’s lead data regulator in the region finally opened an investigation. However that probe remains ongoing.

Whereas the ACCC said today that it will be seeking “declarations, pecuniary penalties, publications orders, and compliance orders” following the federal court ruling. Although it added that the specifics of its enforcement action will be determined “at a later date”. So it’s not clear exactly when Google will be hit with an order — nor how large a fine it might face.

The tech giant may also seek to appeal the court ruling.

Google said today it’s reviewing its legal options and considering a “possible appeal” — highlighting the fact the Court did not agree wholesale with the ACCC’s case because it dismissed some of the allegations (related to certain statements Google made about the methods by which consumers could prevent it from collecting and using their location data, and the purposes for which personal location data was being used by Google).

Here’s Google’s statement in full:

“The court rejected many of the ACCC’s broad claims. We disagree with the remaining findings and are currently reviewing our options, including a possible appeal. We provide robust controls for location data and are always looking to do more — for example we recently introduced auto delete options for Location History, making it even easier to control your data.”

While Mountain View denies doing anything wrong in how it configures location settings — while simultaneously claiming it’s always looking to improve the controls it offers its users — Google’s settings and defaults have, nonetheless, got it into hot water with regulators before.

Back in 2019 France’s data watchdog, the CNIL, fined it $57M over a number of transparency and consent failures under the EU’s General Data Protection Regulation. That remains the largest GDPR penalty issued to a tech giant since the regulation came into force a little under three years ago — although France has more recently sanctioned Google $120M under different EU laws for dropping tracking cookies without consent.

Australia, meanwhile, has forged ahead with passing legislation this year that directly targets the market power of Google (and Facebook) — passing a mandatory news media bargaining code in February which aims to address the power imbalance between platform giants and publishers around the reuse of journalism content.

Daily Crunch: Google Earth gets an update

By Anthony Ha

Google Earth gives users a new look at a changing planet, Facebook tests new business discovery features and Autodesk acquires Upchain. This is your Daily Crunch for April 15, 2021.

The big story: Google Earth gets an update

Google is describing this as Google Earth’s biggest update since 2017, though there’s really just one major addition: A time-lapse mode bringing together satellite photos from the past 37 years, in 3D.

Beyond just being a novel new feature, this mode can reveal how climate change has reshaped our planet. In fact, Google Earth now offers guided tours focused on forest change, urban growth, warming temperatures and more.

The tech giants

Facebook to test new business discovery features in US News Feed — Users will be able to tap on topics they’re interested in underneath posts and ads in their News Feed, allowing them to explore related content from businesses.

Autodesk acquires Upchain — Upchain is a Toronto-based startup that offers a cloud-based product life cycle management service.

Consumer groups and child development experts petition Facebook to drop ‘Instagram for kids’ plan — The letter was written by the Campaign for a Commercial-Free Childhood, an advocacy group that often leads campaigns against big tech and its targeting of children.

Startups, funding and venture capital

Polestar raises $500M from outside investors as EV market grows — This is the first external round for Volvo Car Group’s standalone electric performance brand.

Goldman Sachs leads $23M in funding for Brazilian e-commerce startup Olist — Olist connects small businesses to larger product marketplaces, helping entrepreneurs sell their products to a larger customer base.

Substack announces a $1M initiative to fund local journalists — The newsletter startup will fund independent writers creating local news publications.

Advice and analysis from Extra Crunch

Coinbase’s direct listing alters the landscape for fintech and crypto startups — In Alex Wilhelm’s view, Coinbase reaching a valuation north of $100 billion during its first day of trading was the biggest startup event of the year.

Billion-dollar B2B: cloud-first enterprise tech behemoths have massive potential — Dharmesh Thakker of Battery Ventures writes that a new class of cloud-first, enterprise-tech behemoths have the potential to reach $1 billion in ARR.

How startups can ensure CCPA and GDPR compliance in 2021 — It’s important to enact best data management practices before a legal situation arises.

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

Everything else

Detroit’s native son, billionaire Dan Gilbert, makes the case for his town — The Quicken Loans founder has poured at least $2.5 billion into rehabilitating buildings in the heart of the city.

Can the tech trade show return in 2021? — IFA promises “full-scale, real-life event” for Berlin in September, while MWC reels from the loss of marquee names.

Garry Kasparov launches a community-first chess platform — Kasparovchess will be a platform in which legendary chess players have free reign to share tips and tricks with players from various levels.

The Daily Crunch is TechCrunch’s roundup of our biggest and most important stories. If you’d like to get this delivered to your inbox every day at around 3pm Pacific, you can subscribe here.

You can now pay for BART using an iPhone or Apple Watch

By Greg Kumparak

Good news, Bay Area! Apple Pay now works with Clipper cards.

That means you can now use an iPhone or Apple Watch to pay for BART. Or Muni. Or Caltrain. Or the Ferry! Or (almost) any other transit-related thing you’d otherwise use the plastic Clipper card for.

Clipper has a page outlining the Apple Pay setup process right here.

A few quick but important things to note:

  • Adding an existing Clipper card to an Apple Wallet apparently transfers the funds off that card. At that point, says Clipper, “your plastic card has been deactivated” — so it sounds like it won’t work as a physical backup card.
  • Some people will want to hang on to the plastic cards, regardless: Clipper notes that Bay Area bike share users and anyone using an RTC Discount Card will need to keep the plastic card, even after its deactivated for transit use.
  • Clipper has previously confirmed that support is coming for Google Pay (Android) “this spring”, but today’s rollout seems to support Apple Pay only.

As noted back in February when this was first confirmed as on-the-way, Clipper works with Apple’s “Express Transit” feature. That’s just a fancy way to say that you can tap-to-pay with the digital Clipper card without first needing to punch in your phone’s PIN or using FaceID. On certain newer iPhones, it also lets you keep using the Clipper card for a few hours after your battery has died; a wonderful thing in a pinch, but probably not something you want to rely on regularly.

Hadrian is building the factories of the future for rocket ships and advanced manufacturing

By Jonathan Shieber

If the eight person team behind the new startup Hadrian has their way, they’ll have transformed the manufacturing industry within the next decade.

At least, that’s the goal for the new San Francisco-based startup, founded only last year, which has set its sights on building out a new model for advanced manufacturing to enable the satellite, space ship, and advanced energy technology companies to build the future they envision better and faster.

We view our job as to provide the world’s most efficient space and defense component factory,” said Hadrian founder, Chris Power.

Initially, the company is building factories to make the parts that go on rocket ships, according to Power, but the business has implications for any company that needs bespoke components to make their equipment.

“Let me tell you how bad it is at the moment and what’s going to happen over the next 20 years. Right now everyone in space and defense, [including] SpaceX and Lockheed Martin, outsources their parts and manufacturing to small factories across the country. They’re super expensive, they’re unreliable and they’re completely invisible to the customers,” said Power. “This causes big problems with space and defense manufacturers in the design phase, because the lead time is so long and the iteration time is super long. Imagine running software and being able to iterate on your product once every 20 days? If you can imagine a Gantt chart of how to build a rocket, about 60% of that is buffer time… A lot of the delays in launches and stuff like that happen because parts got delivered three months ago. It’d be like running a McDonalds and realizing that your fries and burger providers could not tell you when the food would arrive.”

It’s hard to overstate the strategic importance of the parts suppliers to the operations of aerospace, defense, and advanced machining companies. As no less an authority on manufacturing than Elon Musk noted in a tweet, “The factory is the product.” It’s also hard to overstate the geopolitical importance of re-establishing the U.S. as a center of manufacturing excellence, according to Hadrian’s investors Lux Capital, Founders Fund, and Construct Capital. Which is one reason why they’re investing $9.5 million into the very early stage business.

“America made massive strategic mistakes in the early 90s which have left our national manufacturing ecosystem completely dilapidated,” said Founders Fund principal Delian Asparouhov. “The only way to get out of this disaster is to re-invent the most basic input into our aerospace and defense supply chains, machining metal parts quickly and with high tolerance. Right now, America’s most innovative company, SpaceX, relies on a network of near-retired machinists to produce space-worthy metal parts, and no one in technology is. focused on solving this.”

 

The factory is the product

— Elon Musk (@elonmusk) January 11, 2021

Power got to understand the problem at his previous company, Ento, which sold workforce management software to blue collar customers. It was there he realized the issue of. the aging workforce and the need for manufacturers to upgrade almost every aspect of their own technology stack. “I realized that the right way to bring technology to the industrial space is not to sell software to these companies, it’s to build an industrial business from scratch with software.”

Initially, Hadrian is focusing all of its efforts on the space industry, where the component manufacturing problem is especially acute, but the manufacturing capabilities the company is building out have broad relevance across any industry that requires highly engineered components.

“The demand for manufacturing from both the large SpaceX and Blue Origin all the way to this growing long tail of companies from Anduril to Relativity to Varda,” said Lux Capital co-founder Josh Wolfe. “Most of these guys are using mom and pop machine shops… [and] those shops are horribly inefficient. They’re not consistent, and they’re not reliable. Between the software automation, the hardware, you can cut down on inefficiency every step of the process… I like to think of value creation as waste reduction… so mundane things like quoting, scheduling, bidding, and planning all the way to the programming of the manufacturing… every one of those things takes hours to tens of hours to days and weeks, so if you can do that in minutes, it’s just a no-brainer. [Hadrian] will be the cutting edge choice for all of the new and explicitly dedicated and focused aerospace and defense companies.”

Power envisions a network of manufacturing facilities that can initially cover roughly 65% of all space and defense components, and will eventually take that number up to 95% of components. Already several of the biggest launch vehicle and satellite manufacturers are in talks with the company to produce hundreds of units for them, Power said. Some of those companies just happen to be in the Construct, Lux, and Founders Fund portfolio.

And the company’s founder sees this as a new way to revitalize American manufacturing jobs as well. “Manufacturing jobs in space and defense can easily be as high paying as a software engineering job at Google,” he said. In an ideal world, Hadrian would like to offer an onramp to high paying manufacturing careers in the 21st century in the same way that automakers provided good union jobs in the twentieth.

“We haven’t built any of this. If you look at the sheer number of people that we need to train and hire on our new technology and new systems, that people problem and that training problem is part of growing our business.”

A render of Axiom’s future commercial space station design.

Google says Google Earth is getting its biggest update since 2017

By Frederic Lardinois

Google Earth now features a timelapse mode that brings together 24 million satellite photos from the last 37 years. And… that’s it. Google says it’s the biggest update to Google Earth — a product you’ve likely forgotten even existed — since its redesign in 2017.

To be fair, Google Earth hasn’t gotten any major new features updates since then. So by default, I guess this qualifies as the biggest update to Earth in a while. It’s worth noting, though, that Google Earth timelapses launched a few years already, but on a dedicated site and only in 2D. Now it’s in 3D. Exciting stuff — for five minutes (or really depressing, if you look at the Earth’s glaciers and rain forests).

Google’s FeedBurner moves to a new infrastructure but loses its email subscription service

By Frederic Lardinois

Google today announced that it is moving FeedBurner to a new infrastructure but also deprecating its email subscription service.

If you’re an internet user of a certain age, chances are you used Google’s FeedBurner to manage the RSS feeds of your personal blogs and early podcasts at some point. During the Web 2.0 era, it was the de facto standard for feed management and analytics, after all. Founded in 2004, with Dick Costolo as one of its co-founders (before he became Twitter’s CEO in 2010), it was acquired by Google in 2007.

Ever since, FeedBurner lingered in an odd kind of limbo. While Google had no qualms shutting down popular services like Google Reader in favor of its ill-fated social experiments like Google+, FeedBurner just kept burning feeds day in and day out, even as Google slowly deprecated some parts of the service, most notably its advertising integrations.

I don’t know that anybody spent a lot of time thinking about the service and RSS has slowly (and sadly) fallen into obscurity, yet the service was probably easy enough to maintain that Google kept it going. And despite everything, shutting it down would probably break enough tools for publishers to create quite an uproar. The TechCrunch RSS feed, to which you are surely subscribed in your desktop RSS reader, is http://feeds.feedburner.com/TechCrunch/, after all.

So here we are, 14 years later, and Google today announced that it is “making several upcoming changes to support the product’s next chapter.” It’s moving the service to a new, more stable infrastructure.

But in July, it is also shutting down some non-core features that don’t directly involve feed management, most importantly the FeedBurner email subscription service that allowed you to get emailed alerts when a feed updates. Feed owners will be able to download their email subscriber lists (and will be able to do so after July, too). With that, Blogger’s FollowByEmail widget will also be deprecated (and hey, did you start this day thinking you’d read about FeedBurner AND Blogger on TechCrunch without having to travel back to 2007?).

Google stresses that other core FeedBurner features will remain in place, but given the popularity of email newsletters, that’s a bit of an odd move.

This startup wants to use technology to help solve Austin’s housing shortage problem

By Mary Ann Azevedo

The sheer volume of people migrating to Austin from all over the country, but particularly from the San Francisco Bay Area, has been making headlines for a while now.

One result of this continued migration is a steady surge in housing prices due to increased demand and low inventory that dropped to nearly zero earlier this year. Now, Homebound, a Santa Rosa, California-based tech-enabled homebuilding startup, is entering the Austin market with the goal of helping ease some of the pain felt in the city by offering an alternative to buying existing homes.

Homebound has raised about $73 million over the years from the likes of Google Ventures, Fifth Wall, Khosla, Sound Ventures, Atomic and Thrive Capital. It raised a $35 million Series B last April and then closed on a $20 million convertible note late last year. CEO Nikki Pechet and Atomic managing partner Jack Abraham founded the company in 2017 after Abraham lost his home to wildfires.

Essentially serving as a virtual general contractor, Homebound combines technology and a network for “vetted” and licensed building “experts” to manage the new home construction from the design phase to completion. The startup has developed tools to track and manage hundreds of unique tasks associated with building a home.

Up until this point, Homebound has been focused on helping homeowners navigate the challenges and complexities of rebuilding after wildfires in California. But this month, Homebound will be expanding to Austin, its first non-disaster market, with the goal of taking learnings from those rebuilds and applying the same “streamlined, tech-enabled building process” to make custom homebuilding an option for local homeowners.

I talked with Homebound’s CEO and co-founder, Nikki Pechet, to learn more.

With Homebound, she said, the company is out to serve as a “next gen” homebuilder to make it possible “for anyone, anywhere to build a home.”

Austin’s housing market is definitely overheated, with homes going 10-30% above asking in some cases (I should know, I live here).

“Homeowners have been reaching out to us from across the country asking us to come to their market,” Pechet said. “We’re already seeing Austin grow faster than any of our other markets did in their early days. It’s going to be a huge market for us.”

It’s a model Pechet envisions replicating in other cities with similar housing supply issues such as Miami, Tampa, Raleigh and Charlotte.

“This is just the start,” Pechet said. “We’re taking the platform to markets across the country to help exactly with this issue.”

The company starts by helping a potential homeowner identify land they want to build on, or help them find a lot among the inventory Homebound has already built up. From there, it can help with everything from architectural plans to design to actual construction via its platform. Homebound offers a set of plans for people to choose from, with varying levels of customization.

Building costs for a typical single-family home in the Austin area will start around $300,000 depending on the size, complexity of house, lot size and location. That does not include land cost. Some people are opting to build second units on existing properties.

“In most cases, people can build a new home for less than they can pay for an existing home just because of the dynamics,” Pechet said.

JXL turns Jira into spreadsheets

By Frederic Lardinois

Atlassian’s Jira is an extremely powerful issue tracking and project management tool, but it’s not the world’s most intuitive piece of software. Spreadsheets, on the other hand, are pretty much the de facto standard for managing virtually anything in a business. It’s maybe no surprise then that there are already a couple of tools on the market that bring a spreadsheet-like view of your projects to Jira or connect it to services like Google Sheets.

The latest entrant in this field is JXL Spreadsheets for Jira (and specifically Jira Cloud), which was founded by two ex-Atlassian employees, Daniel Franz and Hannes Obweger. And in what has become a bit of a trend, Atlassian Ventures invested in JXL earlier this year.

Franz built the Good News news reader before joining Atlassian, while his co-founder previously founded Radiant Minds Software, the makers of Jira Roadmaps (now Portfolio for Jira), which was acquired by Atlassian.

Image Credits: JXL

“Jira is so successful because it is awesome,” Franz told me. “It is so versatile. It’s highly customizable. I’ve seen people in my time who are doing anything and everything with it. Working with customers [at Atlassian] — at some point, you didn’t get surprised anymore, but what the people can do and track with JIRA is amazing. But no one would rock up and say, ‘hey, JIRA is very pleasant and easy to use.’ ”

As Franz noted, by default, Jira takes a very opinionated view of how people should use it. But that also means that users often end up exporting their issues to create reports and visualizations, for example. But if they make any changes to this data, it never flows back into Jira. No matter how you feel about spreadsheets, they do work for many people and are highly flexible. Even Atlassian would likely agree because the new Jira Work Management, which is currently in beta, comes with a spreadsheet-like view and Trello, too, recently went this way when it launched a major update earlier this year.

Image Credits: JXL

Over the course of its three-month beta, the JXL team saw how its users ended up building everything from cross-project portfolio management to sprint planning, backlog maintenance, timesheets and inventory management on top of its service. Indeed, Franz tells me that the team already has some large customers, with one of them having a 7,000-seat license.

Pricing for JXL seems quite reasonable, starting at $1/user/month for teams with up to 10 users. Larger teams get increasingly larger discounts, down to $0.45/user/month for licenses with over 5,000 seats. There is also a free trial.

One of the reasons the company can offer this kind of pricing is because it only needs a very simple backend. None of a customer’s data sits on JXL’s servers. Instead, it sits right on top of Jira’s APIs, which in turn also means that changes are synced back and forth in real time.

JXL is now available in the Atlassian Marketplace and the team is actively hiring as it looks to build out its product (and put its new funding to work).

SambaNova raises $676M at a $5.1B valuation to double down on cloud-based AI software for enterprises

By Ingrid Lunden

Artificial intelligence technology holds a huge amount of promise for enterprises — as a tool to process and understand their data more efficiently; as a way to leapfrog into new kinds of services and products; and as a critical stepping stone into whatever the future might hold for their businesses. But the problem for many enterprises is that they are not tech businesses at their cores and so bringing on and using AI will typically involve a lot of heavy lifting. Today, one of the startups building AI services is announcing a big round of funding to help bridge that gap.

SambaNova — a startup building AI hardware and integrated systems that run on it that only officially came out of three years in stealth last December — is announcing a huge round of funding today to take its business out into the world. The company has closed in on $676 million in financing, a Series D that co-founder and CEO Rodrigo Liang has confirmed values the company at $5.1 billion.

The round is being led by SoftBank, which is making the investment via Vision Fund 2. Temasek and the Government of Singapore Investment Corp. (GIC), both new investors, are also participating, along with previous backers BlackRock, Intel Capital, GV (formerly Google Ventures), Walden International and WRVI, among other unnamed investors. (Sidenote: BlackRock and Temasek separately kicked off an investment partnership yesterday, although it’s not clear if this falls into that remit.)

Co-founded by two Stanford professors, Kunle Olukotun and Chris Ré, and Liang, who had been an engineering executive at Oracle, SambaNova has been around since 2017 and has raised more than $1 billion to date — both to build out its AI-focused hardware, which it calls DataScale and to build out the system that runs on it. (The “Samba” in the name is a reference to Liang’s Brazilian heritage, he said, but also the Latino music and dance that speaks of constant movement and shifting, not unlike the journey AI data regularly needs to take that makes it too complicated and too intensive to run on more traditional systems.)

SambaNova on one level competes for enterprise business against companies like Nvidia, Cerebras Systems and Graphcore — another startup in the space which earlier this year also raised a significant round. However, SambaNova has also taken a slightly different approach to the AI challenge.

In December, the startup launched Dataflow-as-a-service as an on-demand, subscription-based way for enterprises to tap into SambaNova’s AI system, with the focus just on the applications that run on it, without needing to focus on maintaining those systems themselves. It’s the latter that SambaNova will be focusing on selling and delivering with this latest tranche of funding, Liang said.

SambaNova’s opportunity, Liang believes, lies in selling software-based AI systems to enterprises that are keen to adopt more AI into their business, but might lack the talent and other resources to do so if it requires running and maintaining large systems.

“The market right now has a lot of interest in AI. They are finding they have to transition to this way of competing, and it’s no longer acceptable not to be considering it,” said Liang in an interview.

The problem, he said, is that most AI companies “want to talk chips,” yet many would-be customers will lack the teams and appetite to essentially become technology companies to run those services. “Rather than you coming in and thinking about how to hire scientists and hire and then deploy an AI service, you can now subscribe, and bring in that technology overnight. We’re very proud that our technology is pushing the envelope on cases in the industry.”

To be clear, a company will still need data scientists, just not the same number, and specifically not the same number dedicating their time to maintaining systems, updating code and other more incremental work that comes managing an end-to-end process.

SambaNova has not disclosed many customers so far in the work that it has done — the two reference names it provided to me are both research labs, the Argonne National Laboratory and the Lawrence Livermore National Laboratory — but Liang noted some typical use cases.

One was in imaging, such as in the healthcare industry, where the company’s technology is being used to help train systems based on high-resolution imagery, along with other healthcare-related work. The coincidentally-named Corona supercomputer at the Livermore Lab (it was named after the 2014 lunar eclipse, not the dark cloud of a pandemic that we’re currently living through) is using SambaNova’s technology to help run calculations related to some Covid-19 therapeutic and antiviral compound research, Marshall Choy, the company’s VP of product, told me.

Another set of applications involves building systems around custom language models, for example in specific industries like finance, to process data quicker. And a third is in recommendation algorithms, something that appears in most digital services and frankly could always do to work a little better than it does today. I’m guessing that in the coming months it will release more information about where and who is using its technology.

Liang also would not comment on whether Google and Intel were specifically tapping SambaNova as a partner in their own AI services, but he didn’t rule out the prospect of partnering to go to market. Indeed, both have strong enterprise businesses that span well beyond technology companies, and so working with a third party that is helping to make even their own AI cores more accessible could be an interesting prospect, and SambaNova’s DataScale (and the Dataflow-as-a-service system) both work using input from frameworks like PyTorch and TensorFlow, so there is a level of integration already there.

“We’re quite comfortable in collaborating with others in this space,” Liang said. “We think the market will be large and will start segmenting. The opportunity for us is in being able to take hold of some of the hardest problems in a much simpler way on their behalf. That is a very valuable proposition.”

The promise of creating a more accessible AI for businesses is one that has eluded quite a few companies to date, so the prospect of finally cracking that nut is one that appeals to investors.

“SambaNova has created a leading systems architecture that is flexible, efficient and scalable. This provides a holistic software and hardware solution for customers and alleviates the additional complexity driven by single technology component solutions,” said Deep Nishar, Senior Managing Partner at SoftBank Investment Advisers, in a statement. “We are excited to partner with Rodrigo and the SambaNova team to support their mission of bringing advanced AI solutions to organizations globally.”

Apple and Google will both attend Senate hearing on app store competition

By Taylor Hatmaker

After it looked like Apple might no-show, the company has committed to sending a representative to a Senate antitrust hearing on app store competition later this month.

Last week, Senators Amy Klobuchar (D-MN) and Mike Lee (R-UT) put public pressure on the company to attend the hearing, which will be held by the Senate Judiciary Subcommittee on Competition Policy, Antitrust, and Consumer Rights. Klobuchar chairs that subcommittee, and has turned her focus toward antitrust worries about the tech industry’s most dominant players.

The hearing, which Google will also attend, will delve into Apple and Google’s control over “the cost, distribution, and availability of mobile applications on consumers, app developers, and competition.”

App stores are one corner of tech that looks the most like a duopoly, a perception that Apple’s high profile battle with Fortnite-maker Epic is only elevating. Meanwhile, with a number of state-level tech regulation efforts brewing, Arizona is looking to relieve developers from Apple and Google’s hefty cut of app store profits.

In a letter last week, Klobuchar and Lee, the subcommittee’s ranking member, accused Apple of “abruptly” deciding that it wouldn’t send a witness to the hearing, which is set for April 21.

“Apple’s sudden change in course to refuse to provide a witness to testify before the Subcommittee on app store competition issues in April, when the company is clearly willing to discuss them in other public forums, is unacceptable,” the lawmakers wrote.

By Monday, that pressure had apparently done its work, with Apple agreeing to attend the hearing. Apple didn’t respond to a request for comment.

While the lawmakers are counting Apple’s acquiescence as a win, that doesn’t mean they’ll be sending their chief executive. Major tech CEOs have been called before Congress more often over the last few years, but those appearances might have diminishing returns.

Tech CEOs, Apple’s Tim Cook included, are thoroughly trained in the art of saying little when pressed by lawmakers. Dragging in a CEO might work as a show of force, but tech execs generally reveal little over the course of their lengthy testimonies, particularly when a hearing isn’t accompanied by a deeper investigation.

A ‘more honest’ stock market

By Lucas Matney

Hello friends, and welcome back to Week in Review!

Last week, I talked about Clubhouse’s slowing user growth. Well, this week news broke that they had been in talks with Twitter for a $4 billion acquisition, so it looks like they’re still pretty desirable. This week, I’m talking about a story I published a couple days ago that highlights pretty much everything that’s wild about the alternative asset world right now.

If you’re reading this on the TechCrunch site, you can get this in your inbox from the newsletter page, and follow my tweets @lucasmtny.


The big thing

If you successfully avoided all mentions of NFTs until now, I congratulate you, because it certainly does seem like the broader NFT market is seeing some major pullback after a very frothy February and March. You’ll still be seeing plenty of late-to-the-game C-list celebrities debuting NFT art in the coming weeks, but a more sober pullback in prices will probably give some of the NFT platforms that are serious about longevity a better chance to focus on the future and find out how they truly matter.

I spent the last couple weeks, chatting with a bunch of people in one particular community — one of the oldest active NFT communities on the web called CryptoPunks. It’s a platform with 10,000 unique 24×24 pixel portraits and they trade at truly wild prices.

I wrote about the history and legacy of CryptoPunks, a vibrant $200 million NFT marketplace built around trading pixelated characters. There are only 10,000 of them and owning the cheapest one will cost you about $30k. https://t.co/X4iTSl6FjC

— Lucas Matney (@lucasmtny) April 8, 2021

This picture sold for a $1.05 million.

I talked to a dozen or so people (including the guy who sold that one ^^) that had spent between tens of thousands and millions of dollars on these pixelated portraits, my goal being to tap into the psyche of what the hell is happening here. The takeaway is that these folks don’t see these assets as any more non-sensical than what’s going on in more traditional “old world” markets like public stock exchanges.

A telling quote from my reporting:

“Obviously this is a very speculative market… but it’s almost more honest than the stock market,” user Max Orgeldinger tells TechCrunch. “Kudos to Elon Musk — and I’m a big Tesla fan — but there are no fundamentals that support that stock price. It’s the same when you look at GameStop. With the whole NFT community, it’s almost more honest because nobody’s getting tricked into thinking there’s some very complicated math that no one can figure out. This is just people making up prices and if you want to pay it, that’s the price and if you don’t want to pay it, that’s not the price.”

Shortly after I published my piece, Christie’s announced that they were auctioning off nine of the CryptoPunks in an auction likely to fetch at least $10 million at current prices. The market surged in the aftermath and many millions worth of volume quickly moved through the marketplace minting more NFT millionaires.

Is this all just absolutely nuts? Sure.

Is it also a poignant picture of where alternative asset investing is at in 2021? You bet.

Read the full thing.


an illustration of a cardboard ballot box with an Amazon smile on the front

Other things

Here are the TechCrunch news stories that especially caught my eye this week:

Amazon workers vote down union organization attempt
Amazon is breathing a sigh of relief after workers at their Bessemer, Alabama warehouse opted out of joining a union, lending a crushing defeat to labor activists who hoped that the high-profile moment would lead more Amazon workers to organize. The vote has been challenged, but the margin of victory seems fairly decisive.

Supreme court sides with Google in Oracle case
If any singular event impacted the web the most this week, it was the Supreme Court siding with Google in a very controversial lawsuit by Oracle that could’ve fundamentally shifted the future of software development.

Coinbase is making waves
The Coinbase direct listing is just around the corner and they’re showing off some of their financials. Turns out crypto has been kind of hot lately and they’re raking in the dough, with revenue of $1.8 billion this past quarter.

Apple share more about the future of user tracking
Apple is about to upend the ad-tracking market and they published some more details on what exactly their App Tracking Transparency feature is going to look like. Hint: more user control.

Consumers are spending lots of time in apps
A new report from mobile analytics firm App Annie suggests that we’re dumping more of our time into smartphone apps, with the average users spending 4.2 hours a day doing so, a 30 percent increase over two years.

Sonos perfects the bluetooth speaker
I’m a bit of an audio lover, which made my colleague Darrell’s review of the new Sonos Roam bluetooth speaker a must-read for me. He’s pretty psyched about it, even though it comes in at the higher-end of pricing for these devices, still I’m looking forward to hearing one with my own ears.


 

Image Credits: Nigel Sussman

Extra things

Some of my favorite reads from our Extra Crunch subscription service this week:
The StockX EC-1
“StockX is a unique company at the nexus of two radical transitions that isn’t just redefining markets, but our culture as well. E-commerce upended markets, diminishing the physical experience by intermediating and aggregating buyers and sellers through digital platforms. At the same time, the internet created rapid new communication channels, allowing euphoria and desire to ricochet across society in a matter of seconds. In a world of plenty, some things are rare, and the hype around that rarity has never been greater. Together, these two trends demanded a stock market of hype, an opportunity that StockX has aggressively pursued.”

Building the right team for a billion-dollar startup
“I would really encourage you to take some time to think about what kind of company you want to make first before you go out and start interviewing people. So that really is going to be about understanding and defining your culture. And then the second thing I’d be thinking about when you’re scaling from, you know, five people up to, you know, 50 and beyond is that managers really are the key to your success as a company. It’s hard to overstate how important managers, great managers, are to the success of your company.

So you want to raise a Series A
“More companies will raise seed rounds than Series A rounds, simply due to the fact that many startups fail, and venture only makes sense for a small fraction of businesses out there. Every check is a new cycle of convincing and proving that you, as a startup, will have venture-scale returns. Moore explained that startups looking to move to their next round need to explain to investors why now is their moment.”

Until next week,
Lucas M.

And again, if you’re reading this on the TechCrunch site, you can get this in your inbox from the newsletter page, and follow my tweets @lucasmtny.

Google denies Pixel 5a 5G cancelation, confirming it’s coming this year

By Brian Heater

Sometimes you’ve just got to confirm an unannounced product to put the rumors to bed, I guess. That was Google’s strategy this afternoon, following earlier rumors from Android Central that a chip shortage had put the kibosh on the mid-budget phone.

In a comment to TechCrunch, a Google spokesperson noted, “Pixel 5a 5G is not cancelled. It will be available later this year in the U.S. and Japan and announced in line with when last year’s a-series phone was introduced.”

That time frame would put the device’s arrival around late-summer, meaning it won’t arrive in time for Google I/O in May, as some speculated. Interestingly, the company appears to be limiting the device’s availability to two countries — at least at launch. That could, perhaps, be due to earlier-reported component shortages.

As The Verge notes, the company hasn’t been particularly precious when it comes to product announcements. The company took a similar approach ahead of the release of the Pixel. Either way, this isn’t exactly the standard big company approach to rumor denial, which is to either not answer or otherwise deflect.

Google may well be on edge about its Pixel line these days. The phone line hasn’t exactly taken the mobile world be storm, resulting in longstanding rumors that the company is looking to shake things up. That, in part, has seemingly been confirmed by some fairly high-profile exits.

Still, even while there have been issues on the premium side, the company’s budget “a” line has helped buoy its overall numbers. No word yet on specific specs, but the handset is not expected to be a radical departure from its predecessor.

 

APKPure app contained malicious adware, say researchers

By Zack Whittaker

Security researchers say APKPure, a widely popular app for installing older or discontinued Android apps from outside of Google’s app store, contained malicious adware that flooded the victim’s device with unwanted ads.

Kaspersky Lab said that it alerted APKPure on Thursday that its most recent app version, 3.17.18, contained malicious code that siphoned off data from a victim’s device without their knowledge, and pushed ads to the device’s lock screen and in the background to generate fraudulent revenue for the adware operators.

But the researchers said that the malicious code had the capacity to download other malware, potentially putting affected victims at further risk.

The researchers said the APKPure developers likely introduced the malicious code, known as a software development kit or SDK, from an unverified source. APKPure removed the malicious code and pushed out a new version, 3.17.19, and the developers no longer list the malicious version on its site.

APKPure was set up in 2014 to allow Android users access to a vast bank of Android apps and games, including old versions, as well as app versions from other regions that are no longer on Android’s official app store Google Play. It later launched an Android app, which also has to be installed outside Google Play, serving as its own app store to allow users to download older apps directly to their Android devices.

APKPure is ranked as one of the most popular sites on the internet.

But security experts have long warned against installing apps outside of the official app stores as quality and security vary wildly as much of the Android malware requires victims to install malicious apps from outside the app store. Google scans all Android apps that make it into Google Play, but some have slipped through the cracks before.

TechCrunch contacted APKPure for comment but did not hear back.

NLPCloud.io helps devs add language processing smarts to their apps

By Natasha Lomas

While visual ‘no code‘ tools are helping businesses get more out of computing without the need for armies of in-house techies to configure software on behalf of other staff, access to the most powerful tech tools — at the ‘deep tech’ AI coal face — still requires some expert help (and/or costly in-house expertise).

This is where bootstrapping French startup, NLPCloud.io, is plying a trade in MLOps/AIOps — or ‘compute platform as a service’ (being as it runs the queries on its own servers) — with a focus on natural language processing (NLP), as its name suggests.

Developments in artificial intelligence have, in recent years, led to impressive advances in the field of NLP — a technology that can help businesses scale their capacity to intelligently grapple with all sorts of communications by automating tasks like Named Entity Recognition, sentiment-analysis, text classification, summarization, question answering, and Part-Of-Speech tagging, freeing up (human) staff to focus on more complex/nuanced work. (Although it’s worth emphasizing that the bulk of NLP research has focused on the English language — meaning that’s where this tech is most mature; so associated AI advances are not universally distributed.)

Production ready (pre-trained) NLP models for English are readily available ‘out of the box’. There are also dedicated open source frameworks offering help with training models. But businesses wanting to tap into NLP still need to have the DevOps resource and chops to implement NLP models.

NLPCloud.io is catering to businesses that don’t feel up to the implementation challenge themselves — offering “production-ready NLP API” with the promise of “no DevOps required”.

Its API is based on Hugging Face and spaCy open-source models. Customers can either choose to use ready-to-use pre-trained models (it selects the “best” open source models; it does not build its own); or they can upload custom models developed internally by their own data scientists — which it says is a point of differentiation vs SaaS services such as Google Natural Language (which uses Google’s ML models) or Amazon Comprehend and Monkey Learn.

NLPCloud.io says it wants to democratize NLP by helping developers and data scientists deliver these projects “in no time and at a fair price”. (It has a tiered pricing model based on requests per minute, which starts at $39pm and ranges up to $1,199pm, at the enterprise end, for one custom model running on a GPU. It does also offer a free tier so users can test models at low request velocity without incurring a charge.)

“The idea came from the fact that, as a software engineer, I saw many AI projects fail because of the deployment to production phase,” says sole founder and CTO Julien Salinas. “Companies often focus on building accurate and fast AI models but today more and more excellent open-source models are available and are doing an excellent job… so the toughest challenge now is being able to efficiently use these models in production. It takes AI skills, DevOps skills, programming skill… which is why it’s a challenge for so many companies, and which is why I decided to launch NLPCloud.io.”

The platform launched in January 2021 and now has around 500 users, including 30 who are paying for the service. While the startup, which is based in Grenoble, in the French Alps, is a team of three for now, plus a couple of independent contractors. (Salinas says he plans to hire five people by the end of the year.)

“Most of our users are tech startups but we also start having a couple of bigger companies,” he tells TechCrunch. “The biggest demand I’m seeing is both from software engineers and data scientists. Sometimes it’s from teams who have data science skills but don’t have DevOps skills (or don’t want to spend time on this). Sometimes it’s from tech teams who want to leverage NLP out-of-the-box without hiring a whole data science team.”

“We have very diverse customers, from solo startup founders to bigger companies like BBVA, Mintel, Senuto… in all sorts of sectors (banking, public relations, market research),” he adds.

Use cases of its customers include lead generation from unstructured text (such as web pages), via named entities extraction; and sorting support tickets based on urgency by conducting sentiment analysis.

Content marketers are also using its platform for headline generation (via summarization). While text classification capabilities are being used for economic intelligence and financial data extraction, per Salinas.

He says his own experience as a CTO and software engineer working on NLP projects at a number of tech companies led him to spot an opportunity in the challenge of AI implementation.

“I realized that it was quite easy to build acceptable NLP models thanks to great open-source frameworks like spaCy and Hugging Face Transformers but then I found it quite hard to use these models in production,” he explains. “It takes programming skills in order to develop an API, strong DevOps skills in order to build a robust and fast infrastructure to serve NLP models (AI models in general consume a lot of resources), and also data science skills of course.

“I tried to look for ready-to-use cloud solutions in order to save weeks of work but I couldn’t find anything satisfactory. My intuition was that such a platform would help tech teams save a lot of time, sometimes months of work for the teams who don’t have strong DevOps profiles.”

“NLP has been around for decades but until recently it took whole teams of data scientists to build acceptable NLP models. For a couple of years, we’ve made amazing progress in terms of accuracy and speed of the NLP models. More and more experts who have been working in the NLP field for decades agree that NLP is becoming a ‘commodity’,” he goes on. “Frameworks like spaCy make it extremely simple for developers to leverage NLP models without having advanced data science knowledge. And Hugging Face’s open-source repository for NLP models is also a great step in this direction.

“But having these models run in production is still hard, and maybe even harder than before as these brand new models are very demanding in terms of resources.”

The models NLPCloud.io offers are picked for performance — where “best” means it has “the best compromise between accuracy and speed”. Salinas also says they are paying mind to context, given NLP can be used for diverse user cases — hence proposing number of models so as to be able to adapt to a given use.

“Initially we started with models dedicated to entities extraction only but most of our first customers also asked for other use cases too, so we started adding other models,” he notes, adding that they will continue to add more models from the two chosen frameworks — “in order to cover more use cases, and more languages”.

SpaCy and Hugging Face, meanwhile, were chosen to be the source for the models offered via its API based on their track record as companies, the NLP libraries they offer and their focus on production-ready framework — with the combination allowing NLPCloud.io to offer a selection of models that are fast and accurate, working within the bounds of respective trade-offs, according to Salinas.

“SpaCy is developed by a solid company in Germany called Explosion.ai. This library has become one of the most used NLP libraries among companies who want to leverage NLP in production ‘for real’ (as opposed to academic research only). The reason is that it is very fast, has great accuracy in most scenarios, and is an opinionated” framework which makes it very simple to use by non-data scientists (the tradeoff is that it gives less customization possibilities),” he says.

Hugging Face is an even more solid company that recently raised $40M for a good reason: They created a disruptive NLP library called ‘transformers’ that improves a lot the accuracy of NLP models (the tradeoff is that it is very resource intensive though). It gives the opportunity to cover more use cases like sentiment analysis, classification, summarization… In addition to that, they created an open-source repository where it is easy to select the best model you need for your use case.”

While AI is advancing at a clip within certain tracks — such as NLP for English — there are still caveats and potential pitfalls attached to automating language processing and analysis, with the risk of getting stuff wrong or worse. AI models trained on human-generated data have, for example, been shown reflecting embedded biases and prejudices of the people who produced the underlying data.

Salinas agrees NLP can sometimes face “concerning bias issues”, such as racism and misogyny. But he expresses confidence in the models they’ve selected.

“Most of the time it seems [bias in NLP] is due to the underlying data used to trained the models. It shows we should be more careful about the origin of this data,” he says. “In my opinion the best solution in order to mitigate this is that the community of NLP users should actively report something inappropriate when using a specific model so that this model can be paused and fixed.”

“Even if we doubt that such a bias exists in the models we’re proposing, we do encourage our users to report such problems to us so we can take measures,” he adds.

 

Claiming a landmark in fusion energy, TAE Technologies sees commercialization by 2030

By Jonathan Shieber

In a small industrial park located nearly halfway between Los Angeles and San Diego, one company is claiming to have hit a milestone in the development of a new technology for generating power from nuclear fusion.

The 20-year-old fusion energy technology developer TAE Technologies said its reactors could be operating at commercial scale by the end of the decade, thanks to its newfound ability to produce stable plasma at temperatures over 50 million degrees (nearly twice as hot as the sun).

The promise of fusion energy, a near limitless energy source with few emissions and no carbon footprint, has been 10 years out for the nearly 70 years since humanity first harnessed the power of nuclear energy.  But a slew of companies — including TAE, General Fusion, Commonwealth Fusion Systems and a host of others across North America and around the world — are making rapid advancements that look to bring the technology from the realm of science fiction into the real world.

For TAE Technologies, the achievement serves as a validation of the life’s work of Norman Rostoker, one of the company’s co-founders who had devoted his life to fusion energy research and died before he could see the company he helped create reach its latest milestone.

“This is an incredibly rewarding milestone and an apt tribute to the vision of my late mentor, Norman Rostoker,” said TAE’s current chief executive officer, Michl Binderbauer, in a statement announcing the company’s achievement. “Norman and I wrote a paper in the 1990s theorizing that a certain plasma dominated by highly energetic particles should become increasingly better confined and stable as temperatures increase. We have now been able to demonstrate this plasma behavior with overwhelming evidence. It is a powerful validation of our work over the last three decades, and a very critical milestone for TAE that proves the laws of physics are on our side.”

Rostoker’s legacy lives on inside TAE through the company’s technology platform, called, appropriately, “Norman.” In the last 18 months that technology has demonstrated consistent performance, reaching over 50 million degrees in several hundred test cycles.

Six years ago, the company had proved that its reactor design could sustain plasma indefinitely — meaning that once the switch is flipped on a reaction, that fusion reaction can continue indefinitely. Now, the company said, it has achieved the necessary temperatures to make its reactors commercially viable.

It’s with these milestones behind it that TAE was able to raise an additional $280 million in financing, bringing its total up to $880 million and making it one of the best financed private nuclear fusion endeavors in the world.

“The Norman milestone gives us a high degree of confidence that our unique approach brings fusion within grasp technologically and, more important, economically,” Binderbauer said. “As we shift out of the scientific validation phase into engineering commercial-scale solutions for both our fusion and power management technologies, TAE will become a significant contributor in modernizing the entire energy grid.”

The company isn’t generating energy yet, and won’t for the foreseeable future. The next goal for the company, according to Binderbauer, is to develop the technology to the point where it can create the conditions necessary for making energy from a fusion reaction.

“The energy is super tiny. It’s immaterial. It’s a needle in the haystack,” Binderbauer said. “In terms of its energy discernability, we can use it for diagnostics.”

TAE Technologies Michl Binderbauer standing next to the company’s novel fusion reactor. Image Credits: TAE Technologies

Follow the sun

It took $150 million and five iterations for TAE Technologies to get to Norman, its national laboratory scale fusion device. The company said it conducted over 25,000 fully integrated fusion reactor core experiments, optimized using machine learning programs developed in collaboration with Google and processing power from the Department of Energy’s INCITE program, which leverages exascale-level computing, TAE Technologies said.

The new machine was first fired up in the summer of 2017. Before it could even be constructed TAE Technologies went through a decade of experimentation to even begin approaching the construction of a physical prototype. By 2008, the first construction began on integrated experiments to make a plasma core and infuse it with some energetic particles. The feeder technology and beams alone cost $100 million, Binderbauer said. Then the company needed to develop other technologies like vacuum conditioning. Power control mechanisms also needed to be put in place to ensure that the company’s 3 megawatt power supply could be stored in enough containment systems to power a 750 megawatt energy reaction.

Finally, machine learning capabilities needed to be tapped from companies like Google and compute power from the Department of Energy had to be harnessed to manage computations that could take what had been the theorems that defined Rostoker’s life’s work, and prove that they could be made real.

“By the time Norman became an operating machine we had four generations of devices preceding it. Out of those there were two fully integrated ones and two generations of incremental machines that could do some of it but not all of it.”

Fusion energy’s burning problems

While fusion has a lot of promise as a zero-carbon source of energy, it’s not without some serious limitations, as Daniel Jassby, the former principal physicist at the Princeton Plasma Physics Lab noted in a 2017 Bulletin of the Atomic Scientists article.

Jassby wrote:

Earth-bound fusion reactors that burn neutron-rich isotopes have byproducts that are anything but harmless: Energetic neutron streams comprise 80 percent of the fusion energy output of deuterium-tritium reactions and 35 percent of deuterium-deuterium reactions.

Now, an energy source consisting of 80 percent energetic neutron streams may be the perfect neutron source, but it’s truly bizarre that it would ever be hailed as the ideal electrical energy source. In fact, these neutron streams lead directly to four regrettable problems with nuclear energy: radiation damage to structures; radioactive waste; the need for biological shielding; and the potential for the production of weapons-grade plutonium 239—thus adding to the threat of nuclear weapons proliferation, not lessening it, as fusion proponents would have it.

In addition, if fusion reactors are indeed feasible—as assumed here—they would share some of the other serious problems that plague fission reactors, including tritium release, daunting coolant demands, and high operating costs. There will also be additional drawbacks that are unique to fusion devices: the use of a fuel (tritium) that is not found in nature and must be replenished by the reactor itself; and unavoidable on-site power drains that drastically reduce the electric power available for sale.

TAE Technologies is aware of the problems, according to a spokesperson for the firm, and the company has noted the issues Jassby raised in its product development, the spokesperson said.

“All the callouts to tritium is exactly why TAE has been focused on pB-11 as its feedstock from the very beginning (early 90’s). TAE will reach D-T conditions as a natural stepping stone to pB-11, cause it cooks at ‘only’ 100M c, whereas pB-11 is upwards of 1M c,” the spokesperson wrote in a response. “It would seem like a much harder accomplishment to then scale to 1M, but what this milestone proves is the ‘Scaling law’ for the kind of fusion TAE is generating — in an FRC (the linear design of “Norman”, unlike the donut Tokamaks) the hotter the plasma, the more stable it becomes. It’s the opposite of a [Tokamak].  The milestone gives them scientific confidence they can increase temps beyond DT to pB11 and realize fusion with boron — cheap, aneutronic, abundant — the ideal terrestrial feedstock (let’s not get into mining the moon for helium-3!).”

As for power concerns, the TAE fusion reactor can convert a 2MW grid feed into 750MW shots on the machine without taking down Orange County’s grid (and needing to prove it to SCE), and scale power demand in microseconds to mold and course-correct plasmas in real-time, the spokesperson wrote.

In fact, TAE is going to spin off its power management technology into a separate business focused on peak shaving, energy storage and battery management on the grid and in electric vehicles.

A “safer” fusion technology?

The Hydrogen-Boron, or p-B11, fuel cycle is, according to the company, the most abundant fuel source on earth, and will be the ultimate feedstock for TAE Technologies’ reactor, according to the company. But initially, TAE, like most of the other companies currently developing fusion technologies, will be working with Deuterium-Tritium as its fuel source.

The demonstration facility “Copernicus,” which will be built using some of the new capital the company has announced raising, will start off on the DT fuel cycle and eventually make the switch. Over time, TAE hopes to license the DT technology while building up to its ultimate goal.

Funding the company’s “money by milestone” approach are some of the world’s wealthiest families, firms and companies. Vulcan, Venrock, NEA, Wellcome Trust, Google and the Kuwait Investment Authority are all backers. So too, are the family offices of Addison Fischer, Art Samberg and Charles Schwab.

“TAE is providing the miracles the 21st century needs,” said Addison Fischer, TAE board director and longtime investor who has been involved with conservation and environmental issues for decades. Fischer also founded VeriSign and is a pioneer in defining and implementing security technology underlying modern electronic commerce. “TAE’s most recent funding positions the company to undertake their penultimate step in implementing sustainable aneutronic nuclear fusion and power management solutions that will benefit the planet.”

Formation raises $4M led by Andreessen Horowitz to train truly ‘exceptional’ software engineers

By Mary Ann Azevedo

Sophie Zhou Novati worked as a senior engineer at Facebook and then Nextdoor, where she struggled to hire great engineers for her team.

Frustrated, she decided to try training engineers to meet her team’s hiring standards by mentoring at a local coding bootcamp. After two and a half years of mentoring on nights and weekends, Novati decided to turn her passion into a career.

She and her husband, Michael, founded Formation with a couple of goals in mind. For one, they wanted to offer personalized training to help people not just learn to code, but to become “exceptional” software engineers. Sophie was also struck by the diversity of the people she witnessed going through coding bootcamps, but she realized that those graduates weren’t getting access to the same opportunities that students from traditional universities do.

Formation co-founder and CEO Sophia Zhou Navati

Formation co-founder and CEO Sophia Zhou Navati

With Formation, her goal is to personalize the training experience via a remote fellowship program that combines automated instruction with access to a “network of top tier mentors” from companies such as Facebook and Google. After one year in beta, Formation is unveiling its Engineering Fellowship, where every fellow gets a “personalized training plan tailored to their unique career ambitions.” So far, it’s placed just over 30 people in engineering roles at companies such as Facebook, Microsoft and Lyft with an average starting salary of $120,000.

Formation aims to offer an experience beyond bootcamps, which Sophie argues “have gotten too big, too fast, churning hundreds or thousands of students through fixed curriculums without individualized attention.”

The startup attracted the attention of Andreessen Horowitz, which just led its $4 million seed round. Designer Fund, Combine, Lachy Groom, Slow Ventures and engineers from Airbnb, Notion, Rippling and others also participated in the financing.

“The first thing that really struck me about this community is just how diverse it is. Forty-four percent of graduates are reporting that they identify as nonmale, and the percentage of Black and Latinx graduates is nearly double the national average at traditional universities,” Sophie told TechCrunch. “But the problem is that only about 55% of bootcamp grads are getting a job as a software engineer, and of the ones that do, their median salary is only about $65,000. At the same time, companies everywhere are just desperately looking for ways to diversify their talent pool.”

Instead of having students follow a fixed curriculum, Formation leverages adaptive learning technology to build a personalized training plan tailored to each student’s specific skillset and career goals. The platform continuously assesses their skills and adapts their roadmap, according to Sophie.

About half of the people participating in Formation’s program are current engineers already working in the industry in some capacity. 

Connie Chan, general partner at Andreessen Horowitz, said she’s been examining the edtech space for a while, including companies building new tools for teaching and upleveling coding skills. 

Formation stood out to her as the “only true tech-based and scalable solution that optimizes each student’s mastery of important skills.” Its ability to dynamically change based on a student’s performance in particular was compelling.

“The founder-product fit is also super clear — Sophie brings her own best-in-class engineering experience to Formation, as well as her long-time passion for mentoring,” Chan wrote via email.

❌