Category: Tech news

Apple popped out a new pair of AirPods this week. The design looks exactly like the old pair of AirPods. Which means I’m never going to use them because Apple’s bulbous earbuds don’t fit my ears. Think square peg, round hole.

The only way I could rock AirPods would be to walk around with hands clamped to the sides of my head to stop them from falling out. Which might make a nice cut in a glossy Apple ad for the gizmo — suggesting a feeling of closeness to the music, such that you can’t help but cup; a suggestive visual metaphor for the aural intimacy Apple surely wants its technology to communicate.

But the reality of trying to use earbuds that don’t fit is not that at all. It’s just shit. They fall out at the slightest movement so you either sit and never turn your head or, yes, hold them in with your hands. Oh hai, hands-not-so-free-pods!

The obvious point here is that one size does not fit all — howsoever much Apple’s Jony Ive and his softly spoken design team believe they have devised a universal earbud that pops snugly in every ear and just works. Sorry, nope!

A proportion of iOS users — perhaps other petite women like me, or indeed men with less capacious ear holes — are simply being removed from Apple’s sales equation where earbuds are concerned. Apple is pretending we don’t exist.

Sure we can just buy another brand of more appropriately sized earbuds. The in-ear, noise-canceling kind are my preference. Apple does not make ‘InPods’. But that’s not a huge deal. Well, not yet.

It’s true, the consumer tech giant did also delete the headphone jack from iPhones. Thereby depreciating my existing pair of wired in-ear headphones (if I ever upgrade to a 3.5mm-jack-less iPhone). But I could just shell out for Bluetooth wireless in-ear buds that fit my shell-like ears and carry on as normal.

Universal in-ear headphones have existed for years, of course. A delightful design concept. You get a selection of different sized rubber caps shipped with the product and choose the size that best fits.

Unfortunately Apple isn’t in the ‘InPods’ business though. Possibly for aesthetic reasons. Most likely because — and there’s more than a little irony here — an in-ear design wouldn’t be naturally roomy enough to fit all the stuff Siri needs to, y’know, fake intelligence.

Which means people like me with small ears are being passed over in favor of Apple’s voice assistant. So that’s AI: 1, non-‘standard’-sized human: 0. Which also, unsurprisingly, feels like shit.

I say ‘yet’ because if voice computing does become the next major computing interaction paradigm, as some believe — given how Internet connectivity is set to get baked into everything (and sticking screens everywhere would be a visual and usability nightmare; albeit microphones everywhere is a privacy nightmare… ) — then the minority of humans with petite earholes will be at a disadvantage vs those who can just pop in their smart, sensor-packed earbud and get on with telling their Internet-enabled surroundings to do their bidding.

Will parents of future generations of designer babies select for adequately capacious earholes so their child can pop an AI in? Let’s hope not.

We’re also not at the voice computing singularity yet. Outside the usual tech bubbles it remains a bit of a novel gimmick. Amazon has drummed up some interest with in-home smart speakers housing its own voice AI Alexa (a brand choice that has, incidentally, caused a verbal headache for actual humans called Alexa). Though its Echo smart speakers appear to mostly get used as expensive weather checkers and egg timers. Or else for playing music — a function that a standard speaker or smartphone will happily perform.

Certainly a voice AI is not something you need with you 24/7 yet. Prodding at a touchscreen remains the standard way of tapping into the power and convenience of mobile computing for the majority of consumers in developed markets.

The thing is, though, it still grates to be ignored. To be told — even indirectly — by one of the world’s wealthiest consumer technology companies that it doesn’t believe your ears exist.

Or, well, that it’s weighed up the sales calculations and decided it’s okay to drop a petite-holed minority on the cutting room floor. So that’s ‘ear meet AirPod’. Not ‘AirPod meet ear’ then.

But the underlying issue is much bigger than Apple’s (in my case) oversized earbuds. Its latest shiny set of AirPods are just an ill-fitting reminder of how many technology defaults simply don’t ‘fit’ the world as claimed.

Because if cash-rich Apple’s okay with promoting a universal default (that isn’t), think of all the less well resourced technology firms chasing scale for other single-sized, ill-fitting solutions. And all the problems flowing from attempts to mash ill-mapped technology onto society at large.

When it comes to wrong-sized physical kit I’ve had similar issues with standard office computing equipment and furniture. Products that seems — surprise, surprise! — to have been default designed with a 6ft strapping guy in mind. Keyboards so long they end up gifting the smaller user RSI. Office chairs that deliver chronic back-pain as a service. Chunky mice that quickly wrack the hand with pain. (Apple is a historical offender there too I’m afraid.)

The fixes for such ergonomic design failures is simply not to use the kit. To find a better-sized (often DIY) alternative that does ‘fit’.

But a DIY fix may not be an option when discrepancy is embedded at the software level — and where a system is being applied to you, rather than you the human wanting to augment yourself with a bit of tech, such as a pair of smart earbuds.

With software, embedded flaws and system design failures may also be harder to spot because it’s not necessarily immediately obvious there’s a problem. Oftentimes algorithmic bias isn’t visible until damage has been done.

And there’s no shortage of stories already about how software defaults configured for a biased median have ended up causing real-world harm. (See for example: ProPublica’s analysis of the COMPAS recidividism tool — software it found incorrectly judging black defendants more likely to offend than white. So software amplifying existing racial prejudice.)

Of course AI makes this problem so much worse.

Which is why the emphasis must be on catching bias in the datasets — before there is a chance for prejudice or bias to be ‘systematized’ and get baked into algorithms that can do damage at scale.

The algorithms must also be explainable. And outcomes auditable. Transparency as disinfectant; not secret blackboxes stuffed with unknowable code.

Doing all this requires huge up-front thought and effort on system design, and an even bigger change of attitude. It also needs massive, massive attention to diversity. An industry-wide championing of humanity’s multifaceted and multi-sized reality — and to making sure that’s reflected in both data and design choices (and therefore the teams doing the design and dev work).

You could say what’s needed is a recognition there’s never, ever a one-sized-fits all plug.

Indeed, that all algorithmic ‘solutions’ are abstractions that make compromises on accuracy and utility. And that those trade-offs can become viciously cutting knives that exclude, deny, disadvantage, delete and damage people at scale.

Expensive earbuds that won’t stay put is just a handy visual metaphor.

And while discussion about the risks and challenges of algorithmic bias has stepped up in recent years, as AI technologies have proliferated — with mainstream tech conferences actively debating how to “democratize AI” and bake diversity and ethics into system design via a development focus on principles like transparency, explainability, accountability and fairness — the industry has not even begun to fix its diversity problem.

It’s barely moved the needle on diversity. And its products continue to reflect that fundamental flaw.

Many — if not most — of the tech industry’s problems can be traced back to the fact that inadequately diverse teams are chasing scale while lacking the perspective to realize their system design is repurposing human harm as a de facto performance measure. (Although ‘lack of perspective’ is the charitable interpretation in certain cases; moral vacuum may be closer to the mark.)

As WWW creator, Sir Tim Berners-Lee, has pointed out, system design is now society design. That means engineers, coders, AI technologists are all working at the frontline of ethics. The design choices they make have the potential to impact, influence and shape the lives of millions and even billions of people.

And when you’re designing society a median mindset and limited perspective cannot ever be an acceptable foundation. It’s also a recipe for product failure down the line.

The current backlash against big tech shows that the stakes and the damage are very real when poorly designed technologies get dumped thoughtlessly on people.

Life is messy and complex. People won’t fit a platform that oversimplifies and overlooks. And if your excuse for scaling harm is ‘we just didn’t think of that’ you’ve failed at your job and should really be headed out the door.

Because the consequences for being excluded by flawed system design are also scaling and stepping up as platforms proliferate and more life-impacting decisions get automated. Harm is being squared. Even as the underlying industry drum hasn’t skipped a beat in its prediction that everything will be digitized.

Which means that horribly biased parole systems are just the tip of the ethical iceberg. Think of healthcare, social welfare, law enforcement, education, recruitment, transportation, construction, urban environments, farming, the military, the list of what will be digitized — and of manual or human overseen processes that will get systematized and automated — goes on.

Software — runs the industry mantra — is eating the world. That means badly designed technology products will harm more and more people.

But responsibility for sociotechnical misfit can’t just be scaled away as so much ‘collateral damage’.

So while an ‘elite’ design team led by a famous white guy might be able to craft a pleasingly curved earbud, such an approach cannot and does not automagically translate into AirPods with perfect, universal fit.

It’s someone’s standard. It’s certainly not mine.

We can posit that a more diverse Apple design team might have been able to rethink the AirPod design so as not to exclude those with smaller ears. Or make a case to convince the powers that be in Cupertino to add another size choice. We can but speculate.

What’s clear is the future of technology design can’t be so stubborn.

It must be radically inclusive and incredibly sensitive. Human-centric. Not locked to damaging defaults in its haste to impose a limited set of ideas.

Above all, it needs a listening ear on the world.

Indifference to difference and a blindspot for diversity will find no future here.

Autonomous delivery startup Nuro is bursting with ideas since SoftBank invested nearly $1 billion in February, new filings reveal.

A recent patent application details how its R1 self-driving vehicle could carry smaller robots to cross lawns or climb stairs to drop off packages. The company has even taken the step of trademarking the name “Fido” for delivery services.

“We think there’s something neat about that name,” Nuro founder Dave Ferguson told TechCrunch. “It’s friendly, neighborly and embodies the spirit of a helper that brings you things. It wasn’t intended to extend towards literal robot dogs, although some of the legged platforms that others are building could be very interesting for this last 10-foot problem.”

Another section of Nuro’s patent shows the R1 delivering piping hot pizza and beverages, prepared en route in automated kitchens.

“We tried to build a lot of flexibility into the R1’s compartment so we could serve all the applications that people will be able to think of,” Ferguson said. “A coffee machine is actually a pretty good one. If you go to your local barista, those machines are incredibly expensive. Amortizing them over an entire neighborhood makes sense.”

As automated technologies mature, companies are focusing less on simply getting around and more on how services will connect with actual customers. Delivering goods instead of passengers also means fewer regulations to navigate.

That opportunity has prompted a number of companies, including e-commerce and logistics giant Amazon, FedEx, and numerous startups to explore autonomous delivery.  At CES this year, Continental unveiled a prototype dog-shaped robot for last-yard deliveries, while Amazon has unveiled a sidewalk robot called Scout that is already delivering packages to homes.

The first company to scale automated driving and delivery could start building revenue while those aiming for autonomous taxis are stuck in a maze of laws, safety concerns and consumer skepticism.

Origin story

Softbank’s capital allows Nuro’s founders to run with its many ideas. But even in its earliest days, they benefited from an early injection of cash.

Nuro was founded in June 2016 by Ferguson and another former Google engineer, Jiajun Zhu, after they received multi-million dollar payouts from the company’s infamous Chauffeur bonus plan. Chauffeur bonuses were intended to incentivize engineers who stuck with Google’s self-driving car project. However, the plan’s structure meant that anyone who left after the first payout in 2015 would also receive a large lump sum.

Lead engineer Anthony Levandowski appears to have earned over $125 million from the plan. He used some of the money to start Otto, a self-driving truck company that was acquired by Uber and subsequently became the focus of an epic patent and trade secrets theft lawsuit.

Court filings from that case suggest that Ferguson and Zhu received around $40 million each, although Ferguson would not confirm this. (Another Chauffeur alum, Russell Smith, got a smaller payout and quickly joined Nuro as its hardware lead).

Nuro completed its first Series A funding round in China just three months later, in a previously unreported deal that gave NetEase founder Ding Lei (aka William Ding) a seat on Nuro’s board. Ding was China’s first Internet and gaming billionaire, and was reportedly once the wealthiest person in China. However, his business empire, which spans e-commerce, education and pig farming, recently laid off large numbers of staff.

“William has been a board member and a strong supporter from the very start. But he’s not directing company decisions,” says Ferguson.

A second, U.S.-based round in June 2017 raised Nuro’s total Series A funding to $92 million.

A Nuro spinout

Nuro started pilot grocery deliveries last summer with a Kroger supermarket affiliate in the Phoenix suburb of Scottsdale. The pilot initially used modified Toyota Prius sedans and transitioned in December to its R1 vehicle. “We’re super excited about the application area,” says Ferguson. “87 percent of commerce is still local and 43 percent of all personal vehicle trips in the U.S. are for shopping and running errands.”

Meanwhile, Uber’s self-driving truck program, which had begun with the acquisition of Otto, was on its last legs. Although the program was not publicly canned until July 2018, many of its key personnel left in May. The LinkedIn profiles of engineers Jur van den Berg, Nancy Sun and Alden Woodrow show them going straight from Uber to found Ike, another self-driving truck startup, the same month.

When Ike came out of stealth mode in October, Nuro characterized its relationship with the new company as a partnership, where “we gave Ike a copy of our autonomy and infrastructure software and, in exchange, Nuro got an equity stake in Ike.”

In reality, Ike was more of a spinout. California and Delaware business records show that Ike was not incorporated until July, and shared office space with Nuro until at least the beginning of September. Ike’s founding engineers actually worked at Nuro after leaving Uber. Van den Berg can even be seen in a Nuro team photo that was shot in June and reproduced in Nuro’s Safety Report, wearing a Nuro T-shirt.

Ferguson confirmed that all three Ike founders had worked at Nuro before starting Ike.

“We are always looking for opportunities where the tech that we’ve built could help,” Ferguson said. “Trucking was a really good example, but we recognized that as a company, we couldn’t spread ourselves too thin. It made sense for both sides for the Ike co-founders to build their own independent company.”

Ike CEO Woodrow told TechCrunch recently that it’s using Nuro’s hardware designs and autonomous software, as well as data logging, maps and simulation systems. It raised $52 million in its own Series A in February.

Not to be outdone, Nuro quickly followed with an announcement of a $940 million investment by the SoftBank Vision Fund, in exchange for what Ferguson calls a “very, very significant ownership stake.” Nuro had been introduced to SoftBank after talks with Cruise fell through.

Thousands of bots

Apart from robotic dogs, what does the future hold for a newly cash-rich Nuro?

“We’re very excited about the Scottsdale pilot, but it’s basically one grocery store in one ZIP code,” says Ferguson. Shortly after our interview, Nuro announced that it would be expanding its delivery service to four more ZIP codes in Houston, Texas.

“Next year and onwards, we want to start to realize the potential of what we’re building to eventually service millions of people” Ferguson said. We’re aggressively expanding the number of partners we’re working with and we’re working on how we manufacture a vehicle at a large scale.”

Nuro will likely to partner with an established auto OEM to build a fleet of what Ferguson hopes will become tens or hundreds of thousands of driverless vehicles. Last week, it petitioned the National Highway Traffic Safety Administration (NHTSA) for exemptions to safety standards that do not make sense for a driverless vehicle – like having to install a windshield or rearview mirrors.

Nuro told NHTSA that it wants to introduce up to 5,000 upgraded vehicles called the R2X, over the next two years. The electric vehicles would have a top speed of 25 miles per hour and appear very similar to the R1 prototype operating in Arizona and Texas today. The R2X will have 12 high-def cameras, radars, and a top-mounted LiDar sensor. Nuro said it would not sell the vehicle but “own and centrally operate the entire fleet of R2Xs through partnerships with local businesses.”

“Providing services is also very expensive,” Ferguson explained. “Look at Uber or Lyft. As we scale up to the population we’re trying to serve and the number of verticals we’re looking at, it requires capital to operate until we’re profitable, which will not happen this year.”

As I’m sure everyone reading this knows, female-founded businesses receive just over 2 percent of venture capital on an annual basis. Most of those checks are written to early-stage startups. It’s extremely difficult for female founders to garner late-stage support, let alone cash $100 million checks.

Maybe that’s finally changing. This week, not one but two female-founded and led companies, Glossier and Rent The Runway, raised nine-figure rounds and cemented their status as unicorn companies. According to PitchBook data from 2018, there are only about 15 unicorn startups with female founders. Though I’m sure that number has increased in the last year, you get the point: There are hundreds of privately held billion-dollar companies and shockingly few of those have women founders (even fewer have female CEOs)…

Moving on…

YC Demo Days

I spent a good part of the week at San Francisco’s Pier 48 in a room full of vest-wearing investors. We listened to some 200 YC companies make their 120-second pitch and though it was a bit of a whirlwind, there were definitely some standouts. ICYMI: We wrote about each and every company that pitched on day 1 and day 2. If you’re looking for the inside scoop on the companies that forwent demo day and raised rounds, or were acquired, before hitting the stage, we’ve got that too.

IPO corner

Lyft: This week, Lyft set the terms for its highly-anticipated initial public offering, expected to be completed next week. The company will charge between $62 and $68 per share, raising more than $2 billion at a valuation of ~$23 billion. We previously reported its initial market cap would be around $18.5 billion, but that was before we knew that Lyft’s IPO was already oversubscribed. Here’s a little more background on the Lyft IPO for those interested.

Uber: The global ride-hailing business flew a little more under the radar this week than last week, but still managed to grab a few headlines. The company has decided to sell its stock on the New York Stock Exchange, which is the least surprising IPO development of 2019, considering its key U.S. competitor, Lyft, has been working with the Nasdaq on its IPO. Uber is expected to unveil its S-1 in April.

Ben Silbermann, co-founder and CEO of Pinterest, at TechCrunch Disrupt SF 2017.

Pinterest: Pinterest, the nearly decade-old visual search engine, unveiled its S-1 on Friday, one of the final steps ahead of its NYSE IPO, expected in April. The $12.3 billion company, which will trade under the ticker symbol “PINS,” posted revenue of $755.9 million in the year ending December 31, 2018, up from $472.8 million in 2017. It has roughly doubled its monthly active user count since early 2016, hitting 265 million last year. The company’s net loss, meanwhile, shrank to $62.9 million in 2018 from $130 million in 2017.

Zoom: Not necessarily the buzziest of companies, but its S-1 filing, published Friday, stands out for one important reason: Zoom is profitable! I know, what insanity! Anyway, the startup is going public on the Nasdaq as soon as next month after raising about $150 million in venture capital funding. The full deets are here.

Seed money

General Catalyst, a well-known venture capital firm, is diving more seriously into the business of funding seed-stage business. The firm, which has investments in Warby Parker, Oscar and Stripe, announced earlier this week its plan to invest at least $25 million each year in nascent teams.

Deal of the week

Earlier this week, Opendoor, the SoftBank -backed real estate startup, filed paperwork to raise even more money. According to TechCrunch’s Ingrid Lunden, the business is planning to raise up to $200 million at a valuation of roughly $3.7 billion. It’s possible this is a Series E extension; after all, the company raised its $400 million Series E only six months ago. Backers of OpenDoor include the usual suspects: Andreessen Horowitz, Coatue, General Atlantic, GV, Initialized Capital, Khosla Ventures, NEA and Norwest Venture Partners.

Startup capital

• UiPath is raising $400M at a more than $7B valuation
• Ola raises $300M as part of a new EV deal with Hyundai and Kia
• Music startup Splice raises $57.5M to sell samples
• Iterable lands $50M to expand its cross-channel marketing platform
• Guesty, meant for property managers on Airbnb, raises $35M
• Travis Kalanick invests in Kargo, the ‘Uber for Trucks’
• Catch emerges from Y Combinator with $5.1M

Backstage Capital founder and managing partner Arlan Hamilton, center.

Debate

Axios’ Dan Primack and Kia Kokalitcheva published a report this week revealing Backstage Capital hadn’t raised its debut fund in total. Backstage founder Arlan Hamilton was quick to point out that she had been honest about the challenges of fundraising during various speaking engagements, and even on the Gimlet “Startup” podcast, which featured her in its latest season. A Twitter debate ensued and later, Hamilton announced she was stepping down as CEO of Backstage Studio, the operations arm of the venture fund, to focus on raising capital and amplifying founders. TechCrunch’s Megan Rose Dickey has the full story.

Pro rata rights

This week, TechCrunch’s Connie Loizos revisited a long-held debate: Pro rata rights, or the right of an earlier investor in a company to maintain the percentage that he or she (or their venture firm) owns as that company matures and takes on more funding. Here’s why pro rata rights matter (at least, to VCs).

#Equitypod

If you enjoy this newsletter, be sure to check out TechCrunch’s venture-focused podcast, Equity. In this week’s episode, available here, Crunchbase News editor-in-chief Alex Wilhelm and I chat about Glossier, Rent The Runway and YC Demo Days. Then, in a special Equity Shot, we unpack the numbers behind the Pinterest and Zoom IPO filings.

Want more TechCrunch newsletters? Sign up here.

Tesla CEO Elon Musk argued Friday that his Twitter use did not violate a settlement agreement with the U.S. Securities and Exchange Commission and that the agency’s request to have him held in contempt is based on a “radical interpretation” of the order, according to court papers filed in Manhattan federal court.

The SEC has asked a judge to hold Musk in contempt for violating a settlement agreement reached last year over Musk’s now infamous “funding secured” tweet. Under that agreement, Musk is supposed to get approval from Tesla’s board before communicating potentially material information to investors.

Musk contends he didn’t violate the agreement and that the problem lies in the SEC’s interpretation, which he describes as “virtually wrong at every level.” The filing also reveals new details about the settlement negotiations, notably that the SEC sent Musk a draft agreement that would have required him to obtain pre-approval for all public statements related to Tesla, in any format.

Musk and Tesla never agreed to those terms. Instead, Musk says the agreement requires him to comply with Tesla own policy, which would require pre-approval for “written communications that contain, or reasonably could contain, information material to the company or its shareholders.”

The barbs traded via court filings are the latest in an escalating fight between the billionaire entrepreneur and SEC that began last August when Musk tweeted that he had “funding secured” for a private takeover of the company at $420 per share.  The SEC filed a complaint in federal district court in September alleging that Musk lied.

Musk and Tesla settled with the SEC last year without admitting wrongdoing. Tesla agreed to pay a $20 million fine; Musk had to agree to step down as Tesla chairman for a period of at least three years; the company had to appoint two independent directors to the board; and Tesla was also told to put in place a way to monitor Musk’s statements to the public about the company, including via Twitter.

But the fight was re-ignited last month after Musk sent a tweet on February 19 that Tesla would produce “around” 500,000 cars this year, correcting himself hours later to clarify that he meant the company would be producing at an annualized rate of 500,000 vehicles by year end.

The SEC argued that the tweet sent by Musk violated their agreement. Musk has said the tweet was “immaterial” and complied with the settlement.

This post reveals the cost of acquiring a customer on every ad channel my agency has tested. The ad channels include Facebook, Instagram, YouTube, Quora, Google Search, Google Shopping, Snapchat, LinkedIn and others. Using this data, you can reduce your costs by identifying which channels are a likely fit for your own product. Then you can focus on testing just those channels to start. I’m pulling data from my agency’s experience testing 15+ ad channels and running thousands of ads for dozens of Y Combinator startups.

This post leaves you with a prioritized to-do list of which channels might work for your product, and reference points for how much you can expect to pay if you get those channels to work.

Which ad channels should I use?

We focus on three criteria when assessing ad channels:

• Profit — You want to earn at least as much as the cost of acquiring a customer. For example, it’s uncommon to acquire an American customer through Facebook for less than $30 USD. So, your profit per customer should be at least $30 to break even. (In reality, it should likely be 3x more to account for meaningful profit and ad channel volatility.)
• Volume — If your audience doesn’t exist in significant quantities on a given ad channel, it’s likely not worth your time to experiment with it yet. Especially given your effective audience volume is probably smaller than you think: It’s not just a matter of how many people use the channel, but how many who use it also want your product today and can justify its cost.
• Targeting — The best-fitting ad channels are those that let you narrowly target your desired audience. If they can’t do this, you’ll be forced to target broadly, which wastes dollars on the wrong eyeballs. This means your customer acquisition cost (CAC) is more likely to exceed your profit margin.

In short, to succeed with ad channels, your product should earn sufficient profit and have a sufficiently large targetable audience.

Even the most successful tech company is going to have a stumble from time to time. Nintendo’s 45 years in the video game industry is spotted with a few doozies, but none are more infamous than the Virtual Boy. The 1994 portable console was marketed as an early home entry into virtual reality, but in actual reality ended up being little more than a blood-red headache.

Nintendo knew the comparisons to the doomed console would come fast and furiously when it launched its next VR venture, so the company took the time to get it just right. In a sense, Labo VR is a cautious push into the virtual realm. It’s nowhere near the all-in approach of Oculus, Vive or even PlayStation VR, for that matter — but it’s uniquely Nintendo.

Like the first Labo kits, it’s a friendly reminder that Nintendo’s chief job is to surprise and delight, and it happily delivers on both fronts. But just as the Labo piano shouldn’t be mistaken for a real musical instrument, Labo VR ought not be viewed as a real virtual reality.

It’s not just the pop-out cardboard form factor, either. Google made that a perfectly acceptable beginner’s approach to VR. It’s more that Nintendo has taken a very casual approach to all of this. The kit’s virtual reality experience is an extension of Labo itself. It’s no more important than the process of building the headset and various accessories step by step on the app. Or, for that matter, sharing all of the above experiences with others.

During a demo of the new kits in New York this week, Nintendo was quick to point out that the headsets are built without a strap. It claims this was a conscious decision so that the experience can be passed around and shared. I’m sure there are some practical reasons behind this decision as well, but it’s certainly a nice thought.

Virtual reality is, by nature of its form factor, a solitary experience. Labo VR doesn’t have any sort of video-out feature to share the experience on a big screen (for now, at least), so the idea of offering it up in a more social play-and-pass scenario is appealing. This goes double for the fact that, like the original Labo kits, all of the games included fall under the casual banner. The experiences share a common lineage with Nintendo analog titles like Mario Party or Mario Paint.

Your mileage with each title will vary. Certainly some (Bird and Blaster spring to mind) will stay with you longer than others and demand more repeat play. On the whole, each buildable peripheral launches with one (maybe two) compatible games. The good news, however, is that, like Labo, the company packs a lot of controllers (and therefore experiences) into a single kit.

The standard Labo: VR Kit ships with six Toy-Con projects (VR Goggles, Toy-Con Blaster, Toy-Con Camera, Toy-Con Bird, Toy-Con Wind Pedal and Toy-Con Elephant), while the cheaper Starter Set comes with two (Goggles and Blaster). If you go for the latter to dip your toes in the water or just to save on cash, there are a pair of “expansion sets” to get the full experience.

Unlike the last time Nintendo came to town with a Labo press tour, we didn’t actually get any time to build. That said, if previous kits are any indication, that’s half of the fun and value proposition here. Also, the amount of time you’ll spend building varies greatly from project to project — take it from me, someone who spent most of a work morning building that damn piano.

Once built, the VR experience is about on-par with what you’d expect from a Google VR. Again, it’s a set of lenses attached to a hunk of cardboard. This is no Rift or Vive and the immersiveness of your own experience will vary. The graphics are cartoony and oftentimes just large polygons. But a well-crafted casual gaming experience can be enough to pull you out of your own head for a bit. Bird is the best example of this.

The controller clips on the headset, with a Toy-Con popping out the other end like a beak. As a player, you hook your hands on either side of the display and flap along as you play a bird, flying around trees and completing different missions to feed an army of hatchlings. It’s a relaxing reprieve from some of the faster-paced games, as you glide around the skies. Add in the foot-controlled Wind Pedal, and the system delivers a puff of air to your face as you boost your bird, adding to the effect.

Blaster, a big, fun novelty gun, is the most engaging of the bunch. When I ended my demos with some extra time to spare, the Nintendo rep asked me if I wanted to give any of the games another go. The answer was simple. A simple first-person shooter, Blaster pits you against an army of alien blobs. You load the gun by cocking it like a shot-gun, and pull the trigger to an explosive effect.

Honorable mention goes to Doodle, which uses the bizarre elephant-shaped controller. The experience is unique from the rest in that it’s not actually a game, but rather a 3D drawing tool. It’s one of the more clever additions to the pack, though actually drawing on a 3D plane with a cardboard controller shaped like an elephant’s trunk is easier said than done. The implementation is a bit lacking, but it offers interesting insight into where Labo VR might go in the future.

Honestly, I just scratched the surface during my briefing. But there’s little question that Labo VR is a fun and singular experience. There’s also a special screen holder, so users who have rough time with VR can experience a 2D version of the games and accessories. Also, as with the standard Labo kit, Nintendo has bundled in Toy-Con Garage, so users can start building their own games when they tire of the pre-packaged experiences.

If there’s one disappointment in all of this, it’s that it will likely be a while before we see a full standalone VR experience from Nintendo. The idea of playing as Mario, Link and the like in virtual reality is no doubt something of a lifelong dream for plenty of gamers who grew up on the characters. But while Virtual Boy is a quarter-century in the past, the memory still lingers.

Until then, Labo VR is a fully engaging take on VR, and a uniquely Nintendo one, to boot.

Catch up on the most important tech news today in two minutes or less.

More than two years into the Trump administration, the long vacant post of U.S. Chief Technology Officer will be filled. Bloomberg first reported that today Trump is elevating Michael Kratsios, current deputy U.S. CTO, to the nation’s top tech position. Prior to his experience within the Trump administration, Kratsios served as chief of staff at Peter Thiel’s investment firm Thiel Capital and as chief financial officer at another Thiel project, the hedge fund Clarium Capital.

The U.S. CTO role was created during the Obama years and three CTOs have served to date, the last of which was former Googler Megan Smith, known for leading early acquisitions at Google before her move to Google.org.

The CTO position advises the president on tech issues, works to shape tech policy and importantly serves as a link to the private sector. In contrast with his predecessors, Kratsios brings a distinct venture capital-colored perspective to the role, which sits within the White House Office of Science and Technology Policy.

The electric car company says four former employees took proprietary information to their new jobs at Zoox. Oh, and Tesla sued an employee of another rival, too.

TypeScript, Microsoft’s JavaScript Alternative, Continues to Rise in Popularity

European governments have been bringing the hammer down on tech in recent months, slapping record fines and stiff regulations on the largest imports out of Silicon Valley. Despite pleas from the world’s leading companies and Europe’s eroding trust in government, European citizens’ staunch support for regulation of new technologies points to an operating environment that is only getting tougher.

According to a roughly 25-page report recently published by a research arm out of Spain’s IE University, European citizens remain skeptical of tech disruption and want to handle their operators with kid gloves, even at a cost to the economy.

The survey was led by the IE’s Center for the Governance of Change — an IE-hosted research institution focused on studying “the political, economic, and societal implications of the current technological revolution and advances solutions to overcome its unwanted effects.” The “European Tech Insights 2019” report surveyed roughly 2,600 adults from various demographics across seven countries (France, Germany, Ireland, Italy, Spain, The Netherlands, and the UK) to gauge ground-level opinions on ongoing tech disruption and how government should deal with it.

The report does its fair share of fear-mongering and some of its major conclusions come across as a bit more “clickbaity” than insightful. However, the survey’s more nuanced data and line of questioning around specific forms of regulation offer detailed insight into how the regulatory backdrop and operating environment for European tech may ultimately evolve.

Distractions

DNA computers have to date only been able to run one algorithm, but a new design shows how these machines can be made more flexible—and useful.

Hundreds gathered this week at San Francisco’s Pier 48 to see the more than 200 companies in Y Combinator’s Winter 2019 cohort present their two-minute pitches. The audience of venture capitalists, who collectively manage hundreds of billions of dollars, noted their favorites. The very best investors, however, had already had their pick of the litter.

What many don’t realize about the Demo Day tradition is that pitching isn’t a requirement; in fact, some YC graduates skip out on their stage opportunity altogether. Why? Because they’ve already raised capital or are in the final stages of closing a deal.

ZeroDown, Overview.AI and Catch are among the startups in YC’s W19 batch that forwent Demo Day this week, having already pocketed venture capital. ZeroDown, a financing solution for real estate purchases in the Bay Area, raised a round upwards of $10 million at a $75 million valuation, sources tell TechCrunch. ZeroDown hasn’t responded to requests for comment, nor has its rumored lead investor: Goodwater Capital.

Without requiring a down payment, ZeroDown purchases homes outright for customers and helps them work toward ownership with monthly payments determined by their income. The business was founded by Zenefits co-founder and former chief technology officer Laks Srini, former Zenefits chief operating officer Abhijeet Dwivedi and Hari Viswanathan, a former Zenefits staff engineer.

The founders’ experience building Zenefits, despite its shortcomings, helped ZeroDown garner significant buzz ahead of Demo Day. Sources tell TechCrunch the startup had actually raised a small seed round ahead of YC from former YC president Sam Altman, who recently stepped down from the role to focus on OpenAI, an AI research organization. Altman is said to have encouraged ZeroDown to complete the respected Silicon Valley accelerator program, which, if nothing else, grants its companies a priceless network with which no other incubator or accelerator can compete.

Overview .AI’s founders’ resumes are impressive, too. Russell Nibbelink and Christopher Van Dyke were previously engineers at Salesforce and Tesla, respectively. An industrial automation startup, Overview is developing a smart camera capable of learning a machine’s routine to detect deviations, crashes or anomalies. TechCrunch hasn’t been able to get in touch with Overview’s team or pinpoint the size of its seed round, though sources confirm it skipped Demo Day because of a deal.

Catch, for its part, closed a $5.1 million seed round co-led by Khosla Ventures, Kindred Ventures, and NYCA Partners prior to Demo Day. Instead of pitching their health insurance platform at the big event, Catch published a blog post announcing its first feature, The Catch Health Explorer.

“This is only the first glimpse of what we’re building this year,” Catch wrote in the blog post. “In a few months, we’ll be bringing end-to-end health insurance enrollment for individual plans into Catch to provide the best health insurance enrollment experience in the country.”

TechCrunch has more details on the healthtech startup’s funding, which included participation from Kleiner Perkins, the Urban Innovation Fund and the Graduate Fund.

Four more startups, Truora, Middesk, Glide and FlockJay had deals in the final stages when they walked onto the Demo Day stage, deciding to make their pitches rather than skip the big finale. Sources tell TechCrunch that renowned venture capital firm Accel invested in both Truora and Middesk, among other YC W19 graduates. Truora offers fast, reliable and affordable background checks for the Latin America market, while Middesk does due diligence for businesses to help them conduct risk and compliance assessments on customers.

Finally, Glide, which allows users to quickly and easily create well-designed mobile apps from Google Sheets pages, landed support from First Round Capital, and FlockJay, the operator an online sales academy that teaches job seekers from underrepresented backgrounds the skills and training they need to pursue a career in tech sales, secured investment from Lightspeed Venture Partners, according to sources familiar with the deal.

Pre-Demo Day M&A

Raising ahead of Demo Day isn’t a new phenomenon. Companies, thanks to the invaluable YC network, increase their chances at raising, as well as their valuation, the moment they enroll in the accelerator. They can begin chatting with VCs when they see fit, and they’re encouraged to mingle with YC alumni, a process that can result in pre-Demo Day acquisitions.

This year, Elph, a blockchain infrastructure startup, was bought by Brex, a buzzworthy fintech unicorn that itself graduated from YC only two years ago. The deal closed just one week before Demo Day. Brex’s head of engineering, Cosmin Nicolaescu, tells TechCrunch the Elph five-person team — including co-founders Ritik Malhotra and Tanooj Luthra, who previously founded the Box-acquired startup Steem — were being eyed by several larger companies as Brex negotiated the deal.

“For me, it was important to get them before batch day because that opens the floodgates,” Nicolaescu told TechCrunch. “The reason why I really liked them is they are very entrepreneurial, which aligns with what we want to do. Each of our products is really like its own business.”

Of course, Brex offers a credit card for startups and has no plans to dabble with blockchain or cryptocurrency. The Elph team, rather, will bring their infrastructure security know-how to Brex, helping the $1.1 billion company build its next product, a credit card for large enterprises. Brex declined to disclose the terms of its acquisition.

Hunting for the best deals

Y Combinator partners Michael Seibel and Dalton Caldwell, and moderator Josh Constine, speak onstage during TechCrunch Disrupt SF 2018. (Photo by Kimberly White/Getty Images)

Ultimately, it’s up to startups to determine the cost at which they’ll give up equity. YC companies raise capital under the SAFE model, or a simple agreement for future equity, a form of fundraising invented by YC. Basically, an investor makes a cash investment in a YC startup, then receives company stock at a later date, typically upon a Series A or post-seed deal. YC made the switch from investing in startups on a pre-money safe basis to a post-money safe in 2018 to make cap table math easier for founders.

Michael Seibel, the chief executive officer of YC, says the accelerator works with each startup to develop a personalized fundraising plan. The businesses that raise at valuations north of $10 million, he explained, do so because of high demand.

“Each company decides on the amount of money they want to raise, the valuation they want to raise at, and when they want to start fundraising,” Seibel told TechCrunch via email. “YC is only an advisor and does not dictate how our companies operate. The vast majority of companies complete fundraising in the 1 to 2 months after Demo Day. According to our data, there is little correlation between the companies who are most in demand on Demo Day and ones who go on to become extremely successful. Our advice to founders is not to over optimize the fundraising process.”

Though Seibel says the majority raise in the months following Demo Day, it seems the very best investors know to be proactive about reviewing and investing in the batch before the big event.

Khosla Ventures, like other top VC firms, meets with YC companies as early as possible, partner Kristina Simmons tells TechCrunch, even scheduling interviews with companies in the period between when a startup is accepted to YC to before they actually begin the program. Another Khosla partner, Evan Moore, echoed Seibel’s statement, claiming there isn’t a correlation between the future unicorns and those that raise capital ahead of Demo Day. Moore is a co-founder of DoorDash, a YC graduate now worth $7.1 billion. DoorDash closed its first round of capital in the weeks following Demo Day.

“I think a lot of the activity before demo day is driven by investor FOMO,” Moore wrote in an email to TechCrunch. “I’ve had investors ask me how to get into a company without even knowing what the company does! I mostly see this as a side effect of a good thing: YC has helped tip the scale toward founders by creating an environment where investors compete. This dynamic isn’t what many investors are used to, so every batch some complain about valuations and how easy the founders have it, but making it easier for ambitious entrepreneurs to get funding and pursue their vision is a good thing for the economy.”

This year, given the number of recent changes at YC — namely the size of its latest batch — there was added pressure on the accelerator to showcase its best group yet. And while some did tell TechCrunch they were especially impressed with the lineup, others indeed expressed frustration with valuations.

Many YC startups are fundraising at valuations at or higher than $10 million. For context, that’s actually perfectly in line with the median seed-stage valuation in 2018. According to PitchBook, U.S. startups raised seed rounds at a median post-valuation of $10 million last year; so far this year, companies are raising seed rounds at a slightly higher post-valuation of $11 million. With that said, many of the startups in YC’s cohorts are not as mature as the average seed-stage company. Per PitchBook, a company can be several years of age before it secures its seed round.

Nonetheless, pricey deals can come as a disappointment to the seed investors who find themselves at YC every year but because their reputations aren’t as lofty as say, Accel, aren’t able to book pre-Demo Day meetings with YC’s top of class.

The question is who is Y Combinator serving? And the answer is founders, not investors. YC is under no obligation to serve up deals of a certain valuation nor is it responsible for which investors gain access to its best companies at what time. After all, startups are raking in larger and larger rounds, earlier in their lifespans; shouldn’t YC, a microcosm for the Silicon Valley startup ecosystem, advise their startups to charge the best investors the going rate?

One of the hottest Y Combinator startups just raised a big seed round to clean up the mess created by Uber, Postmates and the gig economy. Catch sells health insurance, retirement savings plans and tax withholding directly to freelancers, contractors, or anyone uncovered. By building and curating simplified benefits services, Catch can offer a safety net for the future of work.

“In order to stay competitive as a society, we need to address inequality and volatility. We think Catch is the first step to offering alternatives to the mandate that benefits can only come from an employer or the government,” writes Catch co-founder and COO Kristen Tyrrell. Her co-founder and CEO Andrew Ambrosino, a former Kleiner Perkins design fellow, stumbled onto the problem as he struggled to juggle all the paperwork and programs companies typically hire an HR manager to handle. “Setting up a benefits plan was a pain. You had to become an expert in the space, and even once you were, executing and getting the stuff you needed was pretty difficult.” Catch does all this annoying but essential work for you.

Now Catch is getting its first press after piloting its product with tens of thousands of users. TechCrunch caught wind of its highly competitive seed round closing, and Catch confirms it has raised $5.1 million at a $20.5 million post-money valuation co-led by Khosla Ventures, Kindred Ventures, and NYCA Partners. This follow-up to its $1 million pre-seed will fuel its expansion into full heath insurance enrollment, life insurance and more. Catch is part of a growing trend that sees the best Y Combinator startup fully funded before Demo Day even arrives.

“Benefits, as a system built and provided by employers, created the mid-century middle class. In the post-war economic boom, companies offering benefits in the form of health insurance and pensions enabled familial stability that led to expansive growth and prosperity,” recalls Tyrrell, who was formerly the director of product at student debt repayment benefits startup FutureFuel.io. “Emboldened by private-sector growth (and apparent self-sufficiency), the 1970s and 80s saw a massive shift in financial risk management from the government to employers. The public safety net contracted in favor of privatized solutions. As technological advances progressed, employers and employees continued to redefine what work looked like. The bureaucratic and inflexible benefits system was unable to keep up. The private safety net crumbled.”

That problem has ballooned in recent years with the advent of the on-demand economy, where millions become Uber drivers, Instacart shoppers, DoorDash deliverers and TaskRabbits. Meanwhile, the destigmatization of remote work and digital nomadism has turned more people into permanent freelancers and contractors, or full-time employees without benefits. “A new class of worker emerged: one with volatile, complex income streams and limited access to second-order financial products like automated savings, individual retirement plans, and independent health insurance. We entered the new millennium with rot under the surface of new opportunity from the proliferation of the internet,” Tyrrell declares. “The last 15 years are borrowed time for the unconventional proletariat. It is time to come to terms and design a safety net that is personal, portable, modern and flexible. That’s why we built Catch.”

Catch co-founders Andrew Ambrosino and Kristen Tyrrell

Currently Catch offers the following services, each with their own way of earning the startup revenue:

• Health Explorer lets users compare plans from insurers and calculate subsidies, while Catch serves as a broker collecting a fee from insurance providers
• Retirement Savings gives users a Catch robo-advisor compatible with IRA and Roth IRA, while Catch earns the industry standard 1 basis point on saved assets
• Tax Withholding provides an FDIC-insured Catch account that automatically saves what you’ll need to pay taxes later, while Catch earns interest on the funds
• Time Off Savings similarly lets you automatically squirrel away money to finance “paid” time off, while Catch earns interest

These and the rest of Catch’s services are curated through its Guide. You answer a few questions about which benefits you have and need, connect your bank account, choose which programs you want and get push notifications whenever Catch needs your decisions or approvals. It’s designed to minimize busy work so if you have a child, you can add them to all your programs with a click instead of slogging through reconfiguring them all one at a time. That simplicity has ignited explosive growth for Catch, with the balances it holds for tax withholding, time off and retirement balances up 300 percent in each of the last three months.

In 2019 it plans to add Catch-branded student loan refinancing, vision and dental enrollment plus payments via existing providers, life insurance through a partner such as Ladder or Ethos and full health insurance enrollment plus subsidies and premium payments via existing insurance companies like Blue Shield and Oscar. And in 2020 it’s hoping to build out its own blended retirement savings solution and income-smoothing tools.

If any of this sounds boring, that’s kind of the point. Instead of sorting through this mind-numbing stuff unassisted, Catch holds your hand. Its benefits Guide is available on the web today and it’s beta testing iOS and Android apps that will launch soon. Catch is focused on direct-to-consumer sales because “We’ve seen too many startups waste time on channels/partnerships before they know people truly want their product and get lost along the way,” Tyrrell writes. Eventually it wants to set up integrations directly into where users get paid.

Catch’s biggest competition is people haphazardly managing benefits with Excel spreadsheets and a mishmash of healthcare.gov and solutions for specific programs. Twenty-one percent of Americans have saved $0 for retirement, which you could see as either a challenge to scaling Catch or a massive greenfield opportunity. Track.tax, one of its direct competitors, charges a subscription price that has driven users to Catch. And automated advisors like Betterment and Wealthfront accounts don’t work so well for gig workers with lots of income volatility.

So do the founders think the gig economy, with its suppression of benefits, helps or hinders our species? “We believe the story is complex, but overall, the existing state of the gig economy is hurting society. Without better systems to provide support for freelance/contract workers, we are making people more precarious and less likely to succeed financially.”

When I ask what keeps the founders up at night, Tyrrell admits “The safety net is not built for individuals. It’s built to be distributed through HR departments and employers. We are very worried that the products we offer aren’t on equal footing with group/company products.” For example, there’s a $6,000/year IRA limit for individuals while the corporate equivalent 401k limit is $19,000, and health insurance is much cheaper for groups than individuals.

To surmount those humps, Catch assembled a huge list of angel investors who’ve built a range of financial services, including NerdWallet founder Jake Gibson, Earnest founders Louis Beryl and Ben Hutchinson, ANDCO (acquired by Fiverr) founder Leif Abraham, Totem founder Neal Khosla, Commuter Club founder Petko Plachkov, Playable (acquired by Stripe) founder Tad Milbourn and Synapse founder Bruno Faviero. It also brought on a wide range of venture funds to open doors for it. Those include Urban Innovation Fund, Kleiner Perkins, Y Combinator, Tempo Ventures, Prehype, Loup Ventures, Indicator Ventures, Ground Up Ventures and Graduate Fund.

Hopefully the fact that there are three lead investors and so many more in the round won’t mean that none feel truly accountable to oversee the company. With 80 million Americans lacking employer-sponsored benefits and 27 million without health insurance and median job tenure down to 2.8 years for people ages 25 to 34 leading to more gaps between jobs, our workforce is vulnerable. Catch can’t operate like a traditional software startup with leniency for screw-ups. If it can move cautiously and fix things, it could earn labor’s trust and become a fundamental piece of the welfare stack.

Facebook has given another update on measures it took and what more it’s doing in the wake of the livestreamed video of a gun massacre by a far right terrorist who killed 50 people in two mosques in Christchurch, New Zealand.

Earlier this week the company said the video of the slayings had been viewed less than 200 times during the livestream broadcast itself, and about about 4,000 times before it was removed from Facebook — with the stream not reported to Facebook until 12 minutes after it had ended.

None of the users who watched the killings unfold on the company’s platform in real-time apparently reported the stream to the company, according to the company.

It also previously said it removed 1.5 million versions of the video from its site in the first 24 hours after the livestream, with 1.2M of those caught at the point of upload — meaning it failed to stop 300,000 uploads at that point. Though as we pointed out in our earlier report those stats are cherrypicked — and only represent the videos Facebook identified. We found other versions of the video still circulating on its platform 12 hours later.

In the wake of the livestreamed terror attack, Facebook has continued to face calls from world leaders to do more to make sure such content cannot be distributed by its platform.

The prime minister of New Zealand, Jacinda Ardern told media yesterday that the video “should not be distributed, available, able to be viewed”, dubbing it: “Horrendous.”

She confirmed Facebook had been in contact with her government but emphasized that in her view the company has not done enough.

She also later told the New Zealand parliament: “We cannot simply sit back and accept that these platforms just exist and that what is said on them is not the responsibility of the place where they are published. They are the publisher. Not just the postman.”

We asked Facebook for a response to Ardern’s call for online content platforms to accept publisher-level responsibility for the content they distribute. Its spokesman avoided the question — pointing instead to its latest piece of crisis PR which it titles: “A Further Update on New Zealand Terrorist Attack”.

Here it writes that “people are looking to understand how online platforms such as Facebook were used to circulate horrific videos of the terrorist attack”, saying it therefore “wanted to provide additional information from our review into how our products were used and how we can improve going forward”, before going on to reiterate many of the details it has previously put out.

Including that the massacre video was quickly shared to the 8chan message board by a user posting a link to a copy of the video on a file-sharing site. This was prior to Facebook itself being alerted to the video being broadcast on its platform.

It goes on to imply 8chan was a hub for broader sharing of the video — claiming that: “Forensic identifiers on many of the videos later circulated, such as a bookmarks toolbar visible in a screen recording, match the content posted to 8chan.”

So it’s clearly trying to make sure it’s not singled out by political leaders seek policy responses to the challenge posed by online hate and terrorist content.

Further details it chooses to dwell on in the update is how the AIs it uses to aid the human content review process of flagged Facebook Live streams are in fact tuned to “detect and prioritize videos that are likely to contain suicidal or harmful acts” — with the AI pushing such videos to the top of human moderators’ content heaps, above all the other stuff they also need to look at.

Clearly “harmful acts” were involved in the New Zealand terrorist attack. Yet Facebook’s AI was unable to detected a massacre unfolding in real time. A mass killing involving an automatic weapon slipped right under the robot’s radar.

Facebook explains this by saying it’s because it does not have the training data to create an algorithm that understands it’s looking at mass murder unfolding in real time.

It also implies the task of training an AI to catch such a horrific scenario is exacerbated by the proliferation of videos of first person shooter videogames on online content platforms.

It writes: “[T]his particular video did not trigger our automatic detection systems. To achieve that we will need to provide our systems with large volumes of data of this specific kind of content, something which is difficult as these events are thankfully rare. Another challenge is to automatically discern this content from visually similar, innocuous content – for example if thousands of videos from live-streamed video games are flagged by our systems, our reviewers could miss the important real-world videos where we could alert first responders to get help on the ground.”

The videogame element is a chilling detail to consider.

It suggests that a harmful real-life act that mimics a violent video game might just blend into the background, as far as AI moderation systems are concerned; invisible in a sea of innocuous, virtually violent content churned out by gamers. (Which in turn makes you wonder whether the Internet-steeped killer in Christchurch knew — or suspected — that filming the attack from a videogame-esque first person shooter perspective might offer a workaround to dupe Facebook’s imperfect AI watchdogs.)

Facebook post is doubly emphatic that AI is “not perfect” and is “never going to be perfect”.

“People will continue to be part of the equation, whether it’s the people on our team who review content, or people who use our services and report content to us,” it writes, reiterating yet again that it has ~30,000 people working in “safety and security”, about half of whom are doing the sweating hideous toil of content review.

This is, as we’ve said many times before, a fantastically tiny number of human moderators given the vast scale of content continually uploaded to Facebook’s 2.2BN+ user platform.

Moderating Facebook remains a hopeless task because so few humans are doing it.

Moreover AI can’t really help. (Later in the blog post Facebook also writes vaguely that there are “millions” of livestreams broadcast on its platform every day, saying that’s why adding a short broadcast delay — such as TV stations do — wouldn’t at all help catch inappropriate real-time content.)

At the same time Facebook’s update makes it clear how much its ‘safety and security’ systems rely on unpaid humans too: Aka Facebook users taking the time and mind to report harmful content.

Some might say that’s an excellent argument for a social media tax.

The fact Facebook did not get a single report of the Christchurch massacre livestream while the terrorist attack unfolded meant the content was not prioritized for “accelerated review” by its systems, which it explains prioritize reports attached to videos that are still being streamed — because “if there is real-world harm we have a better chance to alert first responders and try to get help on the ground”.

Though it also says it expanded its acceleration logic last year to “also cover videos that were very recently live, in the past few hours”.

But again it did so with a focus on suicide prevention — meaning the Christchurch video would only have been flagged for acceleration review in the hours after the stream ended if it had been reported as suicide content.

So the ‘problem’ is that Facebook’s systems don’t prioritize mass murder.

“In [the first] report, and a number of subsequent reports, the video was reported for reasons other than suicide and as such it was handled according to different procedures,” it writes, adding it’s “learning from this” and “re-examining our reporting logic and experiences for both live and recently live videos in order to expand the categories that would get to accelerated review”.

No shit.

Facebook also discusses its failure to stop versions of the massacre video from resurfacing on its platform, having been — as it tells it — “so effective” at preventing the spread of propaganda from terrorist organizations like ISIS with the use of image and video matching tech.

It claims  its tech was outfoxed in this case by “bad actors” creating many different edited versions of the video to try to thwart filters, as well as by the various ways “a broader set of people distributed the video and unintentionally made it harder to match copies”.

So, essentially, the ‘virality’ of the awful event created too many versions of the video for Facebook’s matching tech to cope.

“Some people may have seen the video on a computer or TV, filmed that with a phone and sent it to a friend. Still others may have watched the video on their computer, recorded their screen and passed that on. Websites and pages, eager to get attention from people seeking out the video, re-cut and re-recorded the video into various formats,” it writes, in what reads like another attempt to spread blame for the amplification role that its 2.2BN+ user platform plays.

In all Facebook says it found and blocked more than 800 visually-distinct variants of the video that were circulating on its platform.

It reveals it resorted to using audio matching technology to try to detect videos that had been visually altered but had the same soundtrack. And again claims it’s trying to learn and come up with better techniques for blocking content that’s being re-shared widely by individuals as well as being rebroadcast by mainstream media. So any kind of major news event, basically.

In a section on next steps Facebook says improving its matching technology to prevent the spread of inappropriate viral videos being spread is its priority.

But audio matching clearly won’t help if malicious re-sharers just both re-edit the visuals and switch the soundtrack too in future.

It also concedes it needs to be able to react faster “to this kind of content on a live streamed video” — though it has no firm fixes to offer there either, saying only that it will explore “whether and how AI can be used for these cases, and how to get to user reports faster”.

Another priority it claims among its “next steps” is fighting “hate speech of all kinds on our platform”, saying this includes more than 200 white supremacist organizations globally “whose content we are removing through proactive detection technology”.

It’s glossing over plenty of criticism on that front too though — including research that suggests banned far right hate preachers are easily able to evade detection on its platform. Plus its own foot-dragging on shutting down far right extremists. (Facebook only finally banned one infamous UK far right activist last month, for example.)

In its last PR sop, Facebook says it’s committed to expanding its industry collaboration to tackle hate speech via the Global Internet Forum to Counter Terrorism (GIFCT), which formed in 2017 as platforms were being squeezed by politicians to scrub ISIS content — in a collective attempt to stave off tighter regulation.

“We are experimenting with sharing URLs systematically rather than just content hashes, are working to address the range of terrorists and violent extremists operating online, and intend to refine and improve our ability to collaborate in a crisis,” Facebook writes now, offering more vague experiments as politicians call for content responsibility.