Year: 2018

Gripping something with your hand is one of the first things you learn to do as an infant, but it’s far from a simple task, and only gets more complex and variable as you grow up. This complexity makes it difficult for machines to teach themselves to do, but researchers at Elon Musk and Sam Altman-backed OpenAI have created a system that not only holds and manipulates objects much like a human does, but developed these behaviors all on its own.

Many robots and robotic hands are already proficient at certain grips or movements — a robot in a factory can wield a bolt gun even more dexterously than a person. But the software that lets that robot do that task so well is likely to be hand-written and extremely specific to the application. You couldn’t for example, give it a pencil and ask it to write. Even something on the same production line, like welding, would require a whole new system.

Yet for a human, picking up an apple isn’t so different from pickup up a cup. There are differences, but our brains automatically fill in the gaps and we can improvise a new grip, hold an unfamiliar object securely, and so on. This is one area where robots lag severely behind their human models. And furthermore, you can’t just train a bot to do what a human does — you’d have to provide millions of examples to adequately show what a human would do with thousands of given objects.

The solution, OpenAI’s researchers felt, was not to use human data at all. Instead, they let the computer try and fail over and over in a simulation, slowly learning how to move its fingers so that the object in its grasp moves as desired.

The system, which they call Dactyl, was provided only with the positions of its fingers and three camera views of the object in-hand — but remember, when it was being trained, all this data is simulated, taking place in a virtual environment. There, the computer doesn’t have to work in real time — it can try a thousand different ways of gripping an object in a few seconds, analyzing the results and feeding that data forward into the next try. (The hand itself is a Shadow Dexterous Hand, which is also more complex than most robotic hands.)

In addition to different objects and poses the system needed to learn, there were other randomized parameters, like the amount of friction the fingertips had, the colors and lighting of the scene, and more. You can’t simulate every aspect of reality (yet), but you can make sure that your system doesn’t only work in a blue room, on cubes with special markings on them.

They threw a lot of power at the problem: 6144 CPUs and 8 GPUs, “collecting about one hundred years of experience in 50 hours.” And then they put the system to work in the real world for the first time — and it demonstrated some surprisingly human-like behaviors.

The things we do with our hands without even noticing, like turning an apple around to check for bruises or passing a mug of coffee to a friend, use lots of tiny tricks to stabilize or move the object. Dactyl recreated several of them, for example holding the object with a thumb and single finger while using the rest to spin to the desired orientation.

What’s great about this system is not just the naturalness of its movements and that they were arrived at independently by trial and error, but that it isn’t tied to any particular shape or type of object. Just like a human, Dactyl can grip and manipulate just about anything you put in its hand, within reason of course.

This flexibility is called generalization, and it’s important for robots that must interact with the real world. It’s impossible to hand-code separate behaviors for every object and situation in the world, but a robot that can adapt and fill in the gaps while relying on a set of core understandings can get by.

As with OpenAI’s other work, the paper describing the results is freely available, as are some of the tools they used to create and test Dactyl.

A meaty first report by the UK parliamentary committee that’s been running an inquiry into online disinformation since fall 2017, including scrutinizing how people’s personal information was harvested from social media services like Facebook and used for voter profiling and the targeting of campaign ads — and whose chair, Damian Collins — is a member of the UK’s governing Conservative Party, contains one curious omission.

Among the many issues the report raises are privacy concerns related to a campaign app developed by a company called uCampaign — which, much like the scandal-hit (and now seemingly defunct) Cambridge Analytica, worked for both the Ted Cruz for President and the Donald J Trump for President campaigns — although in its case it developed apps for campaigns to distribute to supporters to gamify digital campaigning via a tool which makes it easy for them to ‘socialize’ (i.e. share with contacts) campaign messaging and materials.

The committee makes a passing reference to uCampaign in a section of its report which deals with “data targeting” and the Cambridge Analytica Facebook scandal, specifically — where it writes [emphasis ours]:

There have been data privacy concerns raised about another campaign tool used, but not developed, by AIQ [Aggregate IQ: Aka, a Canadian data firm which worked for Cambridge Analytica and which remains under investigation by privacy watchdogs in the UK, Canada and British Columbia]. A company called uCampaign has a mobile App that employs gamification strategy to political campaigns. Users can win points for campaign activity, like sending text messages and emails to their contacts and friends. The App was used in Donald Trump’s presidential campaign, and by Vote Leave during the Brexit Referendum.

The developer of the uCampaign app, Vladyslav Seryakov, is an Eastern Ukrainian military veteran who trained in computer programming at two elite Soviet universities in the late 1980s. The main investor in uCampaign is the American hedge fund magnate Sean Fieler, who is a close associate of the billionaire backer of SCL and Cambridge Analytica, Robert Mercer. An article published by Business Insider on 7 November 2016 states: “If users download the App and agree to share their address books, including phone numbers and emails, the App then shoots the data [to] a third-party vendor, which looks for matches to existing voter file information that could give clues as to what may motivate that specific voter. Thomas Peters, whose company uCampaign created Trump’s app, said the App is “going absolutely granular”, and will—with permission—send different A/B tested messages to users’ contacts based on existing information.”

What’s curious is that Collins’ Conservative Party also has a campaign app built by — you guessed it! — uCampaign, which the party launched in September 2017.

While there is nothing on the iOS and Android app store listings for the Conservative Campaigner app to identify uCampaign as its developer, if you go directly to uCampaign’s website the company lists the UK Conservative Party as one of it’s clients — alongside other rightwing political parties and organizations such as the (pro-gun) National Rife Association; the (anti-abortion) SBA List; and indeed the UK’s Vote Leave (Brexit) campaign, (the latter) as the DCMS report highlights.

uCampaign’s involvement as the developer of the Conservative Campaigner app was also confirmed to us (in June) by the (now former) deputy director & head of digital strategy for The Conservative Party, Anthony Hind, who — according to his LinkedIn profile — also headed up the party’s online marketing, between mid 2015 and, well, the middle of this month.

But while, in his initial response to us, Hind readily confirmed he was personally involved in the procurement of uCampaign as the developer of the Conservative Campaigner app, he failed to respond to any of our subsequent questions — including when we raised specific concerns about the privacy policy that the app had been using, prior to May 23 (just before the EU’s new GDPR data protection framework came into force on May 25 — a time when many apps updated their privacy polices as a compliance precaution related to the new data protection standard).

Since May 23 the privacy policy for the Conservative Campaigner app has pointed to the Conservative Party’s own privacy policy. However prior to May 23 the privacy policy was a literal (branded) copy-paste of uCampaign’s own privacy policy. (We know because we were tipped to it by a source — and verified this for ourselves.)

Here’s a screengrab of the exchange we had with Hind over LinkedIn — including his sole reply:

What looks rather awkward for the Conservative Party — and indeed for Collins, as DCMS committee chair, given the valid “privacy concerns” his report has raised around the use (and misuse/abuse) of data for political targeting — is that uCampaign’s privacy policy has, shall we say, a verrrrry ‘liberal’ attitude to sharing the personal data of app users (and indeed of any of their contacts it would have been able to harvest from their devices).

Here’s a taster of the data-sharing permissions this U.S. company affords itself over its clients’ users’ data [emphasis ours] — according to its own privacy policy:

CAMPAIGNS YOU SUPPORT AND ALIGNED ORGANIZATIONS

We will share your Personal Information with third party campaigns selected by you via the Platform. In addition, we may share your Personal Information with other organizations, groups, causes, campaigns, political organizations, and our clients that we believe have similar viewpoints, principles or objectives as us.

UCAMPAIGN FRIENDS

We may share your Personal Information with other users of the Platform, for example if they connect their address book to our services, or if they invite you to use our services via the Platform.

BUSINESS TRANSFERS

We may share your Personal Information with other entities affiliated with us for internal reasons, primarily for business and operational purposes. uCampaign, or any of its assets, including the Platform, may be sold, or other transactions may occur in which your Personal Information is one of the business assets of the transaction. In such case, your Personal Information may be transferred.

To spell it out, the Conservative Party paid for a campaign app that could, according to the privacy policy it had in place prior to May 23, have shared supporters’ personal data with organizations that uCampaign’s owners — who the DCMS committee states have close links to “the billionaire backer of SCL and Cambridge Analytica, Robert Mercer” — view as ideologically affiliated with their objectives, whatsoever those entities might be.

Funnily enough, the Conservative Party appears to have tried to scrub out some of its own public links to uCampaign — such as changing link for the developer website on the app listing page for the Conservative Campaigner app to the Conservative Party’s own website (whereas before it linked through to uCampaign’s own website).

As the veteran UK satirical magazine Private Eye might well say — just fancy that! 

One of the listed “features” of the Conservative Campaigner app urges Tory supporters to: “Invite your friends to join you on the app!”. If any did, their friends’ data would have been sucked up by uCampaign too to further causes of its choosing.

The version of the Campaigner app listed on Google Play is reported to have 1,000+ installs (iOS does not offer any download ranges for apps) — which, while not in itself a very large number, could represent exponentially larger amounts of personal data should users’ contacts have been synced with the app where they would have been harvested by uCampaign.

We did flag the link between uCampaign and the Conservative Campaigner app directly to the DCMS committee’s press office — ahead of the publication of its report, on June 12, when we wrote:

The matter of concern here is that the Conservative party could itself be an unwitting a source of targeting data for rival political organizations, via an app that appears to offer almost no limits on what can be done with personal data.

Prior to the last update of the Conservative Campaigner app the privacy policy was simply the boilerplate uCampaign T&Cs — which allow the developer to share app users personal info (and phone book contacts) with “other organizations, groups, causes, campaigns, political organizations, and our clients that we believe have similar viewpoints, principles or objectives as us”.

That’s incredibly wide-ranging.

So every user’s phone book contacts (potentially hundreds of individuals per user) could have been passed to multiple unidentified organizations without people’s knowledge or consent. (Other uCampaign apps have been built for the NRA, and for anti-abortion organizations, for example.)

uCampaign‘s T&Cs are here: https://ucampaignapp.com/privacy.html

Even the current T&Cs allow for sharing with US suppliers.

Given the committee’s very public concerns about access to people’s data for political targeting purposes I am keen to know whether Mr Collins has any concerns about the use of uCampaign‘s app infrastructure by the Conservative party?

And also whether he is concerned about the lack of a robust data protection policy by his own party to ensure that valuable membership data is not simply passed around to unknown and unconnected entities — perhaps abroad, perhaps not — with zero regard for or accountability to the individuals in question.

Unfortunately this email (and a follow up) to the DCMS committee, asking for a response from Collins to our privacy concerns, went unanswered.

It’s also worth noting that the Conservative Party’s own privacy policy (which it’s now using for its Campaigner app) is pretty generous vis-a-vis the permissions it’s granting itself over sharing supporters’ data — including stating that it shares data with: 

• The wider Conservative Party
• Business associates and professional advisers
• Suppliers
• Service providers
• Financial organisations – such as credit card payment providers
• Political organisations
• Elected representatives
• Regulatory bodies
• Market researchers
• Healthcare and welfare organisations
• Law enforcement agencies

The UK’s data watchdog recently found fault with pretty much all of the UK political parties’ when it comes to handling of voter data — saying it had sent warning letters to 11 political parties and also issued notices compelling them to agree to audits of their data protection practices.

Safe to say, it’s not just private companies that have been sticking their hand in the personal data cookie jar in recent years — the political establishment is facing plenty of awkward questions as regulators unpick where and how data has been flowing.

This is also not the only awkward story re: data privacy concerns related to a Tory political app. Earlier this year the then-minister in charge of the digital brief, Matt Hancock, launched a self-promotional, self-branded app intended for his constituents to keep up with news about Matt Hancock MP.

However the developers of the app (Disciple Media) initially uploaded the wrong privacy policy — and were forced to issue an amended version which did not grant the minister such non-specific and oddly toned rights to users’ data — such as that the app “may disclose your personal information to the Publisher, the Publisher’s management company, agent, rights image company, the Publisher’s record label or publisher (as applicable) and any other third parties, for use in conjunction with additional user promotions or offers they may run from time to time or in relation to the sale of other goods and services”.

Of course the Matt Hancock App was a PR initiative of (and funded by) an individual Conservative MP — rather than a formal campaign tool paid for by the Conservative Party and intended for use by hundreds (or even thousands) of Party activists for use during election campaigns.

So while there are two issues of Tory-related privacy concern here, only one loops back to the Conservative Party political organization itself.

The bedroom community of Frisco, Texas might seem like an unusual place to find a self-driving vehicle. But here in this city of nearly 175,000 people, there are seven.

And as of Monday, they’re available for the public to use within a specific sector of the city that has a concentration of retail, entertainment venues and office space.

Drive.ai, an autonomous vehicle startup, launched the self-driving on-demand service Monday that will cover a two-mile route. The service will be operated in conjunction with Frisco TMA, a public-private partnership focused on “last-mile” transportation options. People within this geographic zone can hail a ride using a smartphone app.

Even in their small numbers, the modified Nissan NV200s will be hard to miss. The self-driving vehicles are painted a bright orange with two swooping blue lines — with the words “self-driving vehicle” and “Drive.ai” set in white.

The vehicles, which have been given distinctly human names like Anna, Emma, Bob, Fred and Carl, are equipped with LED screens on the hood and rear, and above the front tires, which will display messages as well as the vehicle’s name to pedestrians.

This isn’t a business enterprise just yet. The service, which is considered a pilot project, is free and will be operational for six months. The program will begin with fixed pick-up and drop-off locations around HALL Park and The Star and then will expand into Frisco Station.

Conway Chen, Drive.ai’s vice president of business strategy, emphasized to TechCrunch that this is designed as an on-demand service, and not a shuttle. When the vehicles are not being used they won’t just keep circling the route, which could cause more traffic congestion, Chen said. Instead they will be able to park along the route.

In the weeks since announcing plans to launch in Frisco, Drive.ai has been tweaking the service, its schedule as well as racking up miles on the road and in simulation. The company said it has logged 1 million simulated miles on its Frisco route. In its simulation, Drive.ai replicates scenarios — taken from its driving logs — the vehicles encountered while driving the route, as well as creating its own scenarios.

As Drive.ai explains in a post on Medium: “It’s like a high tech version of SimCity, where we design the world, and can then replay events and modify their components to explore how our technology responds in unique scenarios. This is a good place to start for the more common things that people do on the roads: navigating tricky intersections, right-of-way decisions, and observing the behaviors of cyclists and pedestrians.”

Drive.ai simulation.

The service, which will operate weekdays from 10 a.m. to 7 p.m., will initially have a safety driver behind the wheel. That person will eventually move to a passenger seat and take on a chaperone role, whose primary responsibility will be to answer questions and make riders comfortable. At some point, Drive.ai will remove the employee from the vehicle completely.

The company also has a remote monitoring feature, called “telechoice,” that allows a human operator to see everything in real-time that the self-driving vehicle can see using HD cameras.

Telechoice is not like the full remote control teleoperation that startup Phantom Auto provides. The telechoice operator can control basic functions like braking, but it cannot take full control of the vehicle or make it accelerate. With Drive.ai’s feature, if “Bob” the self-driving vehicle struggles with a certain situation on the road, the telechoice operator can help it make the right decision.

A dark theme option for YouTube users on Android is in the early stages of rolling out to end users, Google confirmed to TechCrunch, following a number of reports and sightings of the dark mode showing up for users in the app’s settings. The feature has taken a bit longer to launch than expected – YouTube first announced a dark mode for its mobile app back in March, when it launched on iOS. At the time, the company said the dark theme for Android was coming “soon.”

Five months later, well, here it is.

Similar to its iOS counterpart, the dark theme is toggled on or off in the Android app’s Settings. When enabled, YouTube’s usual white background switches to black throughout the YouTube app experience as your browse, search and watch videos.

The dark theme has a variety of benefits for end users. It gives watching videos a more cinematic feel, for starters. And when you’ve been staring at your screen for a long time, it can help you to better focus on the content, and not the controls. It can also help to cut down on glare, and help viewers take in the true colors of the videos they watch, the company previously explained.

Plus, some tests have shown dark themes can save battery life – something that’s particularly useful for YouTube’s 1.8 billion monthly users, who are spending more than an hour per day watching YouTube videos on mobile devices.

Above: Image credits, Imgur user absinth92

YouTube first introduced a dark theme in May 2017, when it debuted a series of enhancements to its desktop website, including its simpler, Material Design-inspired look. At the time, it said a dark theme for mobile was a top request.

The YouTube app isn’t alone in catering to users’ desire for a dark mode. Other high-profile apps have gone this route as well, including Twitter, Reddit, Twitter clients like Tweetbot and Twitterific, Reddit clients like Beam, Narwhal, and Apollo, podcast player Overcast, calendar app Fantastical, Telegram X, Instapaper, Pocket, Feedly and others.

Google told us that the dark theme for YouTube on Android is still in the early phases of a gradual rollout, and it will have more updates about this launch in the “coming weeks.”

The change arrives alongside update a YouTube Community Manager shared in YouTube’s Help Forum about YouTube’s adaptive video player. The player on desktop now removes the black bars alongside 4:3 and vertical videos, by adjusting the viewing area accordingly, they said.

Every artificial intelligence startup or corporate R&D lab has to reinvent the wheel when it comes to how humans annotate training data to teach algorithms what to look for. Whether it’s doctors assessing the size of cancer from a scan or drivers circling street signs in self-driving car footage, all this labeling has to happen somewhere. Often that means wasting six months and as much as a million dollars just developing a training data system. With nearly every type of business racing to adopt AI, that spend in cash and time adds up.

Labelbox builds artificial intelligence training data labeling software so nobody else has to. What Salesforce is to a sales team, Labelbox is to an AI engineering team. The software-as-a-service acts as the interface for human experts or crowdsourced labor to instruct computers how to spot relevant signals in data by themselves and continuously improve their algorithms’ accuracy.

Today, Labelbox is emerging from six months in stealth with a $3.9 million seed round led by Kleiner Perkins and joined by First Round and Google’s Gradient Ventures.

“There haven’t been seamless tools to allow AI teams to transfer institutional knowledge from their brains to software,” says co-founder Manu Sharma. “Now we have over 5,000 customers, and many big companies have replaced their own internal tools with Labelbox.”

Kleiner’s Ilya Fushman explains that “If you have these tools, you can ramp up to the AI curve much faster, allowing companies to realize the dream of AI.”

Inventing the best wheel

Sharma knew how annoying it was to try to forge training data systems from scratch because he’d seen it done before at Planet Labs, a satellite imaging startup. “One of the things that I observed was that Planet Labs has a superb AI team, but that team had been for over six months building labeling and training tools. Is this really how teams around the world are approaching building AI?,” he wondered.

Before that, he’d worked at DroneDeploy alongside Labelbox co-founder and CTO Daniel Rasmuson, who was leading the aerial data startup’s developer platform. “Many drone analytics companies that were also building AI were going through the same pain point,” Sharma tells me. In September, the two began to explore the idea and found that 20 other companies big and small were also burning talent and capital on the problem. “We thought we could make that much smarter so AI teams can focus on algorithms,” Sharma decided.

Labelbox’s team, with co-founders Ysiad Ferreiras (third from left), Manu Sharma (fourth from left), Brian Rieger (sixth from left) Daniel Rasmuson (seventh from left)

Labelbox launched its early alpha in January and saw swift pickup from the AI community that immediately asked for additional features. With time, the tool expanded with more and more ways to manually annotate data, from gradation levels like how sick a cow is for judging its milk production to matching systems like whether a dress fits a fashion brand’s aesthetic. Rigorous data science is applied to weed out discrepancies between reviewers’ decisions and identify edge cases that don’t fit the models.

“There are all these research studies about how to make training data” that Labelbox analyzes and applies, says co-founder and COO Ysiad Ferreiras, who’d led all of sales and revenue at fast-rising grassroots campaign texting startup Hustle. “We can let people tweak different settings so they can run their own machine learning program the way they want to, instead of being limited by what they can build really quickly.” When Norway mandated all citizens get colon cancer screenings, it had to build AI for recognizing polyps. Instead of spending half a year creating the training tool, they just signed up all the doctors on Labelbox.

Any organization can try Labelbox for free, and Ferreiras claims hundreds of thousands have. Once they hit a usage threshold, the startup works with them on appropriate SaaS pricing related to the revenue the client’s AI will generate. One called Lytx makes DriveCam, a system installed on half a million trucks with cameras that use AI to detect unsafe driver behavior so they can be coached to improve. Conde Nast is using Labelbox to match runway fashion to related items in their archive of content.

Eliminating redundancy, and jobs?

The big challenge is convincing companies that they’re better off leaving the training software to the experts instead of building it in-house where they’re intimately, though perhaps inefficiently, involved in every step of development. Some turn to crowdsourcing agencies like CrowdFlower, which has their own training data interface, but they only work with generalist labor, not the experts required for many fields. Labelbox wants to cooperate rather than compete here, serving as the management software that treats outsourcers as just another data input.

Long-term, the risk for Labelbox is that it’s arrived too early for the AI revolution. Most potential corporate customers are still in the R&D phase around AI, not at scaled deployment into real-world products. The big business isn’t selling the labeling software. That’s just the start. Labelbox wants to continuously manage the fine-tuning data to help optimize an algorithm through its entire life cycle. That requires AI being part of the actual engineering process. Right now it’s often stuck as an experiment in the lab. “We’re not concerned about our ability to build the tool to do that. Our concern is ‘will the industry get there fast enough?'” Ferreiras declares.

Their investor agrees. Last year’s big joke in venture capital was that suddenly you couldn’t hear a startup pitch without “AI” being referenced. “There was a big wave where everything was AI. I think at this point it’s almost a bit implied,” says Fushman. But it’s corporations that already have plenty of data, and plenty of human jobs to obfuscate, that are Labelbox’s opportunity. “The bigger question is ‘when does that [AI] reality reach consumers, not just from the Googles and Amazons of the world, but the mainstream corporations?'”

Labelbox is willing to wait it out, or better yet, accelerate that arrival — even if it means eliminating jobs. That’s because the team believes the benefits to humanity will outweigh the transition troubles.

“For a colonoscopy or mammogram, you only have a certain number of people in the world who can do that. That limits how many of those can be performed. In the future, that could only be limited by the computational power provided so it could be exponentially cheaper” says co-founder Brian Rieger. With Labelbox, tens of thousands of radiology exams can be quickly ingested to produce cancer-spotting algorithms that he says studies show can become more accurate than humans. Employment might get tougher to find, but hopefully life will get easier and cheaper too. Meanwhile, improving underwater pipeline inspections could protect the environment from its biggest threat: us.

“AI can solve such important problems in our society,” Sharma concludes. “We want to accelerate that by helping companies tell AI what to learn.”

Nobody really likes meetings — and the few people who do like them are the ones with whom you probably don’t want to have meetings. So when you’ve reached your fill and decide to reschedule some of those obligations, the usual process of trying to find a new meeting time begins. Thankfully, the Google Calendar team has heard your sighs of frustration and built a new tool that makes rescheduling meetings much easier.

Starting in two weeks, on August 13th, every guest will be able to propose a new meeting time and attach to that update a message to the organizer to explain themselves. The organizer can then review and accept or deny that new time slot. If the other guests have made their calendars public, the organizer can also see the other attendees’ availability in a new side-by-side view to find a new time.

What’s a bit odd here is that this is still mostly a manual feature. To find meeting slots to begin with, Google already employs some of its machine learning smarts to find the best times. This new feature doesn’t seem to employ the same algorithms to proposed dates and times for rescheduled meetings.

This new feature will work across G Suite domains and also with Microsoft Exchange. It’s worth noting, though, that this new option won’t be available for meetings with more than 200 attendees and all-day events.

The lack of diversity issue in Silicon Valley touches all aspects of the industry — entrepreneurship, big tech company demographics and venture capital. While tech companies have been a bit more deliberate about fostering diversity and inclusion in the last few years — and have made a small bit of progress — the venture capital industry is still sorely lacking in that area.

Just one percent of venture capitalists are Latinx and only three percent are black. White people, unsurprisingly, make up 70 percent of the venture capital industry, according to a recent analysis by Richard Kerby, a partner at Equal Ventures. Compared to Kerby’s 2016 analysis, women now make up 18 percent of the VC industry versus just 11 percent then. At an intersectional level, black and Latinx women make up zero percent of the venture capital industry.

Meanwhile, 40 percent of VCs went to either Stanford or Harvard.

“Just TWO schools!,” Kerby wrote on Medium. “Why is that? Everyone wants to work with those they are most similar to, and education, gender, and race are attributes that allow people to find similarities in others.”

He added, “With 82% of the industry being male, nearly 60% of the industry being white male, and 40% of the industry coming from just two academic institutions, it is no wonder that this industry feels so insular and less of meritocracy but more of a mirrortocracy.”

So, in addition to a lack of racial and gender diversity, there’s also a lack of cognitive diversity. Be sure to head over to Kerby’s Medium post to learn more.

3D body scanning systems have hit the big time after years of stops and starts. Hot on the heels of Original Stitch’s Bodygram, another 3D scanner, 3DLOOK, has entered into the fray with a $1 million investment to measure bodies around the world.

The founders, Vadim Rogovskiy, Ivan Makeev, and Alex Arapovd, created 3DLOOK when they found that they could measure a human body using just a smartphone. The team found that other solutions couldn’t let them measure fits with any precision and depended on expensive hardware.

“After more than six years of building companies in the ad tech industry I wanted to build something new which was not a commodity,” said Rogovskiy. “I wanted to overcome growth obstacles and I learned that the apparel industry had mounting return problems in e-commerce. 3DLOOK’s co-founders spent over a year on pure R&D and testing new approaches and combinations of different technologies before creating SAIA (Scanning Artificial Intelligence for Apparel) in 2016.”

The team raised $400,000 to date and most recently raised a $1 million seed round to grow the company.

The team also collects “fit profiles” and is able to supply these profiles based on “geographic location, age, and gender groups.” This means that 3DLOOK can give you exact sizes based on your scanned measurements and tell you how clothes will fit on your body. They have 20,000 profiles already and are working with eight paying customers and five large enterprise systems. Lemonade Fashion and Koviem are both using the platform.

“3DLOOK is the first company that managed to build a technology that allows capturing human body measurements with just two casual photos, and plans to disrupt the market of online apparel sales, offering brands and small stores an API for desktop and SDK for mobile to gather clients measurements and build custom clothing proposals,” said Rogovskiy. “Additionally, the company collects the database of human body measurements so that brands could build better clothing for all types of body and solve fit and return problems. It will not only allow stores to sell more apparel, it will allow people get the quality apparel.”

3D scanners have gotten better and better over the years and it’s interesting to see companies being able to scan bodies just from a few photos. While these things can’t account for opinions of taste they can definitely make sure that your clothes fit before you order them.

Let’s be honest. We all knew this was too good to last, right? For a few brief moments, Amazon’s Echo Spot was listed at $0.00. The smart device quickly went out of stock, only to have its price balloon back up to $129. Those who managed to pick one up crossed their fingers in hopes that the company would still make good.

Users (a few of our staff members included) are reporting, however, that the company has begun canceling orders placed within the window. It doesn’t appear to be a blanket cancellation at the moment, but that may well just be the time it takes to process what must have been hundreds of orders, at least. Interestingly, the white version of the device (the one that was momentarily free) is still listed as being “currently unavailable” on the site.

We’re still waiting to hear back from Amazon about what, precisely, went down here. We’ve nudged them again this morning, in hopes of getting more information on the cancellation — and whether the company will, at the very, offer up something as consolation to disappointed bargain hunters.