by tyler | Nov 16, 2023 | Pulse, Tech
Like “Avengers” director Joe Russo, I’m becoming increasingly convinced that fully AI-generated movies and TV shows will be possible within our lifetimes.
A host of AI unveilings over the past few months, in particular OpenAI’s ultra-realistic-sounding text-to-speech engine , have given glimpses into this brave new frontier. But Meta’s announcement today put our AI-generated content future into especially sharp relief — for me at least.
Meta his morning debuted Emu Video, an evolution of the tech giant’s image generation tool, Emu. Given a caption (e.g. “A dog running across a grassy knoll”), image or a photo paired with a description, Emu Video can generate a four-second-long animated clip.
Emu Video’s clips can be edited with a complementary AI model called Emu Edit, which was also announced today. Users can describe the modifications they want to make to Emu Edit in natural language — e.g. “the same clip, but in slow motion” — and see the changes reflected in a newly generated video.
Now, video generation tech isn’t new. Meta’s experimented with it before, as has Google. Meanwhile, startups like Runway are already building businesses on it.
But Emu Video’s 512×512, 16-frames-per-second clips are easily among the best I’ve seen in terms of their fidelity — to the point where my untrained eye has a tough time distinguishing them from the real thing.
Image Credits: Meta
Image Credits: Meta
Well — at least some of them. It seems Emu Video is most successful animating simple, mostly static scenes (e.g. waterfalls and timelapses of city skylines) that stray from photorealism — that is to say in styles like cubism, anime, “paper cut craft” and steampunk. One clip of the Eiffel Tower at dawn “as a painting,” with the tower reflected in the River Seine beneath it, reminded me of an e-card you might see on American Greetings .
Image Credits: Meta
Image Credits: Meta
Even in Emu Video’s best work, however, AI-generated weirdness manages to creep in — like bizarre physics (e.g. skateboards that move parallel to the ground) and freaky appendages (toes legs that curl behind feet and legs that blend into each other). Objects often appear and fade from view without much logic to it, too, like the birds overhead in the aforementioned Eiffel Tower clip.
After much too much time spent browsing Emu Video’s creations (or at least the examples that Meta cherry-picked), I started to notice another obvious tell: subjects in the clips don’t… well, do much. So far as I can tell, Emu Video doesn’t appear to have a strong grasp of action verbs, perhaps a limitation of the model’s underpinning architecture.
Image Credits: Meta
Image Credits: Meta
For example, a cute anthropomorphized racoon in an Emu Video clip will hold a guitar, but it won’t strum the guitar — even if the clip’s caption included the word “strum.” Or two unicorns will “play” chess, but only in the sense that they’ll sit inquisitively in front of a chessboard without moving the pieces.
Image Credits: Meta
Image Credits: Meta
So clearly there’s work to be done. Still, Emu Video’s more basic b-roll wouldn’t be out of place in a movie or TV show today, I’d say — and the ethical ramifications of this frankly terrify me.
Image Credits: Meta
Image Credits: Meta
The deepfakes risk aside, I fear for animators and artists whose livelihoods depend on crafting the sorts of scenes AI like Emu Video can now approximate. Meta and its generative AI rivals would likely argue that Emu Video, which Meta CEO Mark Zuckerberg says is being integrated into Facebook and Instagram (hopefully with better toxicity filters than Meta’s AI-generated stickers), augment rather than replace human artists. But I’d say that’s taking the optimistic, if not disingenuous, view — especially where money’s involved.
Earlier this year, Netflix used AI-generated background images in a three-minute animated short. The company claimed that the tech could help with anime’s supposed labor shortage — but conveniently glossed over how low pay and often strenuous working conditions are pushing away artists from the work.
In a similar controversy, the studio behind the credit sequence for Marvel’s “Secret Invasion” admitted to using AI, mainly the text-to-image tool Midjourney , to generate much of the sequence’s artwork. Series director Ali Selim made the case that the use of AI fits with the paranoid themes of the show, but the bulk of the artist community and fans vehemently disagreed .
Image Credits: Meta
Image Credits: Meta
Actors could be on the chopping block, too. One of the major sticking points in the recent SAG-AFTRA strike was the use of AI to create digital likenesses. Studios ultimately agreed to pay actors for their AI-generated likenesses. But might they reconsider as the tech improves? I think it’s likely.
Adding insult to injury, AI like Emu Video is usually trained on images and videos produced by artists, photographers and filmmakers — and without notifying or compensating those creators. In a whitepaper accompanying the release of Emu Video, Meta says only that the model was trained on a data set of 34 million “video-text pairs” ranging in length from 5 to 60 seconds — not where those videos came from, their copyright statuses or whether Meta licensed them.
Image Credits: Meta
Image Credits: Meta
There’s been fits and starts toward industry-wide standards to allow artists to “opt out” of training or receive payment for AI-generated works to which they contributed. But if Emu Video is any indication, the tech — as so often happens — will soon run far ahead of ethics. Perhaps it already has.
by tyler | Nov 8, 2023 | Pulse, Tech
GM self-driving car subsidiary Cruise issued a recall for 950 vehicles equipped with its autonomous vehicle software following a crash that left a pedestrian, who had initially been hit by a human-driven car, stuck under and then dragged by one of the company’s robotaxis.
The company said in a blog post and in the recall notice filed with the National Highway Traffic and Safety Administration that it issued the recall after an analysis of the robotaxi’s response October 2 found the “collision detection subsystem may cause the Cruise AV to attempt to pull over out of traffic instead of remaining stationary when a pullover is not the desired post-collision response.”
In that October incident, a pedestrian was struck by a human driver and then landed in the adjacent lane where a Cruise robotaxi was driving. The robotaxi initiated its brakes and came to a stop with the pedestrian under the vehicle. The robotaxi then attempted to pull over, dragging the woman some 20 feet.
“Although we determined that a similar collision with a risk of serious injury could have recurred every 10 million – 100 million miles of driving on average prior to the software update, we strive to continually improve and to make these events even rarer. As our software improves, it is likely we will file additional recalls to inform both NHTSA and the public of updates to enhance safety across our fleet,” the company wrote in the blog post.
Cruise and its parent company GM had been under increased scrutiny for weeks following several incidents, including a collision with an emergency response vehicle. The opposition turned into regulatory action, however, following the October 2 event.
The California Department of Motor Vehicles and the state’s Public Utilities Commission pulled all permits that allowed Cruise to commercially operate a fleet of robotaxis on public roads in San Francisco. Two days later, the company paused all driverless testing and operations across its fleet, which included Austin, Houston and Phoenix. It also halted production of its custom-built Cruise Origin vehicles.
The company has been in crisis mode ever since, particularly as new reports have emerged that expose the company’s safety practices and potential flaws in its software. Cruise is also facing federal investigations into how autonomous vehicles interact with pedestrians.
Cruise said in its blog post, confirming earlier reports, it is conducting a search to hire a chief safety officer who will report directly to the CEO, hired Exponent to conduct a technical review and retained law firm Quinn Emanuel to examine Cruise’s response to the October 2 incident, including the company’s interactions with law enforcement, regulators and the media.
by tyler | Nov 6, 2023 | Pulse, Tech
YouTube will begin to experiment with new generative AI features, the company announced today. As part of the premium package available to paid subscribers, YouTube users will be able to try out a new conversational tool that uses AI to answer questions about YouTube’s content and make recommendations, as well as try out a new feature that will summarize topics in the comments of a video.
Both features will become available through YouTube’s hub for new experiments, youtube.com/new, on an opt-in basis. The conversational tool will arrive in the next few weeks while the topic summarization tool will begin testing only with a small group of users who sign up for the experiment through the website.
Explains YouTube, the conventional tool’s responses will be generated by large language models that draw on information from YouTube and the web and are designed to help viewers “dive in deeper” to the content they’re watching. For example, participants will be able to ask questions about the video they’re currently viewing or ask for other things to watch similar to the video using actions like “recommend related content.” These conversations with the AI will be able to take place as the video is streaming, so as not to interrupt playback.
Image Credits: YouTube
Image Credits: YouTube
The feature will be offered initially only in the U.S. and on Android devices via a new “Ask” option on the video’s watch page. YouTube Premium members can opt into this test starting today, but it won’t appear on every YouTube video as of yet. The tool is currently being tested with a select group of users, but will be available as an opt-in for YouTube Premium members in the weeks to come.
The comments summarizer, meanwhile, will leverage generative AI to organize and summarize the topics of discussion in large comment sections underneath a video to allow viewers to catch up on what people have been saying. YouTube believes this feature may also be useful to creators to help them learn what their channel’s subscribers are talking about without having to read through all the comments one by one. This could also help them gain inspiration for new videos based on comment topics, YouTube suggests. Plus, if creators want to remove any comment topics, they’ll be able to delete any comments that appear under certain topics.
Image Credits: YouTube
Image Credits: YouTube
The topics will be summarized only based on published comments, not those held for review or those that contain blocked words or are from blocked users, YouTube also notes.
This experiment, which is now running on a small number of videos, will only be available in English for the time being.
YouTube says the tests will continue over the next several weeks and months as the company gathers feedback on how consumers are using the new features.
Google has been experimenting with AI across its product line, from Search to Workspace to consumer apps like Photos and Assistant and more. On YouTube specifically, the company has also experimented with AI-powered ads that let brands target cultural moments, and has been rolling out a number of creator-focused AI features, including tools to find music for videos , AI-powered dubbing , plus “ Dream Screen,” a generative AI feature that lets you create a background image or video for YouTube Shorts. YouTube Music also recently introduced AI-generated playlists. And YouTube had previously experimented with AI-generated video summaries and quizzes on educational videos .
“These features are experimental and we may not always get it right. That’s why we’re starting small with limited availability and collecting feedback,” said YouTube in its announcement. The company said users will be able to submit feedback directly in the tools themselves.
by tyler | Nov 3, 2023 | Pulse
Consumers are hungry for a new way of social networking, where trust and safety are paramount and power isn’t centralized with a Big Tech CEO in charge… or at least that’s what Mozilla believes. The mission-driven tech company behind the Firefox browser, Pocket reader and other apps is now investing its energy into the so-called “fediverse” — a collection of decentralized social networking applications, like Mastodon, that communicate with one another over the ActivityPub protocol.
The idea is to rethink social networking from the ground up.
Today, social networks are often run by large corporations — like Meta, Snap and Google — where advertisers pay the bills. This has created a world where modern social networks are profit-driven, and consumer needs aren’t always at the forefront, Mozilla believes. This disconnect between what people want and what today’s networks offer hit an inflection point last year, when billionaire Elon Musk bought Twitter, leading to a surge of interest in alternatives to centralized social networking, including apps like Mastodon and Bluesky .
Under Musk, Twitter — now called X — has prioritized “free speech” managed by crowdsourced-based moderation. Its policies have resulted in increased hate speech and toxic content on the platform, some studies have shown .
Mozilla believes there’s a better way forward and, more importantly, that now is exactly the right time to invest in that path. And, as a wholly owned subsidiary of a nonprofit, the company says it’s not motivated by generating earnings for shareholders or returning a VC investment, allowing it to progress with a collaborative approach where it takes in input from a lot of different voices.
In an interview with TechCrunch, Mozilla Senior Director of Content Carolyn O’Hara explained why Mozilla has taken an interest in the fediverse and Mastodon, specifically, and how it expects to experiment in the months ahead.
For background, the company announced its plans to enter the fediverse in December 2022 and this May launched a private beta for Mozilla.social, a Mastodon server — or “instance” in fediverse parlance — that will allow consumers to participate in this new form of social networking where problems like misinformation, harassment and issues around worsening mental health are minimized.
The company’s work on its social efforts is also available publicly on GitHub. (Of interest to technical folks, the company is using a forked version of Elk as its alternative web client for its Mastodon server, which is what gives it its slick look and feel.)
Image Credits: Mozilla (opens in a new window)
Mozilla’s involvement, says O’Hara, came about because the company looked at the history of social media over the past decade and didn’t like what it saw.
“I think that it’s a pretty poor track record by existing companies that are only model motivated by profit and just insane user growth, and are willing to tolerate and amplify really toxic content because it looks like engagement,” she says. “[They] aren’t just putting forward the kind of standards that are good for people, but are just good for their bottom lines.”
Plus, she adds, consumers are now becoming aware of this, which is new.
“Consumers feel that the vibes are off a little bit…these platforms aren’t necessarily working in their best interests, or satisfying them,” O’Hara points out.
For Mozilla, that discontent presented an opportunity to enter the social networking space and compete for consumers’ attention. In its case, however, the broader goal is to help the fediverse itself gain traction, not just its own Mastodon server.
“We’re not seeking to get, like 2 billion users on our instance. We want people to have choice and agency,” says O’Hara, adding that choice is in line with Mozilla’s values.
However, the company is aiming to tackle some of the obstacles that have prevented users from joining and participating in the fediverse so far, including the technical hurdles around onboarding, finding people to follow and discovering interesting content to discuss.
For starters, users will join the Mozilla.social instance with their Mozilla accounts , which also provides access to the Firefox browser, Mozilla’s VPN, Pocket and other products, simplifying access to its suite of tools. The instance isn’t yet open to the public, but is experimenting with hundreds of users for the time being.
“We’re keeping the number intentionally small and we’re prioritizing the groups that we’re proactively reaching out to,” O’Hara explained. “The communities we’ve worked with previously will have first access to the private beta, some of which will include folks from the waitlist.”
The timeline to a public launch isn’t yet determined, as the company plans to experiment with various expansions and features throughout next year.
For example, Mozilla is currently experimenting with a Discover feed that aims to surface engaging content. Over time, it plans to gather more signals from around the fediverse to determine what sort of content people are interacting with. It also sees a role for its read-it-later app Pocket in this experience, as the app’s primary use case is to store links to articles and other content that people have signaled they want to read.
“We are planning to open the beta in phases because we want to ensure that we have a content moderation team and other tools that can be scaled relative to the size of the user base. We’re not in a rush as we think it’s important to be deliberate about launching social media tools,” O’Hara says.
Image Credits: Mozilla
Image Credits: Mozilla
“We have a lot of core competency that we brought over from our Pocket experience and other recommendation products that we put together to at least get that engine revving,” notes O’Hara.
Initially, the company is running a Mastodon account, “ @Pocket_Recommends@mozilla.social ,” where it actively posts recommendations of stories to read. It’s the same sort of content you might find in the Pocket email newsletter, but now distributed on the fediverse.
The company also wants to address the needs of creators and publishers who want to broaden their audience as it moves forward with its fediverse plans.
“From a content discovery standpoint, I’m really interested in how we can seed conversations and seed experiences with really high-quality content — certainly, editorial publisher content,” says O’Hara. “How can we get publishers and content providers involved early?… I think of them as a real constituency for us,” she says. Plus, Mozilla wants to bring in the originators whose posts lead to interesting conversations and keep the network active.
O’Hara says Mozilla is now in active discussions with publishers to understand their needs, including both their social needs and business needs, and how those goals may have changed over the past year. As part of these discussions, Mozilla aims to convince publishers that Mastodon isn’t just another place they have to support, but one that could deliver on their objectives. As it learns from the publishers and other content creators what they want and need, it then intends to use that understanding to build features and drive efforts that can help them reach their audiences.
So far, what Mozilla has learned is that publishers today feel somewhat resigned and pessimistic about social networks. Twitter isn’t driving much traffic and, even though Instagram Threads appears interesting, Meta specifically said it wasn’t going to prioritize news on that platform . What Mozilla wants to accomplish, then, is to help reconfigure the Mastodon onboarding process so that when someone — including a publisher or creator — joins its instance (or the fediverse in general) they’re able to build their audience with more ease.
“The baseline is just better onboarding, better connecting with accounts and communities,” explains O’Hara. “You might be interested in spaces where you can have the kinds of conversations that you want to be able to have; trust and safety practices that make you feel safe in order to participate. And I think also then access to content that spurs those conversations and seeds those conversations. And it all being kind of a delightful user experience — but we’re at the early days of that,” she adds.
Image Credits: Mozilla (opens in a new window)
Another avenue of thought is to create spaces within the fediverse where consumers can have civil and direct conversations within smaller communities, instead of shouting into the void, so to speak. As to what that will look like, is less clear. Will it be different instances or just new ways of forming communities, by building out features or experiences? O’Hara couldn’t say, only noting that experimentation will be needed.
Importantly, trust and safety will also be key to Mozilla’s instance. Its content policies indicate strict measures around hate speech, impersonation, self-harm, harassment, misinformation, violent and sexualized content, and more, in addition to illegal content, like CSAM (child sexual abuse material) and promotion of illegal goods.
But there’s a chance that playing it too safe can make a product less compelling as a Twitter/X alternative. That’s one takeaway, at least, from the recent shutdown of Pebble (formerly T2) , a Twitter clone that put trust and safety at the forefront of its development process. The app never gained more than 20,000 users, and the founders now believe that its approach to moderation may have been correct, but it wasn’t a growth driver. In other words, people may say they want to feel safer when participating online, but clearly, that isn’t their only need.
Mozilla’s instance, still in private testing, has a small team of moderators that will grow when the instance launches to the public in the coming months.
“It’s not just a commitment, it’s beyond foundational for us — that is the thing that is going to be true about this instance,” O’Hara says of Mozilla’s plans to heavily moderate its space. Though Mozilla will also rely on technology to do some of the work, it wants to make sure the human moderators, “feel protected, supported and respected for their work,” she adds.
“We are saying out the gate that this isn’t a neutral platform,” O’Hara stresses. “We think that that’s often used as this crutch to allow, or even amplify, really toxic content in the name of engagement… In some cases, platforms aren’t just brave enough to actually just take things down.”
Building directly on the fediverse isn’t the only way the company is prompting new forms of social networking. The company has also financially backed a startup, Mammoth , the makers of a third-party app for Mastodon.
Mozilla has been leaning into its values around inclusion, dignity, security, agency and community for 25 years, O’Hara says, and those will guide its efforts in the fediverse, as well.
“We’re going to have rules on how people can engage with one another…it’s the fediverse, you have lots of other places to go if you don’t want to abide by those rules,” she adds.
Users can provide Mozilla with feedback via @Social @Mozilla.social or by using the hashtag #mozillasocial within the product.
Mozilla expands its Mastodon investment with beta launch of its own highly moderated server
Mozilla leads Mastodon app Mammoth’s pre-seed funding
by tyler | Nov 1, 2023 | Pulse
HubSpot , the Boston-based marketing software maker and CRM platform, announced today it’s acquiring the B2B data provider Clearbit , to enhance its platform with third-party company data spanning millions of businesses. The deal also brings Clearbit’s over 400,0000 users and 1500-plus business customers to HubSpot, and will eventually see the two platforms combine in order to provide HubSpot’s customer base with expanded data plus actionable insights.
Founded in 2015, Clearbit began as a tool that would help users hunt down email addresses associated with a company, as well as employee information like their name, job title, and other details, like social media accounts. The idea was that this information would help companies better understand their leads and vet their sales. SV Angel and First Round Capital were among the startup’s first backers.
I want to say a big public thanks to the @clearbit team, and especially to my two co-founders Harlow Ward and @mattsornson . With their oversight the company has shipped some massive improvements recently.
— Alex MacCaw (@maccaw) November 1, 2023
Over the years that followed, Clearbit developed into a fuller suite of tools, including those targeted towards sales, marketing, and ops teams, which included integrations with CRM providers like HubSpot and Salesforce. It also offered technology to enrich a company’s leads, contacts, and accounts with additional data, including public data from the web — like company websites and crowdsourced data — as well as its own proprietary data. More recently, it began to leverage LLM (large language model) technologies to convert unstructured data into standardized data sets for B2B teams.
As Clearbit co-founder and CEO Matt Sornson explained earlier this year , “Large Language Models and Generative AI are the step changes in technology that we are using to deliver on our promise to customers. LLMs will disrupt entire industries over the coming years, but I am confident that they will completely disrupt the data industry within 18 months,” he wrote in a blog post, where he laso announced the Clearbit data pipeline had been rebuilt with LLMs at the core.
Thanks to this technology, he said, Clearbit was now able to identify and enrich any company or contact data from any country in any language. That also likely made it a more interesting acquisition target.
The company’s website touted Clearbit’s relationships with not only its acquirer HubSpot, but also other businesses like Twilio Segment, Asana, Intercom, Zenefits, Greenhouse, Chargebee, Lattice, and Frame.io. With HubSpot, specifically, Clearbit has been available to its customers as part of the HubSpot App Marketplace since 20119.
Following the deal’s close, Clearbit will become a subsidiary of HubSpot and will eventually become integrated into its customer platform.
Speaking to the reasons why it was interested in Clearbit, HubSpot explained that gathering company data has gotten easier over the years, but challenges still remain around analyzing and using that data. It believed that by combining Clearbit’s data with HubSpot’s platform, companies would be able to enrich their internal customer data with more real-time external context, explained Yamini Rangan, CEO of HubSpot, in a statement .
“Clearbit has made it its mission to collect rich and useful data about millions of companies. HubSpot’s AI-powered customer platform combined with Clearbit’s data will create a powerful, winning combination for our customers,” she noted.
Clearbit had raised $17 million to date, according to data from Pitchbook , which saw the startup valued at $250 million as of January 2019. Its Series A investors included Bedrock, Battery, Cross Creek, and Zetta Venture Partners.
Deal terms were not immediately available.
“Clearbit has always believed that data is fundamental to the best B2B go-to-market teams,” noted Sornson, in a statement. “By joining forces with HubSpot, the industry’s most loved B2B customer platform, we will unlock a whole new level of value for our customers and help all of B2B grow better.”
HubSpot unveils strategy to integrate AI across the platform
Know more about this deal? Sarah Perez can be reached at sarahp@techcrunch.com or Signal (415) 234-3994
by tyler | Nov 1, 2023 | Pulse, Tech
Sandwiched between rows of warehouses in the South Boston Waterfront, MassRobotics serves as both an incubator and import hub for the city’s vibrant robotics community. Soon, it will add accelerator to that list. This morning the nonprofit announced the forthcoming arrival of its simply named MassRobotics Accelerator.
Applications are now open for the 13-week program, which focuses on early-stage robotics startups — particularly those operating in hardware, components and applications. Companies accepted into the incubator will received $100,000 in non-dilutive funding. The funding comes by way of MassTech, an organization developed to foster technology in the state.
Want the top robotics news in your inbox each week? Sign up for Actuator here .
Ten companies will be included in the round. MassRobotics points out that that it won’t be getting an equity stake in the companies. “We feel it’s the first equity-free robotics accelerator in the U.S.,” the accelerator’s director, Marita McGinn, tells TechCrunch.
Teams will have access to MassRobotics’ facilities, which include tools like CNC machines, laser and waterjet cutters and 3D printers for hardware prototyping. The site also features Industrial Automation and Human Robot Interaction (HRI) labs, and a slew of different robotic hardware platforms. Human mentors will also be available, for both the technical and business sides.
Image Credits: MassRobotics
Image Credits: MassRobotics
The application process “soft launched” October 16. It closes at the end of November. The four-month program kicks off February 5, culminating in a May 2 demo day, held as part of Boston’s Robotics Summit & Expo.
McGinn says there’s a “strong preference” for companies already based in Massachusetts, but the program doesn’t have any sort of quota to that effect.
“We’re really looking for the most interesting companies,” she says. “One of the things we really worked on with MassTech is how to attract companies to come to Massachusetts. In the end, I think it will be a nice collection of Mass-based companies and maybe companies from other parts of the country or international. Companies that are out of state need to have some sort of commitment to Massachusetts, one year post-accelerator.”
McGinn adds that diversity — something too often lacking in the industry — will also be a focal point for applications. “We will absolutely make sure that it looks and feels diverse from every standpoint.”
