Connect with us

Mobile

At last, a camera app that automatically removes all people from your photos – TechCrunch

Published

on

As a misanthrope living in a vibrant city, I’m never short of things to complain about. And in particular the problem of people crowding into my photos, whatever I happen to shoot, is a persistent one. That won’t be an issue any more with Bye Bye Camera, an app that simply removes any humans from photos you take. Finally!

It’s an art project, though a practical one (art can be practical!), by Do Something Good. The collective, in particular the artist damjanski, has worked on a variety of playful takes on the digital era, such as a CAPTCHA that excludes humans, and setting up a dialogue between two Google conversational agents.

The new app, damjanski told Artnome, is “an app for the post-human era… The app takes out the vanity of any selfie and also the person.” Fortunately, it leaves dogs intact.

Of course it’s all done in a self-conscious, arty way — are humans necessary? What defines one? What will the world be like without us? You can ponder those questions or not; fortunately, the app doesn’t require it of you.

Bye Bye Camera works using some of the AI tools that are already out there for the taking in the world of research. It uses YOLO (You Only Look Once), a very efficient object classifier that can quickly denote the outline of a person, and then a separate tool that performs what Adobe has called “context-aware fill.” Between the two of them a person is reliably — if a bit crudely — deleted from any picture you take and credibly filled in by background.

It’s a fun project (though the results are a mixed bag) and it speaks not only to the issues it supposedly raises about the nature of humanity, but also the accessibility of tools under the broad category of “AI” and what they can and should be used for.

You can download Bye Bye Camera for $3 on the iOS App Store.

Source link

Source link

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Mobile

Netflix crackdown, monetizing ChatGPT and bypassing FB’s 2FA • TechCrunch

Published

on

Happy weekend, folks, and welcome back to the TechCrunch Week in Review. Henry here, standing in for a vacationing Kyle Wiggers, who is standing in for a parental-leaving Greg Kumparak. Listen, we’ve got a deep bench, and both blokes will be back very soon. Until then, check out just a few of the top stories from the week.

Want it in your inbox every Saturday AM? You can take care of that right here.

most read

Netflix’s password-sharing crackdown: The streaming giant has grown tired of its customers sharing passwords with friends and loved ones around the world. So this week it announced guidelines designed to keep the passwords close to home. Literally inside the walls of the abode of the account holder.

Monetized ChatGPT: OpenAI this week launched a pilot subscription for its text-generating AI. For $20 a month, subscribers can access more than what the base level gets: access to ChatGPT during peak hours, faster response times and priority access to new features and improvements.

Human or AI?: That is the question, and apparently OpenAI wants to help. The company launched a tool that is designed to distinguish between human-written and AI-generated text, but the success rate is only around 26%. OpenAI did say, though, that when used with other methods, it could help prevent AI text generators from being abused.

Bypassing FB 2FA: Meta created a new centralized system so users could manage their logins for Facebook and Instagram, but a bug could have allowed malicious hackers to switch off 2FA just by knowing a user’s phone number. Yikes. A security researcher from Nepal discovered the bug and reported it to Meta Accounts Center last September. And he got paid.

Salesforce layoffs hit: In January, the company announced the imminent reduction of 10% of its workforce. Not everyone was notified at the time, however. This week, hundreds more of the company’s staff found out the fate of their jobs.

“Spill the tea”: Alphonzo “Phonz” Terrell lost his job at Twitter as its global head of Social & Editorial three months ago and promptly got to work on a new app. Called Spill, the app has already attracted a seed round and 60,000 handle reservations. The app is due to launch in alpha during the first quarter of this year.

Google Fi breach: The company said its cell network provider, Google Fi, confirmed a data breach, which, based on the timing of the notice, was likely related to the recent security incident at T-Mobile that allowed hackers to steal millions of customers’ information.

audio roundup

This week out of the TechCrunch Podcast Network, Equity covered the usual slate of venture and startup funding news, and Mary Ann spoke with Hans Tung, investor and managing partner of GGVC, a venture firm with more than $9 billion in assets under management. On Found, Darrell and Becca talked to Rosie Nguyen, a co-founder and the CMO of Fanhouse, about her journey from content creator to founder and how her experience as a creator informs every product decision at Fanhouse.

TechCrunch+

TC+ subscribers get access to in-depth commentary, analysis and surveys — which you know if you’re already a subscriber. If you’re not, consider signing up. I doubt you’ll regret it. Just check out the highlights from this week:

Not quite secondarily: Becca reports on data this week that shows secondary deals are breaking away from the downturned venture market this year.

Open source startups: Paul Sawers examines a report out this week that explores which commercial open source software startups are growing fast and raising cash.

Go team: Ever wonder which slide is the most important slide in a startup’s pitch deck? Why, it’s the team slide and Haje expresses his surprise at just how many startups fail to tell a good story about their teams. And speaking of pitch decks, Haje brings Laoshi’s $570K angel deck breakdown to you.

Dear Sophie: Immigration Sophie Alcorn answers the question, What H-1B and other immigration changes can we expect this year?

Continue Reading

Mobile

Amazon ramped up content spending to $16.6B in 2022, including $7B on originals • TechCrunch

Published

on

Amazon detailed the costs of its content business during its fourth-quarter earnings on Thursday, citing that its content expenses jumped to $16.6 billion in 2022, a 28% increase from $13 billion in 2021.

According to Chief Financial Officer Brian Olsavsky, around $7 billion of that figure went towards Amazon Originals, live sports programming, and licensed third-party video content included with Prime. In 2021, Amazon had spent $5 billion on those three areas of content, for comparison.

While the company didn’t break down exactly how much it invests in each title, it’s reported that Amazon is spending more than $1 billion annually for its NFL streaming rights. Plus, the first season of “The Lord of the Rings: Rings of Power,” the most-watched Amazon original series worldwide, cost over $500 million.

Streaming services know by now that original content is the key to standing out amongst rivals and reducing churn. Amazon is likely boosting its content investments to better compete with Disney, Netflix, and HBO Max. Disney spends approximately $33 billion on content, while both Netflix and HBO Max spend a reported $18 billion. *Note that a portion of Disney’s figure goes towards sports rights—around $11 billion.) Paramount+ also plans to increase streaming content spending to $6 billion by 2024, it recently said.

Amazon didn’t report subscriber numbers for its streaming business. However, Olsavsky boasted during the earnings call that its Prime Video content is a “strong driver of Prime member engagement and new Prime member acquisition,” Olsavsky said.

For instance, “The Rings of Power” was viewed by over 100 million global viewers with more than 24 billion minutes streamed. The company added that, during its launch window, “The Lord of the Rings” series helped drive more Prime sign-ups worldwide than any previous Prime Video content.

Amazon also touted that Thursday Night Football reached the youngest median age audience of any NFL broadcast package since 2013, and viewership among fans ages 18 to 34 years old increased by 11% compared to the 2021 season.

The company claimed the TNF games had an average audience of 11.3 million viewers. The first exclusive TNF game on Prime Video had 15.3 million viewers. Before the 2022 season began, Amazon expected to reach about 12.5 million viewers per week.

Other original content added to the streamer in 2022 includes “My Policeman” starring Harry Styles, the third season of “Jack Ryan” and the Western drama “The English,” among others.

Amazon is also benefiting from its 2022 acquisition of MGM for $8.5 billion. The company noted that “Wednesday,” the MGM-produced series on Netflix, premiered at No. 1 on Nielsen’s weekly streaming charts and earned two Golden Globe nominations. In December, “Wednesday” became the second most popular English-language series on Netflix, surpassing 1.02 billion total hours viewed in just three weeks since its streaming release. Over 150 million households watched the show.

Prime members in the U.S. also saw the return of HBO Max as a Prime Video Channel offering, giving customers access to approximately 15,000 hours of premium content.

Continue Reading

Mobile

‘Nothing, Forever,’ an AI ‘Seinfeld’ spoof, is the next ‘Twitch Plays Pokémon’ • TechCrunch

Published

on

“So, I was at the store the other day, and as I’m checking out, the cashier asks me if I have any coupons, and I say, ‘No coupon problem!’” recalls a pixelated, barely three-dimensional figure that vaguely resembles Jerry Seinfeld. “So I’m walking down the street, and this guy comes up to me and says, ‘Hey, how’s it going?’ and I say, ‘It’s going coupon!’”

An automated laugh track plays, but the joke doesn’t make sense. Then again, it doesn’t have to make sense.

“Nothing, Forever” is a never-ending, AI-generated spoof of “Seinfeld,” the show about nothing. It’s been streaming on Twitch since December, and until a few days ago, the stream had an average of about four concurrent viewers. Now, at the moment I write this, there are 15,097 people watching a group of badly animated friends — Larry Feinberg, Fred Kastopolous, Yvonne Torres and Zoltan Kalker — cycling through infinite “Seinfeld”-like scenes with very little plot.

The show has been streaming almost non-stop on Twitch since December, but it only reached a wider audience this week, when its creators slowly started promoting the stream on Reddit. Now, “Nothing, Forever” has over 98,000 followers on Twitch, and a Discord with about 6,000 members.

Behind the project are Skyler Hartle, a senior product manager at Microsoft, and Brian Habersberger, a polymer physicist. They call themselves Mismatch Media, though this venture remains a side project.

Aptly, the duo met online while playing “Team Fortress 2” and they kept in touch over time. Four years ago, they started working on creative projects together.

Image Credits: Nothing, Forever (opens in a new window)

“It kind of started its journey as a kind of art project that Brian approached me with, and we ended up collaborating and working on it together and iterating on it over the last four years,” Hartle told TechCrunch. “The show we’re creating is really cool, and scratches that creative itch as just a surreal, fun kind of project, but we saw the merit of generative technology as a tool for broad scale content creation and generation.”

To make “Nothing, Forever,” Hartle and Habersberger use various AI models to generate text, speech, and movements. The “script” of the show comes from an Open AI’s GPT-3 model, Davinci. To voice the characters, they use the Microsoft Azure Cognitive Services speech API, and the visuals are made on the Unity game engine.

“The Unity engine just does a lot of interpretation to basically run the show and inherit all this content, and the voices, and all these kinds of other pieces of direction from what we call ‘the director’ in the cloud,” Hartle said. “And the director dictates what happens on the show from a generative perspective.”

They set out to create a surrealist, never-ending television show, and it simply made sense to base it on “Seinfeld,” a show that has defined the structure of a sitcom.

“A sitcom has a laugh track and a sort of formulaic structure,” Habersberger told TechCrunch. “So when characters are saying things that don’t quite make sense, but the structure is one that you’re very familiar with, it really helps you to interpret and make sense of it, even though the sense isn’t there.”

AI dialogue can get repetitive. Characters are constantly referencing new restaurants and stores to the point that it’s become an in-joke. On a fan-made “Nothing, Forever” bingo generator, the free space is “New thing!”

Per the nascent wiki, some new places include a new type of bagel (it’s shaped like an octopus, and called the octobagel), a new shake shop (they serve pickles in their shakes), a new taco truck (they sell tacos and burgers) and a new all-you-can-eat buffet (nothing but pink flamingo wings!). One of the few interactive items in Larry’s apartment is a microwave, a prop that characters frequently and inexplicably use in ways that have no bearing on the plot — the microwave has spurred a fandom of its own, custom Discord emojis and all.

Image Credits: Nothing, Forever bingo (opens in a new window)

“To take that idea even further, ‘Seinfeld’ is famously the show about nothing. What could be more nothing than a robot, right?” said Habersberger. “And then even further with ‘Seinfeld,’ there was a period of fifteen years or so where you could just turn on the TV, and if you flip through the channels, there’s a good chance it’s on, because it was syndicated […] So I was like, wait, now it can really be always on.”

We’ve all seen far too many AI-generated gimmicks, but the AI isn’t what’s most interesting about “Nothing, Forever.” It’s the community that’s gathered around the stream, making the project feel like this generation’s “Twitch Plays Pokémon.”

Image Credits: Nothing, Forever (opens in a new window)

In 2014, an anonymous Australian streamer set up a channel where fans could collaboratively play “Pokémon Red,” pressing buttons and moving the player character via Twitch chat commands. When enough players got involved, the stream turned into chaos. (The route through Rock Tunnel is confusing enough, but imagine navigating it with thousands of people clamoring to control the character’s movement.)

Yet over the two weeks it took to beat the game, fans built deep lore to explain why the character was behaving so erratically. No, we didn’t keep opening our bag to look at the Helix Fossil because of the chaotic button mashing in the chat; indeed, it was because our player character worshiped the Helix Fossil like a deity.

In the same vein, “Nothing, Forever” fans try to parse dialogue to learn more about the universe of the show. On Discord, fans requested a new channel to keep track of new lore as it develops; one page on the wiki chronicles what we know from past mentions of aliens. The community also keeps track of Larry’s recurring standup jokes, which are pretty bad (“What do you call a bear with no teeth? A gummy bear”), but somehow get funnier the more they’re repeated.

“The way that the chat is engaging, they’re kind of creating their own memes and their own culture,” said Hartle. “We’ve had people who have reached out wanting to be community mods and have been watching the show for eight hours at a time.”

Like “Twitch Plays Pokémon,” the creators of “Nothing, Forever” hope to include audience participation features in the future. Hartle told TechCrunch that there are not currently any interactivity features embedded in the livestream, though some fans began theorizing that they were causing Larry to repeat his jokes by getting excited when he talked about gummy bears again.

Mismatch Media hopes they can repurpose the tech stack behind “Nothing, Forever” into an actual system for creating generative media projects. For now, Hartle and Habersburger are taking things slow with their newfound popularity. They’ve been able to make a bit of money from Patreon and Twitch subscriptions, but it’s still unclear how long it will take for the novelty of “Nothing, Forever” to wear off.

Sudden virality can’t last forever. “Twitch Plays Pokémon” became an ongoing series after the completion of the first game, but fan engagement dropped drastically once the excitement around the initial experiment died down. Now, we just fondly remember it as a time when the internet felt less hellish, capturing the same lightning in a bottle as projects like r/Place.

Similarly, it’s difficult to predict how long “Nothing, Forever” will retain its fanbase, and whether or not people would find more joy if the creators started other constant, AI-based sitcom spoofs.

For now, it’s a simple delight to just click off Twitter and watch Larry tell the same jokes over and over.

“What did the fish say when it swam into a wall?”

“Dam!”

Continue Reading

Trending