Tech News

Today, banks and fintech startups tend not to provide products dedicated to families, specifically, and this has appeared as something of a gap in the market. Meanwhile the general lack of financial education and financial literacy means families are missing out on securing financial prosperity for their families.

GoHenry (which raised $121.2M), which bills itself as “smart banking for kids” has attempted to crack part of this market, but is aimed at kids not families, per say. Meanwhile others chew away at Gen-Zs and parents, such as Greenlight (USA), Spriggy (AUS), Ruuky (DE), Step (US), Current (US), Nosso (UK), Unest (US).

Into this fray has stepped Bling, a startup founded by a 20-year-old, that offers a finance platform aimed specifically at families, which is designed so that parents can do financial planning for their children, from pocket money to first investments.

It’s now raised a €3.5M Seed round of financing from Peak (based out of Amsterdam); La Famiglia
; angels such as Lea-Sophie Cramer, Verena Pausder, Felix Haas (co-founder IDnow), Jakob Schreyer (Co-Founder Orderbird), former ING-Diba CEO Ben Tellings, football world cup winner Andre Schürrle, family ‘influencer’ Carmen Kroll, Angel Invest and Prediction Capital
.

The startup says it is is addressing the estimated €3.3B in pocket money given out in Germany every year, just for children aged six to 13 years, along with the €35B spent in the home market in Germany alone (German census).

The Bling product has educational modules for parents, offers a child payment card, can can cover allowances via chores, for instance.

Founder Nils Feigenwinter start Bling at only 20 years old, and created it because, he says, he was frustrated during his high school years by seeing classmates already getting into personal debt: “After twelve years of school, I looked back and realized: Nice, I can now recognize the Pythagorean theorem and mountain stones, but I have no idea about saving or handling money responsibly,” he said in a statement.

With Bling, parents sign up, but no KYC is necessary because it only operates in a first sub-€150 amount. They create a family account, receive a card, and set up their child’s account. Children learn via modules, set up savings pots, can earn money via errands and chores, and customize their cards.

After that, family members and the community join Bling via links, thus contributing to savings pots and investment plans, managing household spending and prepping for critical financial events.

Bling claims it now has 10,000+ children using a Bling Card as their first personal payment experience 6 months after launch, because it eventually taps into like grandparents, godparents, and friends, using network effects for growth.

The business model for Bling is direct subscription, transactions and fees from financial products, partnerships (first mobile phone plans, insurance etc).

Prior to Bling, Feigenwinter founded three other companies in the youth segment, including Switzerland’s largest student magazine, family merchandise and licensing house, as well as a consultancy agency specialized in young adult topics, leading him to be described as the ‘fintech wunderkind’ by German media. He’s joined by CTO and co-founder Leon Stephan.

With Bling, the fintech startup revolution spreads even to pocket money by Mike Butcher originally published on TechCrunch

Amber Group, a Sequoia- and Temasek-backed crypto trading firm, has closed a hefty $300 million Series C funding round as the collapse of FTX shakes the crypto world.

The news, which the Singapore-based firm announced on Twitter Friday morning, follows on the heels of a Bloomberg report claiming that the crypto trader has ditched a Chelsea FC sponsorship deal and is axing 40% of its staff amid market turmoil.

Like other crypto trading firms, Amber was exposed to the FTX implosion. Less than 10% of its total trading capital was with FTX at the time of the collapse, the company said, “but we did have to rebalance some positions.”

That rebalance strategy has come to light as Fenbushi Capital US, the lead investor in Amber’s latest round, pours money into the crypto market maker to keep its business afloat. Fenbushi Capital also backed Amber’s$100 million Series B round in June 2021.

“While the vast majority of our clients and products remain intact, a few of our specific products would have experienced significant drawdowns as an aftermath of the FTX default, unless we could find ways to further protect those affected clients,” Amber said in a tweet.

“That’s why we reacted quickly to adjust our fundraising strategy. The Series C investors came on board with the understanding that we will be laser-focused on providing best-in-class services to our client base of institutional and high-net-worth investors”

The Series C financing was joined by other crypto-native investors and family offices. Amber was last valued at $3 billion in its $200 million Series B extension round in February. Bloomberg reported Friday that the firm’s valuation has slid under $3 billion.

Amber, which provides liquidity and market-making services mostly in Asia, had traded $1 trillion worth of cryptocurrencies cumulatively as of February with assets under management exceeding $5 billion.

TechCrunch has reached out to Amber regarding the scale of its recent layoff. Sources told us that the trading platform, which provides a mix of institutional and retail services, is cutting a “sizable” portion of its staff.

Amber hinted in a tweet that the layoff would indeed be substantial, as it will be “scaling down our mass consumer efforts and non-essential business lines, in an effort to focus on our core businesses and clients. These have not been easy decisions, and we unfortunately have had to say goodbye to many of our excellent colleagues.”

Update from Amber’s comment: “Unfortunately, difficult but decisive adjustments were needed, and this included an organizational realignment to an estimated 300 staff as well as the prudent decision to cut management salaries, organization-wide annual bonuses and marketing expenses. This is so that we remain resilient amidst the current market environment.”

Crypto trader Amber raises $300M as it seeks protection for FTX-hit customers by Rita Liao originally published on TechCrunch

Fashion startup Virgio, co-founded by Amar Nagaram, former chief executive of Myntra, has raised $37 million in a new financing round as the young firm looks to build “a global fashion brand” from the South Asian market.

Prosus Ventures, Alpha Wave and Accel co-led Virgio’s Series A funding, valuing the one-year-old startup at $161 million (post-money). (Virgio says it was founded this year, but hasn’t disclosed its seed funding round.)

Virgio says consumers’ fashion preferences are changing rapidly and they are not satisfied with the incumbents’ offerings. It is attempting to solve this by streamlining design, manufacturing and purchasing processes in “real-time,” offering Gen-Z and late millennials a platform where they can discover and purchase the new trendy apparels.

Virgio offers a wide-range of selection in casual, party and ethnic wear. Each week it adds new selections and customers receive large discounts and free shipping.

The startup’s approach has prompted some to draw its model’s comparisons with the behemoth Shein.

“Traditionally, the fashion industry has operated on depth and discount models,” said Nagaram, founder and chief executive of Virgio, in a statement.

“At the core of Virgio is the tech foundation, which is always listening to evolving trends on social media platforms and predicting the demand for each trend. This is then fed into our agile and responsive supply chain to enable the trendiest, elegant and yet affordable line in near real-time. Thus, Virgio is pioneering the test and scale method, making runway fashion accessible and affordable for all consumers, while eliminating excess inventory for manufacturers,” said Nagaram, a Flipkart veteran who took over Myntra’s top job in 2019 before leaving the firm last year.

The startup says its full-stack approach is allowing it to condense the typical lead time of eight to ten months to just one month. Its eponymous app has already amassed over 100,000 downloads, it said. (On Sensor Tower, the app reflects very little metrics as it is still too small.)

“There is a large underserved market for branded apparel in India currently. For example, for women, only ~25% of fashion is branded apparel. The fashion industry in India is thus at an inflection point, driven by changing consumer preferences as a result of social media’s influence, an important Gen-Z characteristic to individualize as a form of expression, and the integration of social commerce with social media platforms,” said Ashutosh Sharma, Head of India Investments at Prosus Ventures, in a statement.

“We are confident in Virgio’s tech-first model and believe the startup’s expert founding team is uniquely positioned to capture a large opportunity in India’s fashion industry,” he added.

Scores of other entrepreneurs — including Mukesh Bansal, Kunal Shah, Binny Bansal, Bhavesh Agarwal, Vidit Aatrey, Saif Ali Khan, Sriharsha Majety, and Sameer Nigam — also invested in the round.

Former Myntra chief’s fashion startup Virgio valued at $161 million in new funding by Manish Singh originally published on TechCrunch

On Thursday evening, Twitter suspended a number of prominent journalists on the platform without warning or explanation.

The situation followed the company’s decision to suspend the Twitter account of Mastodon, an open source social media alternative that’s built momentum since Elon Musk took over at the company. Twitter took action against Mastodon after the account linked to the Mastodon page of @ElonJet, a student-made bot that tracks the whereabouts of Musk’s private jet.

At least some of the accounts suspended had shared screenshots and observations about Mastodon’s suspension. Just prior to his suspension, Washington Post reporter Drew Harwell tweeted about Mastodon being kicked off the platform.

Former MSNBC host Keith Olbermann, The New York Times’ Ryan Mac, CNN’s Donie O’Sullivan, Mashable’s Matt Binder, and journalist Aaron Rupar were also suspended Thursday evening. Many of the reporters regularly covered Musk’s takeover of Twitter in recent months.

Rupar weighed in on his suspension on Substack, observing that while he did not know why his account was deactivated, he did share a link to the ElonJet Facebook account in the course of reporting on the subject. Through an alternate account, Mac shared the message he received from Twitter and noted that there was no warning before the permanent suspension.

Some of the suspended accounts shared Mastodon and ElonJet’s Twitter handles as well as images of the tweet that appears to have gotten the former account suspended.

In light of Twitter’s reduced human moderation teams, it’s possible that automated systems enforcing Twitter’s brand new rules against accounts like @ElonJet were overzealous in this instance. But it’s at least as likely that this is a case of Musk directing the moderation process based on his own preferences — we just won’t know until someone at Twitter explains what’s going on.

Update:Musk just weighed in on the suspensions, characterizing them as intentional. “Same doxxing rules apply to “journalists” as to everyone else,” he tweeted in a reply.

TechCrunch has reached out to Twitter’s new head of trust and safety Ella Irwin for an explanation of why the accounts were not given the opportunity to delete the offending tweets, which is standard practice for many Twitter infractions. It’s worth noting that the policy these accounts violated, a prohibition against sharing “live location information,” is only 24 hours old.

This story is developing…

Twitter just banned prominent journalists who cover Elon Musk with no warning by Taylor Hatmaker originally published on TechCrunch

A day after Twitter crafted a new policy to explain its decision to ban an account that tracks Elon Musk’s private jet, the fallout continues.

Twitter apparently suspended its open source competitor Mastodon from the service on Thursday afternoon. Just prior to its suspension, Mastodon (@joinmastodon) tweeted a link to the jet tracking account on its own service, according to archives.

The now-banned Twitter account @ElonJet belongs to Florida student Jack Sweeney, who also operates a number of other flight-tracking bots that curate flight information from public data. Sweeney’s personal account was also suspended from Twitter along with many of the bots, including one that issued updates on Meta CEO Mark Zuckerberg.

In early November, Musk struck a different tone about the account but he’s since backtracked, adjusting Twitter’s platform policies in light of his personal preferences. “My commitment to free speech extends even to not banning the account following my plane, even though that is a direct personal safety risk,” he tweeted. That tweet is now accompanied by community notes explaining the @ElonJet saga.

Musk’s personal and political preferences have guided a number of Twitter policy decisions since the company’s hands-on new owner took over. While Musk initially declared that Twitter would allow any content that isn’t illegal, he’s since disallowed specific accounts for personal reasons.

Musk reinstated a wave of high profile Nazis and white supremacists earlier this month but drew the line at Sandy Hook conspiracy theorist Alex Jones, citing his personal experience of fatherhood.

On Mastodon, a federated, open source Twitter alternative, a single individual can’t set the rules for the whole platform. Mastodon’s servers — separate but open instances of the social network — are run by individuals who can set rules, but users can also decamp to a different server if they don’t agree with those choices.

Twitters suspends Mastodon’s account, making a good case for Mastodon by Taylor Hatmaker originally published on TechCrunch

AI-generated music is already an innovative enough concept, but Riffusion takes it to another level with a clever, weird approach that produces weird and compelling music using not audio but images of audio.

Sounds strange, is strange. But if it works, it works. And it does work! Kind of.

Diffusion is a machine learning technique for generating images that supercharged the AI world over the last year. DALL-E 2 and Stable Diffusion are the two most high-profile models that work by gradually replacing visual noise with what the AI thinks a prompt ought to look like.

The method has proved powerful in many contexts and is very susceptible to fine-tuning, where you give the mostly trained model a lot of a specific kind of content in order to have it specialize in producing more examples of that content. For instance, you could fine-tune it on watercolors or on photos of cars, and it would prove more capable in reproducing either of those things.

What Seth Forsgren and Hayk Martiros did for their hobby project Riffusion was fine-tune Stable Diffusion on spectrograms.

“Hayk and I play in a little band together, and we started the project simply because we love music and didn’t know if it would be even possible for Stable Diffusion to create a spectrogram image with enough fidelity to convert into audio,” Forsgren told TechCrunch. “At every step along the way we’ve been more and more impressed by what is possible, and one idea leads to the next.”

What are spectrograms, you ask? They’re visual representations of audio that show the amplitude of different frequencies over time. You have probably seen waveforms, which show volume over time and make audio look like a series of hills and valleys; imagine if instead of just total volume, it showed the volume of each frequency, from the low end to the high end.

Here’s part of one I made of a song (“Marconi’s Radio” by Secret Machines, if you’re wondering):

You can see how it gets louder in all frequencies as the song builds, and you can even spot individual notes and instruments if you know what to look for. The process isn’t inherently perfect or lossless by any means, but it is an accurate, systematic representation of the sound. And you can convert it back to sound by doing the same process in reverse.

Forsgren and Martiros made spectrograms of a bunch of music and tagged the resulting images with the relevant terms, like “blues guitar,” “jazz piano,” “afrobeat,” stuff like that. Feeding the model this collection gave it a good idea of what certain sounds “look like” and how it might re-create or combine them.

Here’s what the diffusion process looks like if you sample it as it’s refining the image:

And indeed the model proved capable of producing spectrograms that, when converted to sound, are a pretty good match for prompts like “funky piano,” “jazzy saxophone,” and so on. Here’s an example:

https://techcrunch.com/wp-content/uploads/2022/12/funky_sax_to_piano.mp3

But of course a square spectrogram (512 x 512 pixels, a standard Stable Diffusion resolution) represents only a short clip; a three-minute song would be a much, much wider rectangle. No one wants to listen to music five seconds at a time, but the limitations of the system they’d created mean they couldn’t just create a spectrogram 512 pixels tall and 10,000 wide.

After trying a few things, they took advantage of the fundamental structure of large models like Stable Diffusion, which have a great deal of “latent space.” This is sort of like the no-man’s-land between more well-defined nodes. Like if you had an area of the model representing cats, and another representing dogs, what’s “between” them is latent space that, if you just told the AI to draw, would be some kind of dogcat, or catdog, even though there’s no such thing.

Incidentally, latent space stuff gets a lot weirder than that:

No creepy nightmare worlds for the Riffusion project, though. Instead, they found that if you have two prompts, like “church bells” and “electronic beats,” you can kind of step from one to the other a bit at a time and it gradually and surprisingly naturally fades from one to the other, on the beat even:

https://techcrunch.com/wp-content/uploads/2022/12/church_bells_to_electronic_beats.mp3

It’s a strange, interesting sound, though obviously not particularly complex or high-fidelity; remember, they weren’t even sure that diffusion models could do this at all, so the facility with which this one turns bells into beats or typewriter taps into piano and bass is pretty remarkable.

Producing longer-form clips is possible but still theoretical:

“We haven’t really tried to create a classic 3-minute song with repeating choruses and verses,” Forsgren said. “I think it could be done with some clever tricks such as building a higher level model for song structure, and then using the lower level model for individual clips. Alternatively you could deeply train our model with much larger resolution images of full songs.”

Where does it go from here? Other groups are attempting to create AI-generated music in various ways, from using speech synthesis models to specially trained audio ones like Dance Diffusion.

Riffusion is more of a “wow, look at this” demo than any kind of grand plan to reinvent music, and Forsgren said he and Martiros were just happy to see people engaging with their work, having fun and iterating on it:

“There are many directions we could go from here, and we’re excited to keep learning along the way. It’s been fun to see other people already building their own ideas on top of our code this morning, too. One of the amazing things about the Stable Diffusion community is how fast people are to build on top of things in directions that the original authors can’t predict.”

You can test it out in a live demo at Riffusion.com, but you might have to wait a bit for your clip to render — this got a little more attention than the creators were expecting. The code is all available via the about page, so feel free to run your own as well, if you’ve got the chips for it.

Try ‘Riffusion,’ an AI model that composes music by visualizing it by Devin Coldewey originally published on TechCrunch

To get a roundup of TechCrunch’s biggest and most important stories delivered to your inbox every day at 3 p.m. PDT, subscribe here.

The TC crew are a bunch of bookworms. Alex collected the top reads from 2022. Haje’s favorite was “A Deadly Education” (Naomi Novik’s first book in the Scholomance trilogy), which incidentally was recommended by both Alex and Amanda, which was how he found out about the books in the first place. Meanwhile, Christine was unable to put down “Blue Ticket” by Sophie Mackintosh. Get down to your local indie bookstore and sniff some freshly printed wood pulp!

In non-book news, there’s a bushel of stories from TechCrunch for ya! — Christine and Haje

The TechCrunch Top 3

Does anyone else miss paper maps?: A gaggle of companies, including Meta, Microsoft, AWS and TomTom, have partnered with the Linux Foundation to form the Overture Maps Foundation to do a few things: develop interoperable open map data and knock Google’s map dominance down a few pegs. Paul has more.
Another tool to decipher Salesforce: Too much data, not enough time. Thankfully there is Sweep, a new no-code config tool that can make the data in Salesforce more usable. The company raised $28 million to put companies back in control of their own data, Kyle writes.
Going Backstage: Paul also has a story on how Spotify is planning to monetize its open source Backstage project via premium plugins.

Startups and VC

Dakotah Rice, founder of Poolit, tells Mary Ann, “2023 will be a fantastic year in all likelihood for venture and private equity. It’s kind of like you’re starting from a new base.” The company raised millions to turn accredited investors into LPs in VC and private equity funds.

Ventures Platform, a Pan-African early-stage venture capital firm, has closed its fund at $46 million as it looks to double down on “category-leading” companies across the continent, Tage writes. Most of its limited partners in the first close were primarily African based, which was a deliberate effort.

And we have five more for you:

Is it a bird? Is it a plane?: Kirsten reports that Zipline now is the national drone service provider for Rwanda.
A SaaSy hotel lobby: Ingrid covers Mews’ $185 million fundraise for its SaaS-based hotel management platform.
Rough times in VC land: Connie reports on the frustration and anger that erupted as SPV platform Assure dumps users at the curb ahead of holidays.
Support direct to customers: Kate reports that Plugo, an e-commerce support platform for D2C brands in Southeast Asia, picks up a $9 million Series A.
Armored AI: Protect AI lands a $13.5 million investment to harden AI projects from attack, Kyle reports.

Which Instagram ad placement is more cost-effective: Reels, Feed Posts or Stories?

Consumer-facing startups are spending more on platforms like TikTok and Instagram to reach customers, but which ad products offer the best return?

In a case study based on Instagram campaigns for a site that facilitates bookings for freelance beauty professionals, digital marketer Angelina Liparteliani looked at Instagram Reels, Feed Posts and Stories.

Her highly detailed breakdown includes examples of the ads used in various campaigns, the process she used for optimizing creative materials, and a cost-per-click analysis that shows how she reduced CPC from $1.51 to 17 cents.

“Definitely don’t chase trends,” advises Liparteliani. “Diversify your ad strategy, test different ideas and don’t give up if your ad doesn’t show results right away.”

Three more from the TC+ team:

Lighting the fuse: Alex wonders if China’s venture capital market can help it reignite growth.
AI attracts dollars: Companies — and VCs — continue to invest in AI despite market slowdown, reports Kyle.
Where’s the beef: Christine explores that while cultivated beef companies tout sustainability, will they also lead to marketability?

TechCrunch+ is our membership program that helps founders and startup teams get ahead of the pack. You can sign up here. Use code “DC” for a 15% discount on an annual subscription!

Big Tech Inc.

Jacquelyn got a scoop that Coinbase created an asset recovery tool for unsupported Ethereum-based tokens. If you just went “Huh?” we’ll let her explain: “In the past, if you sent assets not supported by Coinbase to a user’s address on the exchange, you’d get a message saying the assets were successfully delivered on-chain, but they didn’t actually go to the receivers’ wallets.” Coinbase supports hundreds of cryptocurrencies, but there are thousands that it doesn’t. ERC-20 tokens are one of them, and the crypto exchange found like 4,000 of them sent to its ledger. Fortunately, it now has a way to help customers recover them.

And we have five more for you:

Breaching into the void: Zack has more information on what LastPass is saying — and what it’s not saying — about its recent data breach notice.
Get your beauty on: Alex Perez-Tenessa, a former Prime Video executive, launched the video shopping app Trendio to feature all things beauty e-commerce, Aisha writes.
Do you want that slice triangle or square?: Rita reports that Lucid is poised to enter China’s crowded EV market after seeing a number of open positions in Shanghai advertised on LinkedIn.
If you’re still not scared off by crypto…: PayPal and MetaMask have something for you. The companies are uniting to create an integration that will make it easier to buy crypto, writes Romain.
Shhhhhh: Microsoft-owned GitHub is now making its secret scanning service available for all public GitHub repos for free, Frederic writes.

Daily Crunch: Major tech firms partner with Linux Foundation to support open map data development by Christine Hall originally published on TechCrunch

Elon Musk’s plan to “operate as a giant distributed utility” is creeping towards reality.

First, Tesla piloted a “virtual power plant” in California, where Powerwall home battery owners were invited to sell electricity back to the grid at peak times to mitigate brownouts. Soon after, the automaker expanded the effort into Australia and Japan, and next it’s coming to Texas — with a twist.

On Thursday, the company announced Tesla Electric, an electric plan exclusively offered in parts of Texas where retail choice is available, like Houston and Dallas. The product turns Tesla into a retail electric provider in the Lone Star state, and allows Powerwall owners to sell some of their excess electricity back to the grid, with Tesla as the go-between.

Crucially, this product is invite-only.If you own one of Tesla’s big batteries and have a home in a deregulated part of Texas, you’ll “see a banner appear on your Powerwall home page in the Tesla app,” Tesla said.

Tesla Electric builds on its vision to grow well beyond automaking, however other car makers are also toying with this idea, including GM, which is working with solar seller SunPower; and Toyota, which said today that it is teaming up with a utility in Texas.

Tesla’s energy business has seen some highs and lows this year. In July, Tesla said its solar business had notched its best quarter in several years. Four months later, a report from Electrek said that Tesla had recently canceled some of its solar reservations and laid off workers in the department.

Tesla Powerwall customers in Texas can now sell their electricity back to the grid by Harri Weber originally published on TechCrunch