The biggest challenge in managing IT infrastructure in mixed (cloud, multicloud, and hybrid) environments is complexity.Read More

Jett: The Far Shore puts the emphasis on understanding alien flora and fauna instead of destroying it, and that can create its own conflict.Read More

Polyarc unveils Moss: Book II for Sony's PlayStation VR platform, and it will be longer than the first game.Read More

Sony announced during today's State of Play that Death Stranding: Director's Cut will launch for PlayStation 5 on September 23.Read More

Arcadegeddon is a new online shooter from Friday the 13th studio, and it launches in early access today on PlayStation 5.Read More

Raffael Marty's keynote at LogPoint's customer conference describes logging history and related areas such as big data and open source.Read More

During today's PlayStation State of Play, Sony revealed that Sifu is now coming out in early 2022. It was originally set for 2021.Read More

Intel's Huma Abidi talks about her DE&I initiatives both in her private life and as part of Intel's ongoing efforts across the company.Read More

According to a new study, the datasets used to train AI models are often labeled according to drastically different standards.Read More

French startup Younited Credit has raised a $170 million funding round. Goldman Sachs is leading the round with existing investors Eurazeo, Bpifrance and AG2R La Mondiale also participating. The company offers several credit products to European consumers. It also has a diversified distribution strategy.

Consumer credit in Europe is completely different from consumer credit in the U.S. Many countries don’t rely on a central credit score system to assess your credit worthiness. Similarly, most people don’t have a credit card. Financial institutions that want to offer credit lines have to evaluate the potential risk behind a credit application. It can be a complicated and tedious process.

Younited Credit differentiates itself from legacy players with a data-driven, AI-based approach. Instead of sending a ton of documents to your banker, Younited Credit tries to automate request processes as much as possible.

The company takes advantage of DSP2 regulation and open banking APIs to ingest data. As the startup has facilitated a huge volume of credit offering, it can also leverage past data for machine learning risk models.

So far, Younited Credit has granted more than €2.4 billion in credit ($2.8 billion at today’s exchange rate). It operates in five European countries. France is still the company’s leading market as Italy, Spain, Portugal and Germany represent 40% of Younited Credit’s revenue.

More recently, the company started embedding its product into third-party products. For instance, banks and fintech companies offer credit products in their apps thanks to partnerships with Younited Credit. Examples include N26, Lydia, Orange Bank and Fortuneo. In 2021, the B2B offering represented 30% of Younited Credit’s net banking income.

Right now, Younited Credit has 440 employees. It plans to hire another 200 people over the next 18 months. The company wants to double down on European markets.

Up next, Younited Credit wants to double down on embedded finance with credit products that appear on the checkout page of popular e-commerce websites and apps. The company will compete with ‘buy now, pay later’ companies, such as Klarna, Floa, Oney, Scalapay, etc.

Named Younited Pay, the company plans to offer a wide range of options with payment terms spread over 3 to 48 months. Some companies are already using Younited Pay, such as Free, Micromania and LDLC.

The startup is offering this payment solution online and in brick-and-mortar stores. Once again, Younited Credit tries to find customers where they are already. And it seems like a smart move as physical points of sales represent over 50% of Younited Pay payments this year.

Many teachers and parents in Indonesia rely on WhatsApp to keep in touch, creating “multiple groups that often become messy and highly ineffective, and result in confusion or lost threads,” says Rizky Aniez, the co-founder and chief executive officer of Gredu. The Jakarta-based startup was created to give everyone involved in the educational process—school administrators, teachers, parents, guardians and students—apps that let them keep track of everything and communicate with one another. Today it is announcing a $4 million Series A, led by Intudo Ventures, an Indonesia-focused venture capital firm, with participation from returning investor Vertex Ventures. 

While some teachers use Google Classroom, Gredu was created to work with Indonesia’s K-12 National Curriculum and Islamic Curriculum programs, used in both private and public schools. The startup is also developing new verticals, including software for preschools and university programs. 

Founded in September 2016, Gredu is now used by more than 400 schools, with a total of 400,000 users.  Its Series A will be used to expand in the Greater Jakarta Region and into major cities throughout Indonesia, plus product development and hiring. 

Gredu’s subscription software is centered around a management system that lets administrators and teachers keep on touch of all their their tasks—including syllabuses, teaching schedules and communicating with parents and students. Aniez told TechCrunch that the onboarding process is simple, and “in an ideal solution, it can be done within hours.” Gredu was designed to be modular, so it can be customized to a school or district’s needs. 

The platform currently has four main parts. Gredu School Management System was created for administrators, while Gredu Teacher lets educators track student attendance, create and score exams and arrange class activities. Gredu Parents enables parents and guardians to keep track of their kids’ performance and talk to teachers. Gredu Student, meanwhile, lets students look up their test scores, attendance records and school activities. 

Gredu launched an Online Assignment feature before COVID-19 and during the pandemic, it added Interactive Class to enable remote learning. Aniez said the company plans to add new features and adapt Interactive Class for other uses once in-person schooling becomes the norm again. “We believe that many of the digitization in schools adopted during the pandemic will continue to be used for the future, changing the way administrators manage schools and improving transparency for local education authorities, teachers and parents,” he added.

Gredu is part of a crop of Indonesian edtech startups that have recently raised funding, including tuition platform InfraDigital; homework help and tutoring app CoLearn; and ErudiFi for education financing. 

In a statement, Intudo Ventures founding partner Patrick Yip said, “Working with school districts and administrators, GREDU provide innovative solutions specifically tailored to enhance the quality, transparency and effectiveness of Indonesia’s education system. We are proud to support GREDU at this critical juncture as they help more schools digitize their operations and create positive impact for students throughout Indonesia.” 

Amsterdam-based challenger bank Bunq has been self-funded by its founder and CEO Ali Niknam for several years. But the company has decided to raise some external capital, leading to the largest Series A round for a European fintech company.

The startup is raising $228 million (€193 million) in a round led by Pollen Street Capital. Bunq founder Ali Niknam is also participating in the round — he’s investing $29.5 million (€25 million) while Pollen Street Capital is financing the rest of the round.

As part of the deal, Bunq is also acquiring Capitalflow Group, an Irish lending company that was previously owned by … Pollen Street Capital.

Founded in 2012, Ali Niknam has already invested quite a bit of money into his own company. He poured $116.6 million (€98.7 million) of his own capital into Bunq — that doesn’t even take into account today’s funding round.

But it has paid off as the company expects to break even on a monthly basis in 2021. The company passed €1 billion in user deposits earlier this year. So why is the company raising external funding after turning down VC firms for so many years?

“Everything has a right time. In the beginning of Bunq, it was important to get a laser user focus in the company. Having to also focus on fundraises and the needs of investors distracts. Bunq now is mature enough to start scaling up significantly, so more capital is welcome,” Niknam said.

In particular, the company expects to acquire smaller companies to fuel its growth strategy. Challenger banks have also represented a highly competitive market over the past years in Europe. It’s clear that there will be some consolidation at some point.

Bunq offers bank accounts and debit cards that you can control from a mobile app. It works particularly well if your friends and family are also using Bunq as you can instantly send money, share a bunq.me payment link with other people, split payments and more.

In particular, if you’re going on a weekend trip, you can start an activity with your friends. It creates a shared pot that lets you share expenses with everyone. If you live with roommates, you can also create subaccounts to pay for bills from that account.

The company offers different plans that range from €2.99 per month to €17.99 per month — there’s also a free travel card with a limited feature set. By choosing a subscription-based business model, the startup has a clear path to profitability as most users are paid users.

Programmatic advertising is a $200 billion global marketplace that is rapidly growing and far-reaching, with Connected TV (CTV) serving as its latest accelerant. Unfortunately, however, it’s also a business sector rife with fraud and consumer privacy abuse, particularly in emerging media forms like CTV and mobile.

Global losses to ad fraud exceeded $35 billion last year, a figure expected to rise to $50 billion by 2025, according to the World Federation of Advertisers. Per the WFA, ad fraud is “second only to the drugs trade as a source of income for organized crime,” but there is no one-size-fits-all ad fraud strategy.

To capitalize on the promise of video advertising in mobile and CTV, and measure ad efficacy with confidence, business leaders must ensure that they’re reaching customers — not bots — and achieving their business goals while remaining compliant with the latest regulations and laws.

There are a few key steps business leaders can take to guard their reputation and their ad spend:

Know the different types of ad fraud in CTV and mobile in-app to better protect your ad spend

It’s important to consider the various ways your ad budgets can be squandered on invalid traffic. Although 78% of U.S. households are now reachable via programmatic CTV advertising, ad fraud rates remain high, at 24% in Q4 2020. Traditional ad fraud attacks, such as spoofing (i.e., pretending to be a different publisher) and fake sites or apps, are being supplanted by more advanced schemes, such as CTV device farms.

Knowing that ad fraud is eating away at your budget is the first step, but business leaders need to understand the different schemes so they can apply the right protection in the right moments.

Looking at reach through a quality lens

Historically, the standard way to measure advertising has been focused on reach. However, reach is now more of a vanity metric if uncoupled from traffic quality.

Striving for reach while ignoring quality creates a prime opportunity for ad fraud. Generating fake traffic to create the illusion of “reach” has become a staple in many ad fraud schemes, with some CTV schemes fabricating up to 650 million bid requests a day per day from bots, per The Drum.

High impression rates that fail to convert into actual sales and pricing anomalies (as compared to a peer group) are compelling harbingers of traffic quality issues.

Because the growing CTV ecosystem commands premium pricing, advertisers may be tempted to seek out deals. However, several leading streaming TV providers, such as XUMO and Philo, have warned advertisers about prices that seem too good to be true, noting that they may be signals of fraudulent activity. Work to identify where traffic is coming from and ask questions when the data looks suspicious.

The ad industry itself is also fighting back by giving tools to business owners meant to stymie ad fraud. There are several industry working groups and watchdog organizations — including the Media Rating Council, Interactive Advertising Bureau and Trustworthy Accountability Group — that accredit certain platforms and suppliers to combat ad fraud. These working groups and organizations also regularly release industry standards and programs designed to address fraudulent activity, such as the Ads.txt initiative meant to help advertisers know they are buying inventory via legitimate third parties. All business owners should utilize certified platforms — along with emerging programs and standards — to stay on top of the latest trends in ad fraud.

Business leaders need to prioritize brand safety and compliance

In addition to navigating the complex world of ad quality, brands must now consider whether the publishers they are working with are brand safe and compliant with the latest consumer privacy and compliance laws.

Pixalate’s May 2021 estimates show that 22% of Apple App Store apps and 9% of Google Play Store apps that serve programmatic advertisements don’t have a privacy policy. This is significant because there are already documented cases of consumer data being misused as part of ad fraud schemes. And 70% of Google Play Store apps have at least one of what Google calls “dangerous permissions,” an increase of 5% in 2020. Additionally, of apps that serve programmatic ads, 80% of Apple App Store apps and 66% of Google Play Store apps count children aged 12 and under as part of their audience, which brings COPPA compliance risks into the equation as well.

There are a couple of things at play when it comes to brand safety that business leaders and brands should be aware of. The most important is that what is deemed “safe” for a brand is solely based on that brand — there is no golden standard because each brand has a different vision, mission and goals. Brand safety is subjective. However, it’s essential for success.

Ad fraud, brand safety and data compliance continually evolve, and leaders must follow the numbers, stay educated on market changes, and invest in the right partnerships to ensure consumers, not bots, are engaging with the most impactful and effective content.

AI teams invest a lot of rigor in defining new project guidelines. But the same is not true for killing existing projects. In the absence of clear guidelines, teams let infeasible projects drag on for months.

They put up a dog and pony show during project review meetings for fear of becoming the messengers of bad news. By streamlining the process to fail fast on infeasible projects, teams can significantly increase their overall success with AI initiatives.

AI projects are different from traditional software projects. They have a lot more unknowns: availability of right datasets, model training to meet required accuracy threshold, fairness and robustness of recommendations in production, and many more.

In order to fail fast, AI initiatives should be managed as a conversion funnel analogous to marketing and sales funnels. Projects start at the top of the five-stage funnel and can drop off at any stage, either to be temporarily put on ice or permanently suspended and added to the AI graveyard. Each stage of the AI funnel defines a clear set of unknowns to be validated with a list of time-bound success criteria.

The AI project funnel has five stages:

1. Problem definition: “If we build it, will they come?”

This is the top of the funnel. AI projects require significant investments not just during initial development but ongoing monitoring and refinement. This makes it important to verify that the problem being solved is truly worth solving with respect to potential business value compared to the effort to build. Even if the problem is worth solving, AI may not be required. There might be easier human-encoded heuristics to solve the problem.

Developing the AI solution is only half the battle. The other half is how the solution will actually be used and integrated. For instance, in developing an AI solution for predicting customer churn, there needs to be a clear understanding of incorporating attrition predictions in the customer support team workflow. A perfectly powerful AI project will fail to deliver business value without this level of integration clarity.

To successfully exit this stage, the following statements need to be true:

In my experience, the early stages of the project have a higher ratio of aspiration compared to ground realities. Killing an ill-formed project can avoid teams from building “solutions in search of problems.”

2. Data availability : “We have the data to build it.”

At this stage of the funnel, we have verified the problem is worth solving. We now need to confirm the data availability to build the perception, learning and reasoning capabilities required in the AI project. Data needs vary based on the type of AI project  —  the requirements for a project building classification intelligence will be different from one providing recommendations or ranking.

Data availability broadly translates to having the right quality, quantity and features. Right quality refers to the fact that the data samples are an accurate reflection of the phenomenon we are trying to model  and meet properties such as independent and identically distributed. Common quality checks involve uncovering data collection errors, inconsistent semantics and errors in labeled samples.

The right quantity refers to the amount of data that needs to be available. A common misconception is that a significant amount of data is required for training machine learning models. This is not always true. Using pre-built transfer learning models, it is possible to get started with very little data. Also, more data does not always mean useful data. For instance, historic data spanning 10 years may not be a true reflection of current customer behavior. Finally, the right features need to be available to build the model. This is typically iterative and involves ML model design.

To successfully exit this stage, the following statements need to be true:

Pakistan, the world’s fifth most populous country, has been slow to adapt to the internet economy. Unlike other emerging economies such as China, India and Indonesia, which have embraced digitization and technology, Pakistan has trailed the region in the adoption of technology and startup formation.

Despite this, investors have dreamed for years of the huge opportunities in unlocking Pakistan’s potential as a digital economy. As a country of 220 million people, almost two-thirds of whom are under the age of 30, Pakistan draws natural comparisons to Indonesia — which has rapidly emerged as one of the most vibrant technology ecosystems outside the U.S. and China.

After years of lagging behind, over the course of the past 18 months, Pakistan’s technology ecosystem has come to life in unprecedented fashion. In 2021, Pakistani startups are on track to raise more money than the previous five years combined. Even more excitingly, a large portion of this capital is coming from international investors from across Asia, the Middle East and even famed investors from Silicon Valley.

The rapid emergence of Pakistan’s technology ecosystem on the international stage has been no accident — it’s the result of a confluence of changing facts on the ground and shifting dynamics in the startup and investing world as a result of the pandemic.

Unlocking Pakistan’s potential

The sudden emergence of Pakistan’s tech ecosystem on the international stage has been driven by three major factors: an improving security situation, quickly growing mobile connectivity, and critical legal changes and deregulation.

As a frontline state and coalition partner in the United States’ invasion of Afghanistan, Pakistan saw fatalities from terrorist violence soar from 295 in 2001 to a peak of over 11,000 in 2009. This climate of instability and violence scared away international business and investors from Pakistan for much of the first two decades of the 21st century.

WellSaid Labs, whose tools create synthetic speech that could be mistaken for the real thing, has raised a $10M Series A to grow the business. The company’s home-baked text-to-speech engine works faster than real time and produces natural-sounding clips of pretty much any length, from quick snippets to hours-long readings.

WellSaid came out of the Allen Institute for AI incubator in 2019, and its goal was to make synthetic voices that didn’t sound so robotic for common business purposes like training and marketing content.

It achieved that first by basing its solution on Tacotron, a speech engine developed by Google and academic researchers. But soon it had built its own that was more efficient, resulted in more convincing voices and could produce clips of arbitrary lengths. Speech engines often trip up after a couple sentences, descending into babble or losing tone, but WellSaid’s read the entirety of Mary Shelley’s “Frankenstein” without a hiccup.

The voices were good enough that they were rated as human or as good as human by listeners — not something you could really say about the usual virtual assistant suspects when they speak more than a handful of words. Not only that, but the speech was generated considerably faster than real time, where other high-quality options often operated at a tenth real time or slower — meaning three minutes of speech would take one minute to generate by WellSaid and half an hour or more by Tacotron.

https://techcrunch.com/wp-content/uploads/2020/09/wellsaid-clip.mp3

Lastly, the system allows for new “Voice Avatars” to be created based on existing voice talent, like a trusted company spokesperson or voiceover artist. Originally about 20 hours of audio was needed to build a model of their quirks and voice style, but now it can do so with as little as two hours, CEO Matt Hocking said.

The company is strictly business-focused right now, which is to say there’s no user-facing app to digitize your voice into an avatar or anything. There are attendant risks and no realistic business model for it, so that’s off the table for now.

Such a realistic voice might still be of enormous help to people with disabilities, however, something Hocking acknowledges but admits they’re not quite ready to tackle yet.

“We are committed to expanding access to this technology so that nonverbal communicators, nonprofits and others can benefit from it,” he said.

In the meantime the company has expanded from its first market, corporate training videos, to marketing, longer copy, interactive products with considerable text and app experiences. One hopes that the talent these avatars are based on are being properly compensated for helping create a digital likeness of their voice.

The oversubscribed $10M round was led by FUSE, with participation from repeat investor Voyager, Qualcomm Ventures LLC and GoodFriends, all of whom were likely impressed by the product and business growth. Synthetic voices have served a handful of popular use cases but content has not been a big one — so there’s plenty of room to grow. The company will invest the money in deepening its product offering and growing the team along with it.

The U.S. SPAC market kept rolling along this week with news that Satellogic will go public on the Nasdaq stock exchange thanks to a merger with a blank check company. The Earth-imagery-focused company is standard SPAC fare, with strong capital needs and distant revenues. It was not alone in pursuing the transaction type Tuesday, with news breaking that Nextdoor will also go public on the Nasdaq via a SPAC.

Nextdoor’s projections, as TechCrunch noted, were more modest and thus more believable than what we’ve seen from many other SPAC-led debuts.

These companies represent the two poles of blank-check-powered public offerings: Some startups taking the SPAC route are more speculative, banking on revenues to come, while others feature more established companies with a history of material revenue growth. It’s easy to find more examples of both varieties. Acorns’ deal fits the established trend. Lidar SPACs? Less so.

The Exchange explores startups, markets and money.

Read it every morning on Extra Crunch or get The Exchange newsletter every Saturday.

Given the breadth of companies pursuing blank-check deals, the SPAC boom isn’t over even if there has been chatter that the party is breaking up. Bessemer partner Mary D’Onofrio told The Exchange, for example, that while the “pace of SPAC IPOs” and combinations have slowed, “there is still $128 billion of SPAC dry powder in the market seeking acquisitions and incentivized to transact.”

Matt Murphy, a partner at Menlo Ventures, helped explain the SPAC pace deceleration that D’Onofrio discussed, telling The Exchange that the pace of SPAC deals “has slowed as they’ve gotten more scrutiny and don’t seem quite as ‘easy’ as they once were.”

But this week’s U.S. SPAC news tells us that blank-check companies are still finding a diverse set of companies to take public. But what about other regions? Unicorns are hardly unique to the U.S. startup ecosystem. Are we seeing similar SPAC interest in Europe?

The Exchange tried to find out, given that we’ve seen huge rounds from the region and a few IPOs over various types. Is the SPAC game afoot in Europe?

Hunting European targets

There’s a huge number of SPACs trading in the United States currently hunting for a deal. And there is historical precedent for U.S.-listed blank-check companies taking on European targets. Global law firm Skadden counts 16 U.S. SPAC-led transactions with European companies from 2015 through February of this year, for example.

“For the past few weeks, we’ve been approached on a recurring basis, much like all known French and European scaleups,” Aircall’s co-founder Jonathan Anguelov told French financial newspaper Les Échos last March (translation: TechCrunch). However, being approached doesn’t necessarily mean that European unicorns are entertaining the offers.

This morning r2c, a startup building a SaaS service around the Semgrep open-source project, announced that it has closed a $27 million Series B. Felicis led the round, which the company said was a pre-emptive deal.

Prior investors firms Redpoint and Sequoia also participated in the fundraising event; r2c last raised a $13 million Series A in October of 2020.

The startup fits into several trends that TechCrunch has explored in recent quarters, including what appears to be a growing number of open-source (OSS) grounded startups raising capital, more rounds coming to exist thanks to investors looking to get the jump on inside rounds before they can form.

On the OSS point, r2c works with Semgrep, which the company likens to a “code-aware grep.” Still confused? Don’t worry, this is all a bit technical, but interesting. Grep is a tool for searching through plain-text that has been around for decades. Semgrep is related, but focused on finding things inside of written code.

Given the sheer volume of code that is written daily in the world, you can imagine that there is an ever-rising demand for finding particular bits of text quickly; Semgrep is an evolution of the original project, that was initially built inside of Facebook.

Per r2c CEO Isaac Evans, however, the project failed to attract much awareness. His startup has built what Evans described to TechCrunch has the “canonical” Semgrep fork, or version, and has crafted a software service around the code to make it easier for other companies to use.

There are many ways to generate revenue from open-source software. Two popular monetization routes are througuh support services or offers to host particular projects. But, R2c is a doing something a bit different. The startup sells a monthly, per-developer subscription (SaaS) that packages a broad set of security-focused rules across different coding languages, allowing companies to easily check their own software for possible security issues.

Or as Evans succinctly explained it, r2c offers something akin to application security in a box.

Focusing on cybersecurity is a reasonable tack for the company. Given the ever-growing number of breaches that the public endures, helping companies leak less data, and suffer fewer intrusions is big business.

You don’t have to pay r2c, however. Semgrep is OSS and the rules associated with various languages are available under a LGPL license — more on that definition here. Developers could build their own version of what the company offers. But, Evans argued, it won’t be ready to help you pick which rules you may want to apply to your code, something that his company is happy to help with for a fee.

From a wide lens, r2c fits into the developer tools category. It is content to land and expand inside of companies, perhaps allowing it a lower cost of acquiring customers than we see at some SaaS startups. But that doesn’t mean that the company won’t go to market to sell its service. Per Evans, the startup has historically underinvested in marketing, something that it may now be able to focus more on thanks to its recent financing.

It is not uncommon to see companies with technically-minded founders initially spend too little on the sales and marketing parts of operating a software business. But our impression after discussing the company’s plans with Evans is that r2c intends to get that part of its house in order.

Evans told TechCrunch that his company took aboard more cash because it doesn’t want to build the best search tool for, say, the C programming language. It wants to go broad, fusing what the CEO described as the “customizability of Semgrep” and wide language support.

Let’s see how quickly the company can staff up, bolster its marketing efforts, and take on enterprise clients. Raising a Series C puts the company somewhere past its startup adolescence, so from here on out we can pester the company for concrete growth numbers.

If you’re a frequent TechCrunch reader, you probably already know about mmhmm, the startup with the name you likely either love or hate. It’s Phil Libin’s second act after Evernote, and it’s a startup born of the pandemic maybe more so than any other, providing improved video chat tools, including automatic background removal and advanced presentation features. The company, which is just over a year old, has now raised a total of around $140 million thanks to a fresh injection of $100 million first reported by Bloomberg on Tuesday, which is somewhat astounding if you remember using the first early beta versions like me.

Startups with silly names raising lots of money is hardly an exceptional occurrence in tech, but Libin’s startup earns extra credit for barely having a name at all (it’s really just a sound).

The company was built on the idea that current video tools really fail to provide users with access to all the potential that modern technology offers, particularly when it comes to presentations. Mmhmm’s core presenter tools help your meetings look more like professional newscasts than warmed-over digital versions of transparency slideshows and whiteboard scrawls, and the company has steadily been adding features and improving its performance through frequent iterations since its founding.

As it stands, mmhmm works in tandem with the existing video services that people use for virtual meetings, including Zoom. But Bloomberg says it’s going to go standalone as well, and introduce a mobile app version. That sounds like a good use of the new funds, which come from SoftBank’s Vision Fund, Sequoia Capital and more.

Even projecting forward to a post-pandemic world where virtual meetings are less important, they’re probably still a permanent part of the working world. But mmhmm’s feature set also seems to almost define the concept of ‘feature, not product’ that is presented as a cautionary tale to startups crafting wings of wax and soaring as high as they can in terms of raises and valuation.

The U.S. Department of Defense canceled a $10 billion JEDI cloud computing project with Microsoft and is pursuing a multicloud deal.Read More