Monthly Archives: May 2021

News: Dabbel gets $4.4M to cut CO2 by automating HVAC for commercial buildings

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

Düsseldorf-based proptech startup Dabbel is using AI to drive energy efficiency savings in commercial buildings. It’s developed cloud-based self-learning building management software that plugs into the existing building management systems (BMS) — taking over control of heating and cooling systems in a way that’s more dynamic than legacy systems based on fixed set-point resets. Dabbel

Düsseldorf-based proptech startup Dabbel is using AI to drive energy efficiency savings in commercial buildings.

It’s developed cloud-based self-learning building management software that plugs into the existing building management systems (BMS) — taking over control of heating and cooling systems in a way that’s more dynamic than legacy systems based on fixed set-point resets.

Dabbel says its AI considers factors such as building orientation and thermal insulation, and reviews calibration decisions every five minutes — meaning it can respond dynamically to changes in outdoor and indoor conditions.

The 2018-founded startup claims this approach of layering AI-powered predictive modelling atop legacy BMS to power next-gen building automation is able to generate substantial energy savings — touting reductions in energy consumption of up to 40%.

“Every five minutes Dabbel reviews its decisions based on all available data,” explains CEO and co-founder, Abel Samaniego. “With each iteration, Dabbel improves or adapts and changes its decisions based on the current circumstances inside and outside the building. It does this by using cognitive artificial intelligence to drive a Model-Based Predictive Control (MPC) System… which can dynamically adjust all HVAC setpoints based on current/future conditions.”

In essence, the self-learning system predicts ahead of time the tweaks that are needed to adapt for future conditions — saving energy vs a pre-set BMS that would keep firing the boilers for longer.

The added carrot for commercial building owners (or tenants) is that Dabbel squeezes these energy savings without the need to rip and replace legacy systems — nor, indeed, to install lots of IoT devices or sensor hardware to create a ‘smart’ interior environment; the AI integrates with (and automatically calibrates) the existing heating, ventilation, and air conditioning (HVAC) systems.

All that’s needed is Dabbel’s SaaS — and less than a week for the system to be implemented (it also says installation can be done remotely).

“There are no limitations in terms of Heating and Cooling systems,” confirms Samaniego, who has a background in industrial engineering and several years’ experience automating high tech plants in Germany. “We need a building with a Building Management System in place and ideally a BACnet communication protocol.”

Average reductions achieved so far across the circa 250,000m² of space where its AI is in charge of building management systems are a little more modest but a still impressive 27%. (He says the maximum savings seen at some “peak times” is 42%.)

The touted savings aren’t limited to a single location or type of building/client, according to Dabbel, which says they’ve been “validated across different use cases and geographies spanning Europe, the U.S., China, and Australia”.

Early clients are facility managers of large commercial buildings — Commerzbank clearly sees potential, having incubated the startup via its early-stage investment arm — and several schools.

A further 1,000,000m² is in the contract or offer phase — slated to be installed “in the next six months”.

Dabbel envisages its tech being useful to other types of education institutions and even other use-cases. (It’s also toying with adding a predictive maintenance functionality to expand its software’s utility by offering the ability to alert building owners to potential malfunctions ahead of time.)

And as policymakers around the global turn their attention to how to achieve the very major reductions in carbon emissions that are needed to meet ambitious climate goals the energy efficiency of buildings certainly can’t be overlooked.

“The time for passive responses to addressing the critical issue of carbon emission reduction is over,” said Samaniego in a statement. “That is why we decided to take matters into our own hands and develop a solution that actively replaces a flawed human-based decision-making process with an autonomous one that acts with surgical precision and thanks to artificial intelligence, will only improve with each iteration.”

If the idea of hooking your building’s heating/cooling up to a cloud-based AI sounds a tad risky for Internet security reasons, Dabbel points out it’s connecting to the BMS network — not the (separate) IT network of the company/building.

It also notes that it uses one-way communication via a VPN tunnel — “creating an end-to-end encrypted connection under high market standards”, as Samaniego puts it.

The startup has just closed a €3.6 million (~$4.4M) pre-Series A funding round led by Target Global, alongside main incubator (Commerzbank’s early-stage investment arm), SeedX, plus some strategic angel investors.

Commenting in a statement, Dr. Ricardo Schaefer, partner at Target Global, added: “We are enthusiastic to work with the team at Dabbel as they offer their clients a tangible and frictionless way to significantly reduce their carbon footprint, helping to close the gap between passive measurement and active remediation.”

 

News: TechCrunch’s Equity podcast wins a Webby Award

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

Today, the fine folks at the Webbys announced that TechCrunch’s flagship podcast, Equity, is the best of its kind in the technology category. We’re stoked! Alex Wilhelm, Natasha Mascarenhas and Danny Crichton sit down in front of their mics multiple times each week to regale dedicated listeners with news and analysis from the money part

Today, the fine folks at the Webbys announced that TechCrunch’s flagship podcast, Equity, is the best of its kind in the technology category. We’re stoked!

Alex Wilhelm, Natasha Mascarenhas and Danny Crichton sit down in front of their mics multiple times each week to regale dedicated listeners with news and analysis from the money part of the startup world.

Led by the deft production touch of Chris Gates and Grace Mendenhall, the entertaining trio goes deep into companies, topics and news multiple times each week – while mostly in a good mood. And they’ve come a long way.

Check out the shortest acceptance speech in TechCrunch history here!

Equity launched in March 2017 in a back room — a veritable closet, if you will! — at our old San Francisco office. Alex would come down once a week from his then-digs at Crunchbase News to join Katie Roof and Matt Lynley. Other hosts on our fun ride have included Kate Clark and our very own Connie Loizos.

We’ve got all kinds of podcast goodies up our award-winning sleeves and we hope you and five of your closest friends keep coming along for the ride.

News: A new book aims to blow up some widely held assumptions about the best founding teams

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

There’s a lot of how-to guidance out there when it comes to starting a company, and much of it has reinforced certain beliefs, including that solo founders don’t get very far on their own, that the most successful founders attend a small circle of top schools, and that the best companies are created by people

There’s a lot of how-to guidance out there when it comes to starting a company, and much of it has reinforced certain beliefs, including that solo founders don’t get very far on their own, that the most successful founders attend a small circle of top schools, and that the best companies are created by people who launched them to solve a personal problem into which they had a particular insight.

Ali Tamaseb, who studied biomedical engineering at Imperial College London, attended business school at Stanford, and founded a wearable tech startup before joining the venture firm DCVC as an investor in 2018, says that lot of that guidance is, well, misguided. Tamaseb says he knows this because over the past four years, to improve his own decision-making, he amassed more than 30,000 data points about so-called “super founders,” from their age when their breakout company was founded to how many competitors they faced from the outset; in doing so, he says, he wound up discovering that much of what is espoused in startup circles is off the mark.

To share some of those learnings, Tamaseb has written a newly published  book called Super Founders: What Data Reveals About Billion Dollar Startups. We talked with him yesterday. Our chat has been edited lightly for length.

TC: Why write this book?

AT: When I was a founder, a lot of my perception was shaped through this lens of what the media tells us. Even now, when I’m on Twitter or Clubhouse, a lot of what I hear feels very different compared with what I see as a venture capitalist. Of course, nobody knows everything, and even the most successful venture capitalists have maybe invested in 10 of these breakout companies in their lifetime. So to get to the ground truth, and because nobody has collected this data because it’s hard, over four years of weekends and evenings, I began to collect [it], ultimately establishing 65 data points per company.

TC: What are some of these data points and how much of them were publicly available?

AT: The data includes the career path of the founder and whether their earlier roles were technical or not;  the founder’s education and degree and the school they attended; what the market looked like when their company was started, including what the defensibility factors were; how many competitors they had; their fundraising history — how much did they raise and when and from whom?

There’s a bit of this data in a Pitchbook or Crunchbase, but nobody had put [these other pieces] into [context]. I was going to their LinkedIn profiles, reading interviews, going to the internet’s Wayback Machine and other archives to read reports to understand what these companies looked like at the time they were founded. I also called a lot of these founders to ask them for answers where I couldn’t find the information. It was a very manual thing.

TC: How many founders did you research?

AT: I researched more than 200 unicorns, [launched by] around 500 founders. Because no study has meaning unless you also collect data on a control group, I selected companies that over the same time period had raised a minimum of $3 million in venture capital as my baseline group. I then compared the two groups based on these 65 different elements.

TC: Your research led you to invest in the primary and urgent care company Carbon Health on behalf of DCVC. How so?

AT: Specifically here, founder Eren Bali had had built a bunch of companies before, and a bunch had failed or succeeded on a smaller level, then his last company was [the edtech giant] Udemy, where he spent four years.

One of the key things that I observed in the data is that it’s all about these small steps — and the small or neutral exits. Around 60% of these “super founders” started something earlier, and many actually lost a bunch of money; just 42% of them had a previous exit of $10 million or more, so the majority had “failed” in the world of venture capital. But [the data suggests that] practice makes perfect.

TC: You also found that solo founders aren’t doomed to run smaller companies, despite some earlier thinking by Y Combinator’s Paul Graham that you need at least two cofounders to do something big.

AT: Right, 20% of the founders in both groups — the unicorn and non unicorn group — were solo founders, so VCs are funding solo founders and they are building billion-dollar companies. Basically, one out of every five unicorn companies has a solo founder. So I think that’s another narrative that gets retold, including on Twitter, but that doesn’t match reality. Flexport, for example, has a solo founder [in Ryan Petersen]. So does CarGurus, which was founded by Langley Steinert, who, by the way, first cofounded TripAdvisor [and more recently founded ApartmentAdvisor].

TC: Your book also asserts that there are plenty of founders of billion-dollar companies that didn’t attend elite American universities.

AT: There are schools that founders attended more than others — Stanford, MIT, Wharton and Harvard — but as many of these founders attended schools that aren’t even at the top 100 [ranked U.S. schools] compared to those who went into the top 10. It’s a barbell distribution. Around 36% went to the top 10 schools, the same percentage went to schools not in the top 100, and there’s another 30% or so in the middle.

TC: Two other observations in the book that are interesting are that half the founding CEOs you researched were non-technical, and only 30% had domain expertise in the industry they were disrupting. The latter may surprise readers particularly.

AT: Yes, 30% of founders in consumer tech and 40% in enterprise tech did not come from the same domain [that their company now operates in]. And I see the same thing in startups that are just now getting funded. What it tells you is that domain expertise is not necessarily correlated to success.

Take Nat Turner of Flatiron Health [a cancer-focused start-up that sold to Roche Group in 2018]. These guys were serial entrepreneurs and they had a bunch of successes before, and they jumped from one industry to another, starting with a pizza delivery company they started in college, where they learned about the restaurant industry and deliveries and logistics. They also sold an ad tech company to Google. Then they go and start this company in the cancer oncology IP and data space, where they didn’t know anything, but they learned as much as anybody after spending two years going and talking with every oncologist they could find in New York to understand the space. So maybe founders apply their tech background to different industries or they apply soft skills like resources and connections to learn about a specific industry rather than coming from that industry.

TC: What did your research tell you about funding and its impacts? We’re seeing companies raise bigger rounds faster than ever before, including from Tiger Global.  

AT: I don’t specifically have any brilliant thoughts on Tiger or anyone else, but these unicorns that I studied — even in their seed round and Series A rounds, they had raised two to three times larger rounds than the companies that did not become billion dollar companies. In fact, 92% of these billion-dollar companies were venture backed and they raised a lot of money, which allowed them to attract better talent and go to market faster. Even from the early stage, this kingmaker strategy kind of works.

News: WalkMe is going public: Let’s stroll through its numbers

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

This afternoon, we’re parsing the WalkMe IPO filing, the second Israel-based tech company to file to go public this week.

Hot off the heels of our look into Marqeta’s IPO filing and dives into SPACs for Bright Machines and Bird, we’re parsing the WalkMe IPO filing. Later this week, Squarespace will direct list and we’ll see IPOs from Oatly and Procore. It’s a super busy time for public debuts of all sorts.

Given how hectic the IPO market is, we’re going to skip our usual throat clearing and dig into WalkMe’s IPO document. As always, we’ll start with a brief overview of its product and then move into discussing its financial performance.

WalkMe is the second Israel-based technology company to file to go public this week: No-code startup Monday.com is also pursuing an American IPO.

Alright! Into the breach.

What does WalkMe do?

WalkMe’s software provides visual overlays on websites that help users navigate the product in question. I base that explanation on my time at Crunchbase, which was a customer during at least part of my time there. WalkMe is popular with marketing teams who want to introduce users to a new or refreshed experience.

Per the company’s F-1 filing, other elements of its service that matter include its onboarding system and what WalkMe calls Workstation, or its “single interface to the applications within an enterprise and simplifies task completion through a natural language conversational interface and automation.” We’re including that last feature because it says “automation,” which, in the wake of the UiPath IPO, is a word worth watching. Investors are.

At a high level, WalkMe is a SaaS business, which means that when we digest its results we are digging into a modern software company. Let’s do just that.

WalkMe’s numbers

From 2019 to 2020, WalkMe grew its revenues from $105.1 million to $148.3 million, a gain of 41%. In its most recent quarter, the company’s growth rate slowed: From Q1 2020 to Q1 2021, WalkMe’s top line grew 25% from $34.2 million to $42.7 million.

In SaaS terms, WalkMe calculates that its annual recurring revenue, or ARR, grew from $131.2 million at the end of 2019 to $164.3 million in 2020. In more granular terms, the company’s ARR grew from $137.8 million to $177.5 million in the first quarters of 2020, and 2021, respectively.

News: Skittish is what you’d get if you crossed Animal Crossing with Clubhouse

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

If the Instagram ads feel like they’re closing in and you can’t bring yourself to toggle your Zoom camera on these days, you’re far from alone. Well into 2021, many of the social apps and virtual chat tools that kept the world connected during the pandemic feel more exhausting than the real-life interactions they’re meant

If the Instagram ads feel like they’re closing in and you can’t bring yourself to toggle your Zoom camera on these days, you’re far from alone. Well into 2021, many of the social apps and virtual chat tools that kept the world connected during the pandemic feel more exhausting than the real-life interactions they’re meant to simulate.

But what if hanging out online was… not miserable?

That’s the idea behind Skittish, a virtual browser-based event platform from XOXO co-founder Andy Baio. Skittish is a playful cross between a social audio chat app like Discord or Clubhouse and a cute video game, replete with round, colorful animal avatars to choose from. Unlike a Zoom call, Skittish is a place — one where its inhabitants can bump into one another, do activities together and wait for serendipity to strike.

Skittish is a natural extension of Baio’s interests, a sort of inviting, lightly indie-gamified space where creative people can showcase their work and hang out. “I think I’m just drawn to places where people can be themselves,” Baio told TechCrunch. “With Skittish, it’s been really important to me that people can engage at the level they’re comfortable with.”

Baio has a reputation for curating social spaces, though previously they were mostly IRL. In 2012, Baio co-created XOXO, a whimsical Portland-based festival for quirky people who make stuff. While the festival took a few years off due to Covid, the event lives on in a bustling online community full of indie game devs, offbeat podcasters and digital artists. Prior to XOXO, Baio worked on Kickstarter pre-launch and went on to serve as the crowdfunding site’s first chief technology officer. (Full disclosure: I’m a former XOXO attendee who is part of the community.)

The aptly named Skittish is meant to create an online social space that doesn’t put people on the spot. In Baio’s ideal virtual world, introverts could circle the periphery while extroverts could plunge right in and hold court at the center, just like they might in real life. That range of social styles that isn’t reflected in virtual environments that are either explicitly for work or modeled after work and it’s enough to inspire dread for a lot of people.

For Baio, audio chat hits a sweet spot. Taking the camera out of the equation makes people feel socially fluid, but audio still evokes a degree of social presence that text can’t compete with.

“There’s an assumption in a lot of virtual events that people want to be on camera all the time with strangers, which feels alien to me,” Baio said. “Skittish is audio by default, and uses spatial audio so that you can hear people around you and lurk a little bit before deciding if you want to jump into a conversation. Socializing anywhere, even online, can be really anxiety-inducing.”

Clubhouse might be synonymous with social audio right now, but its structure still doesn’t appeal to everyone. “I like the casual and conversational approach to audio, but [it] just feels like a series of conference panels and needs a strong moderator to be compelling enough to tune in,” Baio said.

How Skittish works

In Skittish, walking up to a group of people (animals, really — Skittish users can differentiate themselves by choosing one of more than 75 deeply cute animal avatars) allows you to hear a conversation just like you would in real life. Backing away, you’d hear that chatter fade until eventually it wouldn’t be audible any more. To have a more private side conversation, you and a friend (a crocodile, maybe?) could peel off from a cluster of other people and deepen your chat on a virtual walk.

Inside a Skittish room, event participants can walk around, chat with others over a mic, place virtual objects and even hop through portals to other rooms. Anyone running a Skittish space can stream videos and music from YouTube or Soundcloud to a virtual screen. Event organizers can also broadcast themselves or other speakers to the full room, overriding the normal proximity rules that let you hear what’s around you.

Baio doesn’t imagine Skittish as a persistent social space, but instead wants it to provide a flexible, playful platform for all kinds of events, everything from live podcast readings and tabletop games to larger company events. Baio says the core target audience for Skittish is “anybody with a Patreon,” and larger company events will offset costs for creators who use Skittish to connect with their communities. Anyone hosting an event can choose to either populate a virtual space with pre-designed virtual objects (think pirate ships and giant doughnuts) or dream up their own environment from scratch.

By designing a service that only exists when people need it, he hopes to avoid the harassment and toxicity that abounds on big social networks. Skittish will still pack a set of tools that allow a space’s creator to mute, kick or even ban users, but ideally, it won’t need it.

“I’m a big fan of dark social, in general, where people can feel more like themselves and moderation is much more human and manageable,” Baio said.

Building Skittish and what’s next

The pandemic shed new light on what people really want out of online social spaces. Zoom’s novelty wore off quickly, and by late 2020 group video chat felt like an entrenched fixture of virtual work, not virtual play. It shouldn’t be surprising that a gentle social simulator with light multiplayer features emerged as the game of 2020.

“It’s a bit of a cliché, but Animal Crossing: New Horizons became a reliable escape for me during the pandemic, a daily source of comfort and routine when we couldn’t go outside,” Baio said. He was charmed by his first foray into the series’ famously soothing rhythms and the game helped him envision Skittish.

“… I think what inspired me most were the simplicity of the controls and camera, the overall tone of the game, and the social features, limited as they are,” Baio said. “You’re capped at seven visitors and it takes forever for people to fly in, but despite that, it’s just a joyful experience to have a bunch of people over to your island.”

With the Nintendo Switch sold out everywhere and Animal Crossing racing up the console’s all-time sales charts, it was obvious early on that something resonated. People who wouldn’t normally consider themselves gamers bought Switches and spent hours shaking virtual trees, chatting with squirrels and touring friends’ islands for interior design tips. With Skittish, Baio hopes to capture a little of that same magic.

Games that double as social networks are booming right now — and with good reason. For many people it’s more natural to socialize when you’re ambiently doing something else together, whether that’s teaming up for a Fortnite duo, building a viking longhouse in Valheim or sampling user-built games within Roblox.

Socializing online with avatars also lets you express yourself in a meaningful enough way that Epic built an entire business around it, with sales of skins (virtual outfits) and emotes (dance moves and gestures) making up the lion’s share of Fortnite revenue.

Skittish grew out of a $100,000 grant awarded by Grant For The Web, a fund created by Coil, Mozilla and Creative Commons to support projects that incorporate micropayments for online creators. Baio began prototyping Skittish last July, imagining it as a pop-up space for events rather than a persistent virtual world.

Skittish spaces initially accommodated up to 120 mixed voices in a single room, but the audio capacity is even higher now. Though he’s still testing what the new limits might be, Skittish is getting closer to Baio’s goal of hosting 1000-person events. Skittish rooms can now be password protected, invite-only or public, and Baio imagines special “cozy” 3-5 person spaces in the project’s future.

For its high quality spatial audio chat, Skittish uses an API from High Fidelity, the latest project from Second Life creator Philip Rosedale. Amazingly, Second Life added spatial audio to its online worlds all the way back in 2007 — fourteen years ago.

Skittish will host its first paid events this month as a test, with invites to follow after that. Baio plans to rely on paid events for revenue and he’s on the fence about offering a free tier due to moderation concerns and the costs associated with hosting hundreds of simultaneous conversations between virtual elephants, zebras and raccoons.

A walk in Skittish

I met up with Baio in Skittish to chat about the project and it immediately felt less awkward than a Zoom call or Google Hangout. As a noble trash panda, I followed Baio’s owl around the colorful polygonal virtual set like we might have walked around a park having coffee together.

Skittish looks like a video game and you can move around using WASD if you want, but it’s straightforward enough that anyone can get the hang of it right away. The world’s simple graphical style sets a chill, creative vibe and the avatars even have a gentle idle animation, a kind of bounce that brings your respective elephant, raccoon or zebra to life.

Like experiences I’ve had in a few other innovative avatar-based virtual worlds (AltspaceVR comes to mind), the sense of really being there, just hanging out, feels revelatory. Multiplayer games have been miles ahead of traditional social networks on this phenomenon for ages; it’s no wonder that Fortnite and Minecraft are de facto social networks for a huge swath of younger people. In Skittish, the high quality spatial audio and playful sense of presence offer something similarly transporting.

Virtual owls aside, Baio says says Skittish will be a success when people start make real connections there that follow them beyond the virtual world he’s created.

“Just like the events I’ve run in real life, I’ll know it’s working when I hear stories about people meeting each other in a playful environment and making new friends,” Baio said.

News: Google updates Firebase with new personalization features, security tools and more

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

At its I/O developer conference, Google today announced a slew of updates to its Firebase developer platform, which, as the company also announced, now powers over 3 million apps. There’s a number of major updates here, most of which center around improving existing tools like Firebase Remote Config and Firebase’s monitoring capabilities, but there are

At its I/O developer conference, Google today announced a slew of updates to its Firebase developer platform, which, as the company also announced, now powers over 3 million apps.

There’s a number of major updates here, most of which center around improving existing tools like Firebase Remote Config and Firebase’s monitoring capabilities, but there are also a number of completely new features here as well, including the ability to create Android App Bundles and a new security tool called App Check.

“Helping developers be successful is what makes Firebase successful,” Firebase product manager Kristen Richards told me ahead of today’s announcements. “So we put helpfulness and helping developers at the center of everything that we do.” She noted that during the pandemic, Google saw a lot of people who started to focus on app development — both as learners and as professional developers. But the team also saw a lot of enterprises move to its platform as those companies looked to quickly bring new apps online.

Maybe the marquee Firebase announcement at I/O is the updated Remote Config. That’s always been a very powerful feature that allows developers to make changes to live production apps on the go without having to release a new version of their app. Developers can use this for anything from A/B testing to providing tailored in-app experience to specific user groups.

With this update, Google is introducing updates to the Remote Config console, to make it easier for developers to see how they are using this tool, as well as an updated publish flow and redesigned test results pages for A/B tests.

What’s most important, though, is that Google is taking Remote Config a step further now by launching a new Personalization feature that helps developers automatically optimize the user experience for individual users. “It’s a new feature of [Remote Config] that uses Google’s machine learning to create unique individual app experiences,” Richards explained. “It’s super simple to set up and it automatically creates these personalized experiences that’s tailored to each individual user. Maybe you have something that you would like, which would be something different for me. In that way, we’re able to get a tailored experience, which is really what customers expect nowadays. I think we’re all expecting things to be more personalized than they have in the past.”

Google is also improving a number of Firebase’s analytics and monitoring capabilities, including its Crashlytics service for figuring out app crashes. For game developers, that means improved support for games written with the help of the Unity platform, for example, but for all developers, the fact that Firebase’s Performance Monitoring service now processes data in real time is a major update to having performance data (especially on launch day) arrive with a delay of almost half a day.

Firebase is also now finally adding support for Android App Bundles, Google’s relatively new format for packaging up all of an app’s code and resources, with Google Play optimizing the actual APK with the right resources for the kind of device the app gets installed on. This typically leads to smaller downloads and faster installs.

On the security side, the Firebase team is launching App Check, now available in beta. App Check helps developers guard their apps against outside threats and is meant to automatically block any traffic to online resources like Cloud Storage, Realtime Database and Cloud Functions for Firebase (with others coming soon) that doesn’t provide valid credentials.

The other update worth mentioning here is to Firebase Extensions, which launched a while ago, but which is getting support for a few more extensions today. These are new extensions from Algolia, Mailchimp and MessageBird, that helps bring new features like Algolia’s search capabilities or MessageBird’s communications features directly to the platform. Google itself is also launching a new extension that helps developers detect comments that could be considered “rude, disrespectful, or unreasonable in a way that will make people leave a conversation.”

News: Google Cloud launches Vertex AI, a new managed machine learning platform

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

At Google I/O today Google Cloud announced Vertex AI, a new managed machine learning platform that is meant to make it easier for developers to deploy and maintain their AI models. It’s a bit of an odd announcement at I/O, which tends to focus on mobile and web developers and doesn’t traditionally feature a lot

At Google I/O today Google Cloud announced Vertex AI, a new managed machine learning platform that is meant to make it easier for developers to deploy and maintain their AI models. It’s a bit of an odd announcement at I/O, which tends to focus on mobile and web developers and doesn’t traditionally feature a lot of Google Cloud news, but the fact that Google decided to announce Vertex today goes to show how important it thinks this new service is for a wide range of developers.

The launch of Vertex is the result of quite a bit of introspection by the Google Cloud team. “Machine learning in the enterprise is in crisis, in my view,” Craig Wiley, the director of product management for Google Cloud’s AI Platform, told me. “As someone who has worked in that space for a number of years, if you look at the Harvard Business Review or analyst reviews, or what have you — every single one of them comes out saying that the vast majority of companies are either investing or are interested in investing in machine learning and are not getting value from it. That has to change. It has to change.”

Wiley, who was also the general manager of AWS’s SageMaker AI service from 2016 to 2018 before coming to Google in 2019, noted that Google and others who were able to make machine learning work for themselves saw how it can have a transformational impact, but he also noted that the way the big clouds started offering these services was by launching dozens of services, “many of which were dead ends,” according to him (including some of Google’s own). “Ultimately, our goal with Vertex is to reduce the time to ROI for these enterprises, to make sure that they can not just build a model but get real value from the models they’re building.”

Vertex then is meant to be a very flexible platform that allows developers and data scientist across skill levels to quickly train models. Google says it takes about 80% fewer lines of code to train a model versus some of its competitors, for example, and then help them manage the entire lifecycle of these models.

The service is also integrated with Vizier, Google’s AI optimizer that can automatically tune hyperparameters in machine learning models. This greatly reduces the time it takes to tune a model and allows engineers to run more experiments and do so faster.

Vertex also offers a “Feature Store” that helps its users serve, share and reuse the machine learning features and Vertex Experiments to help them accelerate the deployment of their models into producing with faster model selection.

Deployment is backed by a continuous monitoring service and Vertex Pipelines, a rebrand of Google Cloud’s AI Platform Pipelines that helps teams manage the workflows involved in preparing and analyzing data for the models, train them, evaluate them and deploy them to production.

To give a wide variety of developers the right entry points, the service provides three interfaces: a drag-and-drop tool, notebooks for advanced users and — and this may be a bit of a surprise — BigQuery ML, Google’s tool for using standard SQL queries to create and execute machine learning models in its BigQuery data warehouse.

“We had two guiding lights while building Vertex AI: get data scientists and engineers out of the orchestration weeds, and create an industry-wide shift that would make everyone get serious about moving AI out of pilot purgatory and into full-scale production,” said Andrew Moore, vice president and general manager of Cloud AI and Industry Solutions at Google Cloud. “We are very proud of what we came up with in this platform, as it enables serious deployments for a new generation of AI that will empower data scientists and engineers to do fulfilling and creative work.”

News: Google adds foldable-focused Android developer updates

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

Things have been a bit quiet on the foldables front of late, but plenty of parties are still bullish about the form factor’s future. Ahead of today’s big I/O kickoff, Samsung (undoubtedly the most bullish of the bunch) posted a bunch of metrics this morning, noting, The global outlook is just as impressive. This year

Things have been a bit quiet on the foldables front of late, but plenty of parties are still bullish about the form factor’s future. Ahead of today’s big I/O kickoff, Samsung (undoubtedly the most bullish of the bunch) posted a bunch of metrics this morning, noting,

The global outlook is just as impressive. This year alone, the foldables market is expected to triple over last year — a year in which Samsung accounted for three out of every four foldable smartphones shipped worldwide.

Part of anticipating growth in the category is ensuring that the software is ready it. Samsung has been tweaking things for a while now on its end, and at I/O in 2018, Google announced that it would be adding support for foldable screens. Recent rumors have suggested that the company is working on its own foldable Pixel, but even beyond that, it’s probably in the company’s best interest to ensure that Android plays nicely with the form factor.

“We studied how people interact with large screens,” the company said in today’s developer keynote. This includes a variety of different aspects, including where users place their hands while using the device — which can be a bit all over the place when dealing with different applications in different orientations and form factors. Essentially, you don’t want to, say, put buttons where people generally place your hands.

The list of upgrades includes the ability to resize content automatically, without overly stretching it out to fit multiple panels. All of this is no doubt going to be a learning curve as foldables end up in the hands of more users. But at very least, it signals Google’s continued view of foldables as a growing category. It’s also one of multiple updates today that involve the company working more closely with Samsung.

The two tech giants also announced a joint Wear OS/Tizen play early today.

News: Google updates its cross-platform Flutter UI toolkit

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

Flutter, Google’s cross-platform UI toolkit for building mobile and desktop apps, is getting a small but important update at the company’s I/O conference today. Google also announced that Flutter now powers 200,000 apps in the Play Store alone, including popular apps from companies like WeChat, ByteDance, BMW, Grab and DiDi. Indeed, Google notes that 1

Flutter, Google’s cross-platform UI toolkit for building mobile and desktop apps, is getting a small but important update at the company’s I/O conference today. Google also announced that Flutter now powers 200,000 apps in the Play Store alone, including popular apps from companies like WeChat, ByteDance, BMW, Grab and DiDi. Indeed, Google notes that 1 in 8 new apps in the Play Store are now Flutter apps.

The launch of Flutter 2.2 follows Google’s rollout of Flutter 2, which first added support for desktop and web apps in March, so it’s no surprise that this is a relatively minor release. In many ways, the update builds on top of the features the company introduced in version 2 and reliability and performance improvements.

Version 2.2 makes null safety the default for new projects, for example, to add protections against null reference exceptions. As for performance, web apps can now use background caching using service workers, for example, while Android apps can use deferred components and iOS apps get support for precompiled shaders to make first runs smoother.

Google also worked on streamlining the overall process of bringing Flutter apps to desktop platforms (Windows, macOS and Linux).

But as Google notes, a lot of the work right now is happening in the ecosystem. Google itself is introducing a new payment plugin for Flutter built in partnership with the Google Pay team and Google’s ads SDK for Flutter is getting support for adaptive banner formats. Meanwhile, Samsung is now porting Flutter to Tizen and Sony is leading an effort to bring it to embedded Linux. Adobe recently announced its XD to Flutter plugin for its design tool and Microsoft today launched the alpha of Flutter support for Universal Windows Platform (UWP) apps for Windows 10 in alpha.

News: Google launches the first beta of Android Studio Arctic Fox

• May 18, 2021
• by:   
• in:  Blog 
• note:  no comments 

At its I/O developer conference, Google today announced the first beta of the next version of its Android Studio IDE, Arctic Fox. For the most part, the idea here is to bring more of the tooling around building Android apps directly into the IDE. While there is a lot that’s new in Arctic Fox, maybe

At its I/O developer conference, Google today announced the first beta of the next version of its Android Studio IDE, Arctic Fox. For the most part, the idea here is to bring more of the tooling around building Android apps directly into the IDE.

While there is a lot that’s new in Arctic Fox, maybe the marquee feature of this update is the integration of Jetpack Compose, Google’s toolkit for building modern user interfaces for Android. In Android Studio, developers can now use Compose Preview to create previews of different configurations (think themes and devices) or deploy a preview directly to a device, all while the layout inspector makes it easier for developers to understand how (and why) a layout is rendered the way it is. With Live Updates enabled any change is then also directly streamed to the device.

The team also integrated the Android Accessibility Test Framework directly into Android Studio to help developers find accessibility issues like missing content descriptions or a low contrast in their designs.

Just like with some of the updates to Android itself, the team is also looking at making it easier to develop for a wider range of form factors. To build Wear OS apps, developers previously had to physically connect the watch to their development machine or go through a lot of steps to pair the watch. Now, users can simply pair a watch and phone emulator (or physical phone) with the new Wear OS Pairing feature. All this takes now is a few clicks.

Also new on the Wear OS side is a new heart rate sensor for the Wear OS Emulators in Android Studio, while the Android Automotive emulator gains the ability to replay car sensor data to help those developers with their development and testing workflow.

Android Studio users who work on a Mac will be happy to hear that Google is also launching a first preview of Android Studio for the Apple Silicon (arm64) architecture.