Technology

Former TokBox head Ian Small is replacing Chris O&Neill as CEO of Evernote, the note-taking and productivity app company said this morning. In a blog post, Small said that the leadership change was announced to employees this morning by Evernoteboard. &We are all hugely appreciative of the energy and dedication Chris has shown over the last three years, and in particular for putting Evernote on solid financial footing so we can continue to build for the future,& he wrote.

Small added, &When Stepan Pachikov founded Evernote, he had a vision for how technology could augment memory and how an app could change the way we relate to information at home and at work. Evernote has been more successful at making progress towards Stepandreams than he could have imagined, but Stepan and I both think that there is more to explore and more to invent.&

O&Neill had been EvernoteCEO since 2015, when he took over the position from co-founder Phil Libin. Small previously served as CEO of TokBox, which operates the OpenTok video calling platform, from 2009 to 2014, and then as its chairman from 2014 to July of this year.

O&Neilldeparture as CEO is the latest significant leadership shift for Evernote, which has withstood several key executive departures over the last few months. In early September, we reported that the company had lost several senior executives, including CTO Anirban Kundu, CFO Vincent Toolan, CPO Erik Wrobel, and head of HR Michelle Wagner, as it sought funding in a potential down-round from the unicorn valuation it hit in 2012. According to TechCrunchsources, Evernote had struggled to grow its base of paid users and active users, as well as enterprise clients, for the last six years.

Then a few weeks later, Evernote announced that had to lay off 54 people, or about 15 percent of its workforce. O&Neill wrote a blog post about the companyfuture growth strategy, including streamlining specific functions like sales so it could focus on product development and engineering.

Social e-commerce startup Goxip raised $5 million in January, and now the Hong Kong-based business has brought in more cash with a strategic $1.4 million investment from financial services company Convoy. Existing backers including Chinese photo app company Meitu also took part.

Convoy offers a range of services that include asset management, insurance and other investment options. Hong Konglargest financial advisory with over 100,000 customers, Convoy isn&t in great shape now. It has been in crisis over legal action and a corruption investigation that is centered around a former company director.

The companyshares remain suspended on the Hong Kong Stock Exchange although it recently made appointments aimed at modernizing its business and this deal is likely another part of that strategy. Convoyportfolio of strategic investments includes Nutmeg in the UKand IrelandCurrencyfair, which bought up Convoypayments arm.

Goxip said it will use the capital and the new relationship with Convoy to offer more installment-based financing options on its service, which is akin to a ‘shoppable Instagram& that has a focus on high-end fashion.

The company already counts major retailers likeNet-a-Porter, Harrods, and ASOS and brands that includelike Nike, Alexander McQueen and Topshop. To date, Goxip has helped customers find outfits and buy them but now, with Convoy, it wants to offer payment plans using a virtual credit card, Goxip co-founder and CEO Juliette Gimenez told TechCrunch.

With 600,000 monthly users and average orders of $300, Goxip is getting close to breaking even,Gimenez said, but she is hopeful that offering staggered payment options over varying periods such as 6-12 monthswill serve Goxip well as it expands in Southeast Asia where typical consumers spend less. That&ll happen soon after the company opened an office in Bangkok ahead of an imminent launch in Thailand, its second expansion after Malaysia.

Goxip has just opened an office in Bangkok ahead of an upcoming launch in Thailand

Beyond geographical additions,Goxip has also branched out into influencer marking this year with recently launchedRewardSnap service. Similar to Rewardstyle in the U.S, it will enableinternet influencers — and particularly those on Instagram — to partner with brands and make money through referrals to their audience.Gimenez said that 150 influencers with a total following of over 14 million have signed up, including Elly Lam who has 140,000 followers on Instagram

Instagram is beefing up its commerce focus — with the addition of a shopping tab and new management at the wheel — butGimenez said she isn&t phased. She points to the fact that Facebook, which owns Instagram, hasn&t been able to make e-commerce work in Asia, while the simplicity of Rewardsnap and its connection to the Goxip service, makes it highly defensible even asInstagram ups its shopping game.

Update: The original version of this article has been updated to reflect that RewardSnap has already launched, and that Elly Lam has 140,000 followers.

Google and a group of game cetologists have undertaken an AI-based investigation of years of undersea recordings, hoping to create a machine learning model that can spot humpback whale calls. Itpart of the companynew &AI for social good& program thatrather obviously positioned to counter the narrative that AI is mostly used for facial recognition and ad targeting.

Whales travel quite a bit as they search for better feeding grounds, warmer waters and social gatherings. But naturally these movements can be rather difficult to track. Fortunately, whales call to each other and sing in individually identifiable ways, and these songs can travel great distances underwater.

So with a worldwide network of listening devices planted on the ocean floor, you can track whale movements — if you want to listen to years of background noise and pick out the calls manually, that is. And thathow we&ve done it for quite a while, though computers have helped lighten the load. Googleteam, in partnership with NOAA, decided this was a good match for the talents of machine learning systems.

These AI (we employ the term loosely here) models are great at skimming through tons of noisy data for particular patterns, which is why they&re applied to voluminous data like that from radio telescopes and CCTV cameras.

In this case the data was years of recordings from a dozen hydrophones stationed all over the Pacific. This data set has already largely been investigated, but Googleresearchers wanted to see if an AI agent could do the painstaking and time-consuming work of doing a first pass on it and marking periods of interesting sound with a species name — in this case humpbacks, but it could just as easily be a different whale or something else altogether.

Spectrograms of whale song, left, an unknown &narrow-band& noise, center, and the recorderown hard disk drive, right.

Interestingly, but not surprisingly in retrospect, the audio wasn&t analyzed as such — instead, the audio was turned into images it could look for patterns in. These spectrograms are a record of the strength of sound in a range of frequencies over time, and can be used for all kinds of interesting things. It so happens that they&re also well studied by machine learning and computer vision researchers, who have developed various means of analyzing them efficiently.

The machine learning model was provided with examples of humpback whale calls and learned how to identify them with reasonable accuracy in a set of sample data. Various experiments were conducted to suss out what settings were optimal — for instance, what length of clip was easy to process and not overlong, or what frequencies could be safely ignored.

The final effort divided the years of data into 75-second clips, and the model was able to determine, with 90 percent accuracy, whether a clip contained a &humpback unit,& or relevant whale sound. Thatnot a small amount of error, of course, but if you trust the machine a bit you stand to save quite a bit of time — or your lab assistanttime, anyway.

A second effort relied on whatcalled unsupervised learning, where the system sort of set its own rules about what constituted similarity between whale sounds and non-whale sounds, creating a plot that researchers could sort through and find relevant groups.

Visualization of how the unsupervised model classified various sounds. The blue ones represent humpback calls.

It makes for more interesting visualizations but it is rather harder to explain, and at any rate doesn&t seem to have resulted in as useful a set of classifications as the more traditional method.

As with similar applications of machine learning in various scholarly fields, this isn&t going to replace careful observation and documentation but rather augment them. Taking some of the grunt work out of science lets researchers focus on their specialties rather than get bogged down in repetitive stats and hours-long data analysis sessions.