Tech / Telecom news — 21 Nov 2016



A new report from Adobe concludes that app installs have almost stopped growing, with only +6% yoy, in contrast with launches of the most popular, already installed apps, which keep increasing (at +62% yoy for the Top 20). Consistently with this, many apps are being discarded very quickly after the first install (Story)

Devices / Wearables

Snapchat’s “Spectacles” start to look like a success already, according to press comments. They’ve just been launched in selected locations, through Minion-looking vending machines, but they seem to be generating demand momentum, just what the company needs before its coming IPO (for which docs were recently filed) (Story)


Startup Trim has built a bot aimed at negotiating better rates with utility companies’ customer reps, and they’re testing it as a tool to reduce the bill at (beloved provider) Comcast, with 70% success and an average reduction of $10/month. Trim suggests this is just the beginning, and plans to keep improving functionalities (Story)


BT’s CEO claims, quite expectedly, that current uncertainty on a potential Openreach structural separation is bad for UK’s economy, as they’re “ready” to invest £6bn over next three years and this process is distracting them. BT is still negotiating with Ofcom, proposing alternative ways to keep Openreach at “arm’s length” (Story)

French startup Sigfox, whose almost immediate death had been predicted by the usual vendor-biased conservative network pundits, just got $160m more of financing, with Salesforce among the new investors. This seems to be linked to a plan to expand their network technology to 60 more countries (Story)

Startup Starry, which aims to use radio access on millimetre waves to provide ultra high speed broadband and compete vs cable, keeps plans to start commercial service in 2017, and just invited some journalists to a pilot deployment in Boston to serve 1,000 customers with a single station, covering an area of 1.5Km radius (Story)


It may sound like tech innovation is advancing fast, but the analysis here suggests we may be stuck with respect to long term innovation enablers, as most recent advances would depend on breakthroughs from 40–50 years ago. Author links this to a shift from (long term) public funded research, to (short term) corporate R&D (Story)


Are Facebook’s actual AI capabilities too weak? Either that or they don’t really want to solve the fake news problem, according to some comments. Facebook just described work under way for this, under public pressure, including “social signalling” of inaccurate content and (of course) improving the AI algorithms (Story)

Meanwhile Google, viewed as the AI leaders, just opened an experimental site to reinforce this public perception. The site lets people “play” with deep learning algorithms that try to guess what you’re sketching (in real time) or compose music. This is also a promotion of the recently-added AI capabilities at Google’s Cloud (Story)


Facebook’s management might be not so confident in future revenue growth, as they revealed a plan to buy back shares for up to $6bn. The plan will start in 1Q17 but doesn’t have an expiration date (so it will be open for Facebook to execute flexibly, depending on market conditions and alternative investment opportunities) (Story)