A Blog by Jonathan Low


Jan 18, 2018

Are Solar Roads the Highway of the Future - Or a Dead End?

Technically feasible but probably not practical until performance of solar road panels improves, wireless transmission of power becomes scalable and costs come down significantly. In other words, a definite possibility for the future, but not immediately. JL

Thomas Hornigold reports in Singularity Hub:

Given that the land is already covered, environmental disruption is lower. The power stations don’t suffer from remoteness problems. With LEDs, you have efficient street-lighting. With wireless power transmission, cars could be powered by roads. (But) the cost of replacing US roadways with solar roads would come in at $56 trillion. Chinese transparent asphalt can withstand 10 times more pressure than normal asphalt. If you spend $56 trillion on solar roads, you’d cover 7.5 x 1010 m2 with panels, generate 600 GW of electrical power - the electricity consumed by the US today.

A Popular Crime-Predicting Algorithm Doesnt Perform Better Than Random, Untrained Humans

There could be problems with the design, the data, the preprogrammed bias of the coders - or perhaps more importantly -  the assumption that algorithms are inherently superior to humans at all forms of decision-making. JL

Issie Lapowsky reports in Wired:

Risk assessment algorithms, used in states from California to New Jersey, crunch data about a defendant’s history—like age, gender, and prior convictions—to help courts decide who gets bail, who goes to jail, and who goes free. Underlying the conversation about algorithms was this assumption that algorithmic prediction was inherently superior to human prediction. "There was no difference between people responding to an online survey for a buck and this commercial software being used in the courts."

Lego Links With Tencent To Develop Online Games For China - And Elsewhere?

Despite efforts to connect its tangible products with the increasingly intangible, screen-oriented focus of childrens' play, Lego has struggled to convert its strategy into profits.

The partnership with Chinese tech giant Tencent may help jumpstart that initiative, but the larger question may be what Lego has to surrender in terms of intellectual property and future direction to do so. JL

Thuy Ong reports in The Verge, image from Classic-castle.com:

The partnership will include the creation of a dedicated area for kids to watch Lego videos on Tencent’s platform, as well as the development of Lego-branded games. The Danish toymaker will also publish and operate Lego Boost, a simple, programmable, and connected robotics kit released in June last year. The two companies will explore co-developing a Chinese version of the social networking platform Lego Life, which is already available in other countries.

General Motors' Driverless Cars Could Be In Use Next Year

Moving fast - and hopefully not breaking things...JL

Neal Boudette reports in the New York Times:

G.M. submitted a petition to the United States Department of Transportation seeking permission to begin operating fully autonomous cars — without steering wheels or pedals — in a commercial ride-hailing service next year. The vehicle, the Cruise AV, could be put into production on a standard assembly line once approval was granted by the federal government and states where the cars would operate.

Following News Feed Change, Facebook's Loss Is Twitter's Gain, But Analysts Are Bullish

There has been growing concern among securities analysts that social media generally and Facebook specifically face significant risk of government regulation due to the growing evidence that the indiscriminate distribution of false or misleading news over their networks has influenced socio-economic and political outcomes.

While Twitter's stock has risen relatively as a potential competitive beneficiary, Facebook's has remained fairly steady because of the perception among analysts that the risk of regulation may now decrease. JL

Janko Roettgers reports in Variety and Stock News reports:

Twitter’s stock price rose just as Facebook lost 4.5% following CEO Mark Zuckerberg’s announcement that the company would tweak its news feed. Twitter is benefitting from the notion that Facebook is closing the door on news since Twitter has spent much of the past two years to reinvent itself as the place where users go to “see what’s happening.” (But some Wall Street analysts) "are positive about these changes because we believe they will be economically neutral and they lower the risk that FB gets regulated.”

IBM and Maersk Joint Venture Uses Blockchain To Optimize Supply Chain Efficiency

The significance of the announcement is that this is a practical application of blockchain technology by two of the leaders in their respective fields - tech and logistics - which means their experience should create a more productive, glitch-free process for those who use it - and, potentially, a first mover advantage for themselves. JL

Richard Milne reports in The Financial Times:

The Danish shipping conglomerate and US (tech co.) estimate businesses spend a fifth of the cost to transport goods on processing documents and administration. The distributed ledger technology behind blockchain create(s) an unchangeable record of transactions along a supply chain shared in real time with whichever companies are necessary. The significance is huge for any transaction that has multiple parties. Blockchain would allow companies at different stages of the supply chain to see the information they need about each transaction. It could set a standard for the digitalisation of supply chains.

Jan 17, 2018

Google To Make Building Artificial Intelligence Algorithms As Easy As Drag and Drop

Smart strategy. Expand access for AI in order to grow the market - and associated revenues for Google. JL

Dave Gershgorn reports in Quartz:

Cloud AutoML (is) a cloud service that allows average folks to train a custom AI algorithm. The service will start with still images, but Google plans to expand this to all of its AI offerings. Google is trying to expand the group of people who can access and use AI software, rather than offer new capabilities. People who don’t know how to code can build and deploy an algorithm on their own. AutoML edits a customer’s algorithm to optimize it for a specific task, like recognizing images of car models or yogurt brands.