A Blog by Jonathan Low

 

Sep 19, 2018

New AI Strategy Mimics How the Brain Learns To Smell

Using all the human senses to figure out how to better empower AI. JL


Jordana Cepelewicz reports in Quanta:

Scientists trying to gain a better understanding of how organisms process chemical information have uncovered coding strategies that seem especially relevant to problems in AI. Olfactory circuits bear striking similarities to more complex brain regions that have been of interest in the quest to build better machines. Computer scientists are now beginning to probe those findings in machine learning contexts.

Future Robot Taxis Could Charge Themselves

At least one clever solution to a problem inhibiting growth in of electric cars. JL


Elizabeth Woyke reports in MIT Technology Review:

There are only 20,000 charging stations for electric vehicles in the US. (There are more than 125,000 gas stations for conventional cars.) Major economies (China, Europe, India, and the US) will need to invest $55 billion in charging infrastructure by 2031 to support the 140 million electric vehicles that will be on the road then. Wireless charging called magnetic resonance takes energy from the electric grid, feeds it into a copper coil on the ground, creating a magnetic field. When a coil attached to the bottom of the car comes within range of this field, an electric current is generated, to charge the car’s battery

Google's Chinese Search To Censor Results, Provide Only Gov't Approved Data

Don't be evil versus don't miss a billion new customers. JL

Tim Cushing reports in Tech Dirt:

Google's prototypes of the search engine linked the search app on a user’s Android smartphone with their phone number. This means individual people’s searches could be easily tracked – and any user seeking out information banned by the government could potentially be at risk. (It) contains a blacklist of terms like "human rights," "student protest," and "Nobel Prize." Dragonfly was “ hardcoded to force [Chinese-provided] data.”

Home Buying Goes Algorithmic

As anyone who has bought or sold a home is likely to say, 'if only it were that easy.' JL 

Ryan Dezember reports in the Wall Street Journal:

Adapting (a) system for valuing mortgage-backed securities to churn out acquisition leads, estimate renovation costs and predict rental yields for a plan to use pattern-recognition software to identify mispriced homes. Besides screening for quantifiable characteristics like age, number of rooms, square footage, school district, property taxes and flood-zone status, it attempts to measure qualitative aspects and uses algorithms to predict future value. (But) “going from 40,000 houses to 12 is a machine problem.Going from 12 to one is a human problem.”

Amazon Is Stuffing Research Result Pages With Ads - And They're Working

Pay to play goes digital. JL

Rani Molla reports in Re/code:

Sponsored ads allow vendors to bid to have their products show up when consumers type in a related search term. “Nobody is scrolling beyond the first page when they do a search.” If you’re Duracell, you can pay to have your product show up above search results when someone types in “batteries.” Spending on sponsored products in Amazon’s search increased 165% in the second quarter of 2018. The competition for brands to bid on their own or others’ keywords is leading toward an “escalating race where the trends are to spend more money to buy more ads to have better visibility on Amazon.”

Why We Need To Update Financial Reporting For the Digital Era

Value is increasingly more intangible than tangible. And financial reporting fails to capture it. That leads to misallocation of resources, misinterpretation of crucial data and misrepresentation of performance. JL

Vijay Govindarajan and colleagues report in Harvard Business Review:

Financial statements fail to capture the value created by digital companies. The market caps of Apple, Alphabet, Amazon, and Microsoft exceed $3 trillion. Their combined assets of $944 billion are an order of magnitude lower than the combined assets of the largest 3,177 companies in 1986. Financial capital is assumed to be unlimited, while human capital is in short supply. Risk is now a feature, not a bug. Investors are paying more attention to ideas and options than to earnings. Accounting is no longer considered a value-added function. Analysts increasingly rely on non-GAAP metrics. 

Sep 18, 2018

Fixing Fast Fashion's Waste - With Biodegradable Clothes

'Fast fashion' is efficient because it creates new products quickly in response to fads or early successes, but it also generates uses a lot of water, chemicals and other materials, while creating tremendous waste due to turnover.

The innovative use of materials created from natural organisms may provide the same speed - but eliminate the cost and excess that fast processes produce. JL


Erica Cirino reports in Scientific American:

When a piece of clothing wears out or goes out of fashion, it gets tossed in the trash; clothes made up 9% of all solid waste produced in the U.S. Innovators are using live organisms to grow biodegradable textiles, creating environmentally friendly materials in the laboratory—and producing near-complete items without factory assembly. Organisms can be grown to fit molds, producing the precise amount of textile needed and dyes from seeds and leaves eliminates waste at the production phase.