zoiaorg

Tech, Strategy, and other interesting things by Roberto Zoiaby Roberto Zoia

Strategy and Technology

Short-term Effectiveness vs Long-term Resilience

by Leave a Comment

Seth Godin, writing about Self, Community and Motivation, explains that in the short term, you can get people to do what’s needed by replacing a “rational, generous, community mindset” with “an immediate and self-focused desire to be safe.”

In the current pandemic, public health officials didn’t appeal to you should wear masks to protect strangers from getting sick, but by implying “if you touch someone, you will die almost instantly and quite horribly. And people, already frightened, embraced the feeling.”

(…) short-term market needs are often efficiently filled by short-term selfish behavior. Resilience comes from a longer-term and more community-focused outlook.

The question is: Once people catch the virus and get through it (as most people will) and recover (as more than 9 out of 10 will), what will replace the selfish panic?

Resorting to short-term motivation can be, in some circumstances, an effective hack. If you resort to it too often, it teaches people to be selfish in the long-term.

Un paso cada día

by Leave a Comment

Me gusta este comercial de Nike porque el mensaje de la héroe no es solamente un poderoso “sí puedo”. No se trata de un voluntarismo vacío. Sus obstáculos son reales y patentes, y la meta que se propone ambiciosa incluso sin sus limitaciones. Y es claro que no lo logró de la noche a la mañana.

Sí podemos. Todos podemos. Un paso adelante cada día.

One Step Forward Every Day

by Leave a Comment

I like this Nike commercial because the message of the hero is not just a powerful “I can”. It’s not pure voluntarism. Her obstacles are real. And her goal is a goal that would be difficult even without her limitations. And it’s clear that she didn’t achieve her dreams in one day.

Yes, we can. We all can. Step by step.

Amazon, las tiendas sin cajero y la carrera por al supremacía en AI

by Leave a Comment

Amazon acaba de anunciar que licenciará su tecnología de tiendas sin cajero a otros operadores de retail y de tiendas de conveniencia bajo el nombre de Just Walk Out. Just Walk Out es tecnología que se apoya fuertemente en deep learning, sensores, y reconocimiento de video.

Abrir tiendas físicas es una operación intensiva en capital y de bajo margen neto. Si Amazon hubiera seguido ese camino, como ha hecho con sus 26 tiendas Amazon Go, hubiera tenido además que operar las tiendas, en un mercado en el que los competidores tienen años de experiencia. En cambio, licenciar la tecnología a otros operadores le permitirá mayores márgenes sobre una facturación menor, y una penetración más rápida, además de convertir a sus competidores en potenciales clientes.

Pero hay un punto más importante. Como dice Enrique Dans, los datos se han convertido en el nuevo elemento que define la competitivdad. En deep learning no hay mejor data que más data1, y en la carrera por la supremacía en Inteligencia Artificial, el que tenga más data para entrenar sus algoritmos de deep learning llevará la delantera.

Las tiendas que implementen Just Walk Out generarán una cantidad importante de data para Amazon. Si bien Amazon ha dicho “solo” recolectará la data indispensable para generar el recibo de pago del cliente, esto incluye la información de sensores, videocámaras y demás que necesite para operar el sistema. Este volumen de información es significativo.

En otras palabras, para Amazon no se trata solo de crecer una línea de negocio licenciando la tecnología, sino de armar un circuito en que la misma operación de las tiendas que implementen Just Walk Out recolectará data de campo que permita entrenar mejor al sistema, para beneficio de Amazon y sus clientes. Esto le dará a Amazon una ventaja significativa sobre sus competidores.

¿Debería quizá Waymo, la división de vehículos de conducción autónoma de Google/Alphabet seguir una estrategia similar? ¿No sería mejor que licenciara su tecnología en vez de seguir construyendo su propia flora de automóviles, lo que le permitiría, en definitiva, aumentar la velocidad de recolección de datos para entrenar sus sistemas de AI?

(Este artícilo también está disponible en inglés.)

Featured image photo by SounderBruce – Own work, CC BY-SA 4.0.

  1. cfr Kai-Fu Lee, AI Superpowers: China, Silicon Valley, and the New World Order. Kai-Fu Lee is the founder of Sinovation Ventures, founder director of Microsoft Research Asia, and past-president of Google China.  ↩

Amazon, cashier-less stores, and the race for AI supremacy

by Leave a Comment

Amazon just announced that it is licensing Just Walk Out, its cashier-less convenience store technology, to other convenient stores and retailers. Just Walk Out is a technology that relies heavily on deep learning (AI), sensor fusion, and visual recognition.

Opening brick and mortar stores is not only capital-intensive, but has low net margins. If Amazon had chosen to continue growing this business lines this way, like it has with its 26 Amazon Go stores, it would also have had to operate the stores. On the other hand, licensing its technology to third parties make sense because it allows faster penetration and higher margins on lower revenues. Also, it converts former competitors into potential customers.

But there is more. As Enrique Dans writes, _the new element that defines competitiveness today is data. In deep learning, there’s no data like more data1. In the race for AI supremacy, the company who has more data to train its deep learning algorithms has the lead.

The stores that implement Just Walk Out will generate a significant amount of data for Amazon. Even if Amazon has said that it “only” will collect the data needed for generating the customer’s receipt, this included detailed data from sensors, videocameras, and other information needed to run the system. The volume of information is significant.

For Amazon, Just Walk Out is not just the natural growth of one of its business lines. It’s about building a flywheel where the operation of the stores that use Amazon’s technology will generate the data that will be used to train the system itself, in benefit of Amazon and its clients. This will give Amazon a significant head start over its competitors in this field.

What about Waymo, Google’s self-driving car project? Wouldn’t it be better for Waymo if instead of trying to build its own car fleet, it would license its technology to car makers?

(This article is also available in Spanish.)

(Featured image photo by SounderBruce – Own work, CC BY-SA 4.0.)

  1. cfr Kai-Fu Lee, AI Superpowers: China, Silicon Valley, and the New World Order. Kai-Fu Lee is the founder of Sinovation Ventures, founder director of Microsoft Research Asia, and past-president of Google China.  ↩