August 2017 Summaries

Alejandro (Alex) Jaimes, DigitalOcean's Head of R&D, discusses the importance of leveraging existing public models and combining public and proprietary datasets to gain a competitive advantage in AI. Not all companies have access to vast amounts of data, but they can still benefit from using existing models and strategies to simplify the manual labeling process and create synthetic datasets. The development of AI models requires significant efforts, and many companies focus on specific verticals or functionalities that work well in practice rather than building a "human-like" AI system. As the AI industry advances, we can expect to see more functionalities coming from vendors and open source initiatives, with cloud-based services becoming increasingly popular due to their ability to quickly scale and accommodate constant changes.

Dave "Dizzy" Smith, senior director of engineering at DigitalOcean, discusses the challenges of managing distributed teams and offers strategies for creating environments where people can do their best work. He emphasizes the importance of effective communication, which he breaks down into three dimensions: asynchronous (e.g., email), synchronous (e.g., video conferencing), and face-to-face. To counteract the forces that make it difficult for distributed teams to unite around a purpose, Smith suggests staying aligned by sharing a contextual view of the business world, staying in touch through regular check-ins and scheduling, and keeping an eye on the horizon by avoiding getting overwhelmed by details. By focusing on creative effective communication practices, managers can foster communication and create environments that support success.