Episode Summary
Executive Summary: Daniel Whitenack and Chris Benson discuss how to set practical AI learning goals for 2019. They recommend choosing resources that fit a learner’s goals and background, balancing theory with coding practice, and applying new skills through personally meaningful projects. They also emphasize peer communities, meetups, conferences, and freely available tools and courses as ways to keep learning in a fast-moving field.
Main Topics: Choose learning resources to fit your goals: The hosts distinguish research-focused learning from practical implementation and argue that learners should first decide what kind of AI work they want to pursue. They suggest matching resources to existing experience and weaknesses, such as building mathematical foundations or gaining coding practice. Research careers and formal education: A PhD is not a general requirement for working in AI. The hosts recommend pursuing doctoral study when research itself is a passion, while noting that programs such as the OpenAI Fellows Program, Google AI Residency, and Allen AI Institute pre-doctoral internship can offer alternative research pathways. Hosts’ 2019 learning resolutions: Daniel plans to deepen his understanding of deep-learning methods for NLP, using resources including NLP with PyTorch, a Udacity PyTorch course, and TensorFlow-focused videos. Chris plans to strengthen the connection between software engineering and AI through AWS machine learning, Google’s machine learning crash course, and hands-on robotics projects. Learning by building meaningful projects: The hosts say books and courses are not enough: learners should apply new skills in projects tied to their interests. Daniel proposes exploring unsupervised machine translation for low-resource languages; Chris describes projects involving animal advocacy, computer vision, robotics, and activities with his daughter. Finding data and project opportunities: The discussion recommends looking to AI-for-good organizations such as DrivenData and DataKind for project ideas and competitions, reusing or adapting research code from GitHub, and using Google Dataset Search to locate public datasets, including combinations of datasets for a use case. Community as a learning resource: Meetups, conferences, and online communities can provide relationships and conversations that complement presentations and course materials. Chris highlights the Atlanta Deep Learning Meetup, while Daniel hopes to engage with NLP research communities and conferences.
Key Arguments: Learning should begin with a clear goal: research and practical application are both valuable, but they call for different paths and resources. A PhD is not necessary for most AI work; it is most appropriate for people motivated to conduct research, while fellowships and residencies can provide other routes into research settings. Resources should address the learner’s weaker side. Developers may need more mathematics and theory, while people from scientific or academic backgrounds may benefit most from coding practice and practical projects. Courses and books provide a foundation, but hands-on projects help knowledge stick and reveal how methods work in real situations. Personal interest can sustain learning and make AI useful beyond work, whether the project concerns language access, animal welfare, robotics, or family activities. Data discovery and human connection are important parts of applied AI: public dataset search tools can broaden project possibilities, while meetups and conferences help learners build useful relationships.
Data Points: Target year for the hosts’ resolutions: 2019 - The episode focuses on learning goals and resources for the coming year. NLP with PyTorch book release date: January 2019 - Daniel identifies the forthcoming O’Reilly book as a resource for studying deep-learning approaches to NLP. Chris’s AWS usage history: Since 2010 - He cites his long-standing use of AWS as a reason to study Amazon’s machine-learning offerings. Atlanta Deep Learning Meetup launch: Beginning of 2017 - Chris says he started the meetup to create a community focused specifically on deep learning. Atlanta Deep Learning Meetup membership: About 2,200 people - Chris gives the approximate size of the group at the time of the episode. Typical attendance at a meetup event: About 100 people - Chris says events commonly draw this many attendees. Meetup event frequency: Pretty much every month - The Atlanta group holds events on a near-monthly basis.
Pivotal Quotes: "You don't need a PhD to work in AI." — Daniel Whitenack: He clarifies that doctoral study is not a prerequisite for entering the AI field. "At some point, you need to actually apply what you're learning and work on kind of a side project or an interesting project of some type." — Daniel Whitenack: He explains why project work is an essential complement to books and courses. "Getting in a room and talking to other people who share that passion and interest from other organizations or academia, there is so much value in those conversations and creating those relationships" — Chris Benson: He argues that the relationships formed at meetups can be as valuable as the formal presentations.
Implications: AI learners can make steady progress without following a single credential-driven path: define a goal, select resources suited to their background, build something they care about, and learn with peers. For the field, accessible tools and communities can widen participation and help translate research into practical applications.