Week 7

 

Part One: Team Final Research Video Project Planning

Our team communicated and divided tasks effectively for the final video project. We used Trello to stay organized, assigning roles and setting deadlines for each portion of the presentation. Meetings were held over Google Meet, which made it easy for us to share progress and brainstorm together in real time. For our slides we used Canva because it provides clean, professional templates and supports collaborative editing.

Overall, the collaboration went smoothly. To improve future large collaborative projects I would schedule a few more brief in person meetings to build rapport and momentum. I would also use shorter, focused check ins rather than longer meetings to keep us on task and respectful of everyone’s time. Finally, setting clearer expectations early for research and presentation sections helped reduce confusion and improved workflow.

Part Two: Reflection on TED Talks and the Harvard Podcast

This week I reflected on three sources: the TED Talk titled "AI is Dangerous, but not for the Reasons You Think" by Sasha Luccioni, the TED Talk "Grit: The Power of Passion and Perseverance" by Angela Duckworth, and the Harvard Thinking podcast episode "The Promise and Peril of AI." Each offered distinct but complementary insights about technology, ethics, and personal development.

Sasha Luccioni, AI is Dangerous but not for the Reasons You Think

Sasha Luccioni redirected attention away from speculative doomsday scenarios to the current, tangible harms AI is creating. Key points that stood out to me include the environmental cost of training large models, the lack of consent when creative works are used as training data, and the ways model outputs reproduce social bias. The environmental examples were striking: training one open model consumed the equivalent energy of 30 homes for a year and emitted about 25 tons of carbon dioxide, while some models can emit many times more. Luccioni’s practical response was to develop measurement tools such as CodeCarbon to estimate energy and emissions, and to support artist led tools like Have I Been Trained to help creators see whether their work was included in training sets. She also developed the Stable Bias Explorer to surface patterns of representation in image generation systems. Her call to measure impacts, disclose them, and create governance tools made me think concretely about responsibilities I will have as a software engineer to design and choose models that respect both people and the planet.

Angela Duckworth, Grit: The Power of Passion and Perseverance

Angela Duckworth’s research emphasizes that sustained effort and long term passion predict success more reliably than raw talent or IQ. Grit is defined as the stamina to stick with goals over years, and Duckworth showed this trait matters across many domains, from military training to spelling bees to classroom persistence. What resonated for me was the practical implication that skills and success improve through sustained practice and mindset work, not only through innate ability. Duckworth highlights growth mindset interventions as one promising approach for building perseverance, though she also notes that the science of how to reliably build grit is still developing. As a student and future engineer I took away a renewed commitment to consistent practice, to embracing failure as a learning opportunity, and to applying small, measurable changes that reinforce persistence.

Harvard Thinking Podcast, The Promise and Peril of AI

The Harvard podcast offered a broad, nuanced discussion about how AI’s promises and risks are often uncertain and context dependent. A few themes stood out. First, the speakers argued that ethics must be embedded throughout the design process rather than treated as an add on. Second, they emphasized the importance of observation and monitoring as AI systems scale into social institutions, using the metaphor of a car with brakes, accelerator, dashboard, and airbags to illustrate needed safeguards and visibility. Third, the episode discussed inequality and power, noting that while AI can sometimes empower less skilled workers, it can also concentrate benefits for those who control capital and data. Finally, the podcast stressed the complementary roles of academia and industry, and the need for public investment in oversight and monitoring that matches the scale of technological change.

How I relate...

Taken together these three pieces shaped my thinking in two main ways. On the technical and societal side, Sasha Luccioni and the Harvard panel reinforced that measuring impact matters. As a future software engineer I will prioritize transparent measurement of environmental, social, and fairness impacts and advocate for tools and practices that make those impacts visible to teammates and stakeholders. I will support embedded ethics in development cycles and push for monitoring and remediation strategies when systems are deployed.

On the personal side, Angela Duckworth’s work on grit complements the technical obligations by reminding me that long term improvement requires consistent effort and resilience. I want to adopt small habits that build persistence, iterate on solutions when they fail, and keep learning about both engineering best practices and ethical design.

In project work I will combine these lessons by advocating for concrete impact metrics when we choose models, by scheduling regular, focused team check ins to maintain momentum, and by committing to the steady practice required to build the skills and judgment needed to do this work well.