Human-Centered Tech: Finding Balance in the AI Revolution (Day 4 at SXSW)

I needed a break after diving into Quantum Computing, robots, and space for three straight days. Today was all about watching the keynotes from the comfort of my hotel room, giving my feet (and brain) a much-needed rest from the conference floor chaos.

Today’s sessions pivoted toward the human side of technology—how we might deploy AI in our workplaces, what our careers could look like in this rapidly evolving landscape, and how to build more humane social experiences online. Here’s what caught my attention:

Ian Beacraft: How Not to Screw Up an AI Transformation

Beacraft outlined five crucial areas for successful AI transformation:

1. Adaptability: With skills having a shelf life of less than two years, companies need “surge skilling” – rapid skill acquisition and transfer – rather than frontloaded education. Learning budgets must scale with technology.
2. Multiple Intelligences: Leaders must become followers and followers must become leaders. Somatic intelligence (listening to your body) becomes crucial as we navigate change and uncertainty in the AI era.
3. Small Data is the New Big: High-signal, proprietary data is where competitive advantage lies. “LLMs will make you average” because they’ve commodified 99% of human knowledge, making the remaining 1% of proprietary knowledge invaluable. You need to train your AI on the specifics of your work, sitting that layer on top of the LLM
4. New Metrics: The AI era demands measurements beyond efficiency, like breakthrough rates, new systems developed, and new tools deployed. “What gets measured gets made.”
5. Creative Generalist 2.0: Teams need to be restructured around AI augmentation by defining core roles, identifying AI enhancement opportunities, consolidating similar responsibilities, and evaluating team structures holistically.

We’re trying to navigate new terrain with outdated maps.

Bluesky Rising: Putting Users Back in Control

Jay Graber, Bluesky CEO, presented an alternative to traditional social platforms that aligns with the day’s human-centered theme:

• Protocols vs. Platforms: Building on open protocols rather than closed platforms has allowed Bluesky to scale to 32 million users with just 21 team members. As Graber explained, “We want to make social media more like the Web”—open, interconnected, and user-driven.
• “Billionaire-proof”: As a public benefit corporation focused on open protocols, Bluesky’s structure means users can take their identity and data elsewhere if the company makes poor decisions. “If we start doing a shitty thing with our app, people can build a new app or connection tool and your identity and information travels with you,” Graber noted.
• User-controlled moderation: Bluesky allows users to select or create their own content moderation services, defining their experience rather than having it dictated from above. For example, users can choose services that label AI-generated images or filter political content.
• AI content consent: Recognizing that user data sovereignty extends to AI training, Bluesky is developing frameworks for user consent at both account and post levels. This approach contrasts sharply with the typical corporate practice of scraping user content without permission.

After watching both keynotes, I couldn’t help but notice the complementary visions: Beacraft advocating for human adaptability and unique knowledge in the workplace, while Graber champions user agency and choice in our social spaces. Both speakers seem to understand that technology should empower rather than replace human judgment.

Tomorrow is my last day at the show. I’ll be back to Space and Robots with a touch of AI.

Featured Photo by Ben Sweet on Unsplash