In traditional software, correctness is binary—code either works or it doesn't. But in the age of AI, we've entered a new paradigm where correctness exists on a spectrum. Discover how modern developers measure and validate AI system outputs, from token probabilities to semantic similarity scores, and learn why narrow, focused AI agents often outperform their general-purpose counterparts.
The infusion of AI and LLM capabilities into software development is driving a paradigm shift on multiple levels. Design patterns are evolving – from harnessing in-context learning and retrieval strategies to new agent-based models – making software more adaptive and intelligent by design.
Joining a letter exchange group leads to the broader realization that sometimes, the path to better creative output leads through deliberate inefficiency.
Effective prompt engineering is an art as much as it is a science. Programmers can ensure quality LLM output in their apps by following established prompting frameworks.
Apple is failing to implement artificial intelligence in a way that plays to their greatest strengths.
The desert, rolling hills of sand, and the subtle movement of time.
On the interplay of melodies and inner voices, the profound complexity beneath the surface of our lives, and slowing down to listen.
The intersection of tides, the rhythm of life, Bach, and personal exploration.
Breathtaking landscapes, sanctity in scarcity, art, and humanity’s impact on nature. The transformative power of place on an artist’s work and spirit, a reminder of how the land and its beauty can become a lifelong passion.
The unlikely convergence of Ansel Adams, rubato, and a coffee advertisement.
A bad bottle of Bordeaux teaches a profound lesson.
The next trillion-dollar tech company won't be built by creating the best AI model - it will be built by controlling how, where, and why people use AI.