Welcome back to Fresh From the Labs! Shilpa, Jared, and Kevin dive into another week of AI exploration at Pioneer Square Labs.

The conversation kicks off with a deep dive into OpenAI's newly released o3 model. Kevin shares his initial impressions after putting it through its paces, highlighting its impressive intelligence, speed, and surprisingly adept web search capabilities. Jared discusses using o3 for complex planning and research, noting its power but also the challenge of its high-level output potentially requiring a "dumb it down" prompt and raising new questions about high-fidelity hallucination. We explore the nuances of interacting with these increasingly sophisticated models, including the pros and cons of using AI memory and chat history. Do you keep it on for context, or turn it off for truly fresh brainstorming?

We then zoom out to discuss how AI use cases are evolving. Referencing recent studies and our own experiences, we explore the shift from brainstorming towards deep research, analysis, and even companionship/therapeutic uses (though we debate our own usage patterns!). A key theme emerges: using AI to rapidly expand professional skill sets, tackling tasks outside our core expertise – from generating marketing campaigns to coding proofs-of-concept. Is this genuine skill-building, or are we leaning on "blind trust"?

This leads to a bigger discussion about AI as an "unblocker" – a tool that helps us overcome hurdles and incrementally build expertise, potentially changing how we approach learning and complex projects. Finally, inspired by recent optimistic takes, we put on our future-gazing hats: what could the world look like in 10 years if AI continues its trajectory? We speculate on everything from the end of disease and automated chores freeing up creative time, to a more efficient distribution of information impacting everything from markets to civic engagement, while acknowledging the necessary conversations around job displacement and societal adaptation.

Join us for insights, experiments, personal anecdotes (including why Jared turns off his chat history!), and a look at both the practical present and the potential future of AI.

Links:

HBR Article: How People Are Really Using Gen AI in 2025

60 Minutes: Demis Hassabis Interview

Dive deep into the latest advancements shaking up the AI landscape! In this episode, we unpack the significant release of Meta's Llama 4, exploring its capabilities and the internal dynamics at Meta that drive such innovation.

Join the conversation as we tackle the persistent challenges of evaluating AI model performance and discuss industry trends. We then shift focus to the exciting potential of AI agents integrated directly into operating systems – could this revolutionize how we interact with our devices?

The discussion also covers the evolution of voice assistants towards more natural interaction, the crucial role AI is playing in business transformation and product development, and the necessary balance between automation and human oversight. Get insights into prototyping with AI, navigating its impact on the future of work, and understanding the practicalities of implementing these powerful tools. Tune in for a comprehensive look at where AI models are heading and how they're reshaping our digital world.

Chapters

Introduction and Excitement for Llama 4

02:18 Insights from Adam on Llama 4 Launch

05:09 Customer Feedback and Model Evaluation

09:14 Challenges in Model Evaluation and Industry Trends

11:35 Upcoming Events and Opportunities

12:14 Exploring Agents in Operating Systems

21:13 Exploring Content-Based Search and Portability

24:41 The Future of Local Models and Integration

27:53 Advancements in Voice Assistants

32:18 The Role of AI in Business Transformation

36:18 Navigating AI's Impact on Employment

42:52 The Future of Work with AI

Welcome back to "Fresh from the Labs," where the Pioneer Square Labs team (Shilpa, Jared, Kevin) is joined by special guest Brian Fioca (formerly PSL, now heading to OpenAI) to discuss their latest AI experiments and observations.

Following up on their previous discussion about "vibe coding," the team shares insights from a recent internal hackathon where non-engineers used tools like Lovable to build applications. They cover the excitement and empowerment users felt, the ability to leverage specific skills like design, but also the real-world limitations encountered – debugging difficulties without code visibility, the need for structured prompts to avoid feature bloat, and hitting technical walls requiring engineering expertise (like Superbase RLS setup).

The conversation shifts to the recent advancements in AI image generation, spurred by OpenAI's updates, discussing the improved usability through iterative refinement and its potential as a tool for creatives rather than a replacement. They also touch on the rapidly increasing user expectations for AI capabilities.

Diving into the business side, the group explores the high valuations and fundraising rounds of AI startups like Cursor, questioning the economics behind subsidized API calls, the true cost of compute at scale, the potential for RLHF data to build moats, and whether open-source AI models and tools (like Aider, Augment, Stagehand, Open Manus) can effectively compete with vertically integrated systems. They also discuss emerging AI-driven business models in B2B, services, and custom software development.

In this first episode of Fresh From the Labs, we (Shilpa, Jared, and Kevin from the product team at Pioneer Square Labs) kick things off by sharing why we’re starting this podcast: to give you a window into the AI experiments we're running day to day, and to think out loud with fellow builders.

We dive into what’s impressed us lately — especially the leap in image generation from ChatGPT-4o. Jared shares how a friend used it to create tabletop miniatures, and we talk about how tools like DALL·E have gone from novelty to genuinely useful. That “toy to tool” moment becomes a theme.

We also unpack our recent move to bring in Mac Studios with massive memory to experiment with running local models. We talk through the implications of local-first AI: privacy, cost, and control — and what that means for startups and enterprise.

The conversation shifts to “vibe coding”: how tools like Cursor, Claude, and Replit are changing how we prototype and build software, what works well, and where it still falls apart — especially when it comes to debugging and keeping things clean.

There’s some speculation, a few tactical tips, and a lot of curiosity. Plus, a quick preview of what we’re playing with next.