In this HCI Webinar, I talk with Juan Betancourt about how companies can operationalize agility for all teams. Juan Betancourt, CEO of Humantelligence, is a visionary leader with a lifelong commitment to reshaping the business landscape for the better. Having observed the limitations of conventional human capital management systems during his time in the software industry, Juan recognized a need for innovation. It was this realization that led him to Humantelligence, where he saw the potential to revolutionize productivity, motivation, and employee retention while making it accessible to all. With a track record of revitalizing global brands like Puma and overseeing the US division of Décathlon, Juan's expertise is unmatched. A Harvard economics graduate with an MBA from The Wharton School, Juan is committed to making work better for all and actively engages in community leadership roles.

This research examines ARC-AGI-3, a 2026 benchmark designed to test an AI’s ability to solve novel problems without prior training or instructions. While current frontier models excel at specialized tasks within their training data, they struggle significantly with the "unknown unknowns" presented in this interactive test, whereas humans succeed easily. The research argues that true artificial general intelligence is defined by the efficiency of acquiring new skills rather than just performing learned tasks. Because of this intelligence gap, organizations are advised to automate only verifiable domains while relying on human judgment for strategic and creative roles. Ultimately, the research suggests that while AI is a powerful tool for structured work, it still lacks the flexible adaptability inherent to human cognition.

The ARC-AGI-3 benchmark, introduced in March 2026, reveals a striking and fundamental performance gap between human intelligence and the most advanced AI systems. While humans solve 100% of the novel, rule-free problem environments presented by ARC-AGI-3, leading AI models achieve less than 1% success, highlighting the vast divide in fluid intelligence. Unlike prior AI tests, ARC-AGI-3 drops agents into unfamiliar, interactive worlds with no instructions or explicit rewards, requiring them to independently explore, hypothesize, and adapt quickly to new rules and goals. Humans excel at this because of their innate ability to apply broad, abstract knowledge and reason fluidly in novel situations. In contrast, current AI systems rely heavily on pattern recognition from vast training data, excelling only when problems resemble prior examples. They struggle in truly novel contexts where no prior data or clear feedback is available. Highlights 🚨 ARC-AGI-3 reveals a massive 100-to-1 performance gap between humans and leading AI on novel problem-solving tasks. 🧠 The benchmark tests fluid intelligence: the ability to learn and adapt quickly in completely new environments without instructions. 👾 AI agents face “flying blind” environments, with no rules or goals, requiring exploration, hypothesis-testing, and inference from scratch. ⏱ Humans solve these tasks rapidly (median 7.4 minutes), while AI struggles due to reliance on pattern recognition from training data. ⚠️ The results challenge the belief that human-level general intelligence in AI is near; current AI lacks genuine adaptive problem-solving ability. 📊 The study proposes a three-tiered AI investment framework, matching AI deployment to task complexity and ambiguity. 🤝 The central lesson: combine AI’s strengths with human creativity and judgment, protecting uniquely human skills for the most complex tasks. Key Insights 🚀 The Fluid Intelligence Gap Defines AI’s Limits: ARC-AGI-3 isolates fluid intelligence—reasoning and adapting to new situations without prior knowledge—as the critical barrier for AI. While current AI systems master pattern matching and memorization from extensive data, they fail to independently discover rules or goals in entirely unfamiliar settings. This gap highlights a fundamental shortcoming, not a mere engineering challenge, but a core architectural and cognitive limitation of today’s AI. Fluid intelligence is the essence of human problem-solving, creativity, and learning, explaining why AI stumbles in groundbreaking, ambiguous, or unpredictable scenarios. 🧩 Novelty and Sparse Feedback Are AI’s Achilles’ Heel: Unlike many AI benchmarks providing explicit rewards or clear feedback, ARC-AGI-3 environments offer sparse, delayed, or ambiguous signals. This lack of immediate reinforcement prevents AI from using standard trial-and-error learning effectively. Humans, however, naturally infer causal relationships and long-term consequences, constructing mental models rapidly. This insight reveals why many AI systems excel in closed, well-defined domains but collapse when feedback is minimal or noisy, a common condition in real-world problems. 📚 Crystallized vs. Fluid Intelligence in AI: The distinction between crystallized intelligence (accumulated knowledge and experience) and fluid intelligence (innate reasoning and adaptability) is central. Current AI excels at crystallized intelligence through massive datasets—text, images, code—but lacks fluid intelligence. This explains their brilliance in known domains (e.g., language translation, code generation) but near-zero effectiveness in completely new, unstructured tasks. This insight implies that future AI breakthroughs require architectures capable of genuine reasoning beyond pattern recognition. 💡 Implications for AI Investment and Deployment: The three-tier model presented offers a practical framework for organizations: Tier 1: Fully automate routine, rule-based tasks with clear success metrics—for example, invoice processing, data categorization—where AI delivers high ROI and manageable risk. Tier 2: Use AI as an augmentation tool for ambiguous or semi-structured work requiring human judgment, such as drafting marketing emails or summarizing research, emphasizing human-AI complementarity. Tier 3: Preserve human leadership for complex, novel, and strategic tasks demanding creativity, adaptability, and deep interpersonal skills, which AI currently cannot replace. This framework helps balance enthusiasm with realism, optimizing AI’s impact while safeguarding human value.

Abstract: As artificial intelligence systems demonstrate increasing proficiency across specialized domains, the fundamental question persists: how close are we to genuine artificial general intelligence? This article examines the introduction of ARC-AGI-3, an interactive benchmark designed to measure agentic intelligence through exploration, goal inference, and adaptive planning in novel environments. Unlike predecessor benchmarks that focused on static pattern recognition, ARC-AGI-3 evaluates systems on their ability to autonomously navigate "unknown unknowns" without explicit instructions or prior exposure. With frontier AI systems scoring below 1% while humans achieve 100% success rates as of March 2026, this benchmark reveals a critical capability gap. Drawing on intelligence theory, organizational learning frameworks, and research on adaptive systems, this article explores what ARC-AGI-3 reveals about current AI limitations, the distinction between domain-specific automation and general intelligence, and the organizational implications of building truly adaptive intelligent systems. The analysis offers evidence-based insights for leaders navigating AI implementation while highlighting the distance remaining before artificial general intelligence becomes reality. See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

This research examines ARC-AGI-3, a 2026 benchmark designed to test an AI’s ability to solve novel problems without prior training or instructions. While current frontier models excel at specialized tasks within their training data, they struggle significantly with the "unknown unknowns" presented in this interactive test, whereas humans succeed easily. The research argues that true artificial general intelligence is defined by the efficiency of acquiring new skills rather than just performing learned tasks. Because of this intelligence gap, organizations are advised to automate only verifiable domains while relying on human judgment for strategic and creative roles. Ultimately, the research suggests that while AI is a powerful tool for structured work, it still lacks the flexible adaptability inherent to human cognition. See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

In this HCI Webinar, I talk with Christie Garcia about her book, YOUR EGO IS SHOWING. Christie Garcia, author of YOUR EGO IS SHOWING: How Ego Management Unlocks Authentic Confidence & Meaningful Success, is the founder and Ego Management expert at Mindful Choice. She is a seasoned leadership coach, speaker, skillful facilitator, and a distinguished contributor to Forbes Coaches Council. With a notable career spanning two decades, Garcia brings a wealth of expertise in the realms of sales, talent acquisition, leadership development, and Ego Management.

In this HCI Webinar, I talk with Dean E. Carter about employee experience design. Dean E. Carter, newly named CEO at Instill.ai and co-author of EMPLOYEE EXPERIENCE DESIGN, has over two decades of experience as an executive officer for renowned, and Fortune50 companies, such as Patagonia and Sears, as well as a board director for private and publicly traded companies. His views on employee experience and future of work are frequently featured in global publications, podcasts, corporate events, and mainstage keynotes.

This research explores how artificial intelligence competencies are fundamentally transforming the modern labor market by creating significant salary premiums and hiring advantages. Research indicates that workers possessing AI skills can earn up to 25% more than their peers and enjoy better access to non-monetary benefits like remote work and flexible leave. To remain competitive, organizations are shifting toward skills-based hiring and internal reskilling programs rather than relying solely on traditional university degrees. The research emphasizes that the economic success of AI depends less on the technology itself and more on an organization’s ability to build human capability and literacy. Ultimately, the research provides a strategic framework for businesses to manage talent scarcity and foster inclusive growth in an increasingly automated economy. OUTLINE: 00:00:00 - AI Skills and Your Paycheck 00:00:29 - From VIP Pass to Seat at the Table 00:01:32 - What Are AI Skills, Really? 00:02:40 - Why Companies Are Paying More 00:03:33 - Your Roadmap and Call to Action 00:04:55 - Build Together – Workers and Companies