Fixing AI Blindspots: Hamilton Mann on Digital Transformation

Communication is more than text. In this clip, Germán León, AI Maven and Founder at Helvetica Digital, reflects on the connection between communication and AI. If 80 percent of human communication is nonverbal, reducing AI to typed prompts misses the bigger opportunity. The real shift lies in designing systems that understand context, behavior, and multiple signals, not just words. As AI evolves, the focus should not be on writing better prompts, but on designing interactions that augment human capability. This is part of a longer conversation in our Forwardism in Focus series. Watch the full conversation and read the article: https://lnkd.in/drPKmYMF #Framna #Forwardism #AI #DigitalProducts #ProductDesign

"People have integrated AI to the point where it becomes nearly impossible for them to imagine writing an email themselves." In a compelling discussion with Dr. Cornelia Walther, we explore how unchecked AI use is driving cognitive decline, weakening critical thinking, decision-making, and even meaning at work, as more people default to machine judgment. Listen to the full discussion: https://gag.gl/xLIw2U Thomson Reuters Institute #AI #ClarityPodcast #GenAI #TRInstitute

"People have integrated AI to the point where it becomes nearly impossible for them to imagine writing an email themselves." In a compelling discussion with Dr. Cornelia Walther, we explore how unchecked AI use is driving cognitive decline, weakening critical thinking, decision-making, and even meaning at work, as more people default to machine judgment. Listen to the full discussion: https://gag.gl/u3mn5H Thomson Reuters Institute #AI #ClarityPodcast #GenAI #TRInstitute

"People have integrated AI to the point where it becomes nearly impossible for them to imagine writing an email themselves." In a compelling discussion with Dr. Cornelia Walther, we explore how unchecked AI use is driving cognitive decline, weakening critical thinking, decision-making, and even meaning at work, as more people default to machine judgment. Listen to the full discussion: https://gag.gl/u3mn5H Thomson Reuters Institute #AI #ClarityPodcast #GenAI #TRInstitute

"People have integrated AI to the point where it becomes nearly impossible for them to imagine writing an email themselves." In a compelling discussion with Dr. Cornelia Walther, we explore how unchecked AI use is driving cognitive decline, weakening critical thinking, decision-making, and even meaning at work, as more people default to machine judgment. Listen to the full discussion: https://gag.gl/u3mn5H Thomson Reuters Institute #AI #ClarityPodcast #GenAI #TRInstitute

"People have integrated AI to the point where it becomes nearly impossible for them to imagine writing an email themselves." In a compelling discussion with Dr. Cornelia Walther, we explore how unchecked AI use is driving cognitive decline, weakening critical thinking, decision-making, and even meaning at work, as more people default to machine judgment. Listen to the full discussion: https://gag.gl/DqxJtc Thomson Reuters Institute #AI #ClarityPodcast #GenAI #TRInstitute

"People have integrated AI to the point where it becomes nearly impossible for them to imagine writing an email themselves." In a compelling discussion with Dr. Cornelia Walther, we explore how unchecked AI use is driving cognitive decline, weakening critical thinking, decision-making, and even meaning at work, as more people default to machine judgment. Listen to the full discussion: https://gag.gl/DqxJtc Thomson Reuters Institute #AI #ClarityPodcast #GenAI #TRInstitute

Excited to share our latest work on LLM fine-tuning. As models scale, it’s often assumed they become harder to optimize. Our research challenges that. We show that fine-tuning is governed not by the full parameter space, but by a small set of high-impact directions. This structural view enables simpler, more efficient optimization—and helps explain common dynamics like rise–then–decay. Why this matters: • Performance may depend more on optimization structure than model size • Simpler methods (e.g. perturbation or population-based) can remain effective at scale • Rise–then–decay may be a predictable transition—not instability—informing better early stopping and training strategies • Points toward a shift from thinking about scale → structure in AI systems 📄 Read the blog: https://cgnz.at/6008QtdAn 📄 Read the paper: https://cgnz.at/6001Qtdfl #AI #MachineLearning #LLM #Quantization #EvolutionStrategies #AIResearch

AI can accelerate research, surface patterns, and show what’s possible. But it can’t deliver results on its own. At Powerhouse, proven outcomes come from customized solutions, not templates. Our team brings industry experience, execution, and real-world insight to translate strategy into work that actually performs. Because results aren’t one-size-fits-all. Read the full breakdown here: https://ow.ly/XW5J50YofNm #PowerhousePlanning #PowerhouseServices #PowerfulTeam #InnovationInAction #AI #PHvsAI

Throughout March, we’ll be sharing perspectives on innovation in action, including insights from our Powerhouse vs. AI resource and how thoughtful, human-led planning continues to make the difference in an increasingly digital world. Follow along as we explore what innovation looks like when expertise leads and technology supports. https://ow.ly/aIAE50Yoflt #PowerhousePlanning #PowerhouseServices #PowerfulTeam #InnovationInAction #AI #PHvsAI