Platforms like Google AI Ultra and Anthropic Claude are exciting to me because they offer powerful, intuitive, and extensible tools that help people at any level visualize and achieve their goals. And these platforms are designed by hardcore researchers.

Case in Point: I have had the pleasure of working in healthcare for 25 years as a technologist, entrepreneur, and researcher. While serving as a research fellow in psychosocial oncology at Sloan-Kettering Cancer Center in the mid-1990s, the World Wide Web was still in its infancy and already very overhyped. We began witnessing people affected by cancer creating binders of printouts from various sources on the web: some good, some poor, and some irrelevant or just plain “snake oil.” Our goal was and still is to better understand where our patients’ information-seeking motivations originate, how they use the web, and to never underestimate the power of hope.

I quickly became fascinated with how the Web spreads information and how misinformation can be effectively countered through evidence-based education and support. I started creating UX designs and flowcharts for potential websites whenever I had a free moment. Unfortunately, from 1996 to the mid-2000s, our ideas were much more ambitious than the technology could support.

Ultimately, I shifted away from coding to focus on strategic planning, client engagement, and business growth for my previous company, EthosCE LMS. Today, I have access to powerful Google AI tools, a vivid imagination for new models and use cases to come, and the ability to build them.

A Simple Example

• “Wouldn’t it be great if I could create a truly real-time, interactive patient case study that uses a first-person ‘Choose Your Own Adventure’ format, where the case is dynamically generated from multiple data sources, including peers, EMR, LMS, and external apps?”

• The virtual patient case app would follow semi-autonomous and autonomous workflows through the internal and external accreditation compliance process, manage learner transcripts, analytics, communication with 3rd parties via APIs, etc.

• “Rinse, Wash, and Repeat.” Start the semi-autonomous process over again, creating new, smarter adventures derived from existing learner group data analytics, new treatment guidelines, benchmarking, etc.

In conclusion, if the Gemini LLM framework with agentic AI had existed in 2000, I likely would have saved years on inputting, testing, debugging, and dealing with insomnia, instead of building rich applied AI applications in professional and consumer health. Still, I wouldn’t trade it in.

Definitely check out the post below, “Every Google AI Tool in 2026.” Thank you, Daria, for this comprehensive overview. Great stuff!

Shocker: “AI Brain Fry” results from mental fatigue caused by overusing or mismanaging AI tools beyond one’s cognitive limits. I usually dislike Web 2.0-style buzzwords (e.g., EpicDigital.ai :), but this term and its early definition seem to fit. Workers are now being pushed to operate at an unsustainable level of cognitive effort, while AI can handle the same tasks without needing breaks, sleep, or social interaction. Recent studies by researchers at MIT, Harvard Business Review, and the Boston Consulting Group have identified common causes, signs, and consequences of this phenomenon.

My favorite quote: “It was like I had a dozen browser tabs open in my head, all fighting for attention.” hashtag#AI hashtag#WorkBurnout hashtag#AIBrainFry hashtag#CognitiveStress hashtag#Productivity

Check out the complete article: https://lnkd.in/eAaB79XE

When I first saw the headline above, my immediate reaction was “Wow. Now that is a rather misleading Web 2.0-ish headline. Either we have fallen for the most well-orchestrated trillion-dollar scam in human history, and AI is really vaporware. Or, there are widespread shortcomings in the strategic leadership and knowledge needed to effectively leverage AI and its various frameworks.” It would take a lot of effort to produce a near-100% failure rate on return on investment (ROI). Let’s go with the latter and add some context.

Common traits contributing to unsuccessful AI initiatives include:

1. Lack of AI strategic leadership and planning across the enterprise. For example, “AI First” doesn’t happen by simply announcing it at a meeting. It’s a mindset and approach that requires ongoing support and commitment from key stakeholders.

2. Avoid conducting deep research on AI frameworks, resource allocations, KPIs, and case studies to meet defined goals.

3. Having the wrong stakeholders at the table, working with unclear goals and objectives that are not aligned with the overall business.

4. Unwillingness to pause an initiative (in progress) to recalibrate the project plan, risk, and trajectory as “stress fractures” begin to appear.

5. Ignoring Albert Einstein’s words: “We cannot solve our problems with the same level of thinking that created them.”

Yes. Artificial Intelligence (AI) frameworks are rapidly evolving, and some tools are unreliable, ineffective, or simply benign. However, many AI breakthroughs are truly transformative and astonishing, influencing every stratum of our society. Given this new reality, we cannot blame technology alone for failures without also recognizing the strengths and limitations of humans in generating immediate, widespread ROI success.

A significant lack of essential AI knowledge and leadership skills is a big challenge. Two widely publicized reports, The GenAI Divide: State of AI in Business 2026 (MIT) and PwC 2026 Global CEO Survey, show that up to 96% of companies have not yet achieved significant ROI. However, 94% of CEOs plan to continue or increase funding, viewing AI as a long-term strategic priority rather than a quick profit.

Primary Causes of Negative ROI

The lack of ROI in AI is rarely due to the models themselves. Instead, it is often the result of systemic structural, and organizational challenges:

1. The Data Foundation Crisis: 73% of data leaders identify data quality and fragmentation as primary barriers.

2. “The Perpetual Pilot: So, Close. Yet So Far.” Approximately 80-85% of AI initiatives fail to reach deployment. Many start as “standalone” prototypes that cannot scale or lack a clear way to fit into existing enterprise workflows. Hmmmm... one might say that “Building prototypes should be formal parts of application design and development.” Stakeholders can then “kick the tires” on key assumptions, timelines, and functionality in a safe, stable environment. The problem with prototyping is:

   1. Stakeholders may not understand the “cause and cascade effect” of an issue and move on to something easier to fix;

   2. Continuous changes to the project scope are not reflected in the final prototype design, leading to conflicts where they shouldn’t be.

   3. Who likes delivering bad news? Nobody. But failing to inform your team of an issue and to seek help understanding its cause or limitations before a prototype goes live is, at the very least, your professional responsibility. At worst, it could lead to a pink slip when things go “boom.”

3. The Productivity Paradox: While 96% of execs expect productivity gains, 77% of employees report that AI has actually increased their workload due to the time required for reviewing AI output and learning new tools.

4. Workflow Misalignment: Only 21% of organizations have redesigned their business processes to align with AI capabilities; most simply “overlay” AI on top of inefficient, legacy processes.

What the “Successful 5%” Are Doing Right

The small number of companies seeing measurable ROI shares specific traits:

• Hyper-Focus on Pain Points: They steer clear of broad “AI for Everything” strategies and instead focus on one high-value, specific business bottleneck (e.g., reducing vendor disputes or automating invoice processing).

• Customization Over Commoditization: They move beyond generic “out-of-the-box” bots, building customized systems that learn from company-specific data and adapt to internal workflows.

• Strategic Partnerships: Successful firms are twice as likely (67% vs 33%) to use external partnerships for developing solutions instead of relying only on in-house development.

• Back-Office Prioritization: Prioritize AI efforts on unglamorous back-office automation where ROI is easiest to measure, instead of focusing on high-visibility marketing or front-end tools.

In closing, I would love to hear your thoughts and feedback on this topic. I am personally a big fan of prototyping, and I believe it is one of the most essential (often overlooked) phases of successful AI initiatives.

The Thinking Game is one of the top documentaries detailing the evolutionary development path towards achieving Artificial General Intelligence (AGI). The film spanning five years follows prodigy Sir Demis Hassabis, CEO and co-founder of Google DeepMind as he and his research team tackle extraordinary complex challenges in AI, physics, and chemistry. At age 47, his work on protein folding earned him the 2024 Nobel Prize in Chemistry (visit AlphaFold website) having catalogued most protein structures that are the building blocks of life’s molecules. Within healthcare, AlphaFold has exponentially reduced the cost and time associated with new drug research. AlphaFold is currently used by over 2 million researchers worldwide.

MIT Program Summary: “The rules of leadership are changing. Today's middle and upper managers face a challenge: fast-moving technologies like generative AI, smart automation, and digital twins offer unprecedented opportunities to transform business, but simply investing in high-value technology projects is no longer enough.

Success in this new era requires a fundamental shift in mindset and capability. You need to understand not just what these technologies can do, but how to lead differently in a world where transformation is emergent, culture matters more than projects, and engaging your people is what makes it all work.

Watch the Webinar discussion and Q&A on leading AI-powered transformation with George Westerman, Principal Research Scientist and Senior Lecturer in Information Technology. George and his colleagues pioneered the study of digital transformation. Now they can show you how to transform in today’s fast-moving world of intelligent machines continuously.”

In this 60-minute virtual event, you will:

• Learn about 7 driving principles that define AI-era leadership

• Gain insight into key emerging technologies that can create new transformation opportunities for your organization

• Build greater confidence to identify and create value, rather than just reacting to digital change

In this podcast, Conor Grennan, Chief Architect of AI at NYU, shares his non-technical approach to successful AI adoption across market sectors. He explains the difference between digital transformation and actual behavior change and discusses how to prepare organizational leadership for long-term success and reduce the risk of failure.

Watch Video Discussion

Time stands still. I remember having a visceral reaction to the interview, thinking, 'Wow. The Administration truly doesn't understand enough to care." Scientific breakthroughs are more than a "flash in the pan" moment. When research answers one question, it often raises many new questions that need answers in different populations and directions. Shutting everything down for a month likely means losing years of progress toward saving lives.