In the Loop: Week Ending 4/26/25

Each week, I curate the most important developments across the AI landscape — from major product launches to policy shifts to real-world challenges around ethics and scale. This week’s stories include Microsoft’s bold move with AI agents, accelerating investment in AI infrastructure, and new tensions between innovation and regulation. Whether you're leading marketing, strategy, or product, these trends are reshaping the way we all work.

Sam Altman Says Politeness Is Costing Us Big in Compute Power & Environmental Impact

In one of my favorite articles from last week (h/t NickAbrahams), OpenAI CEO Sam Altman revealed that users saying “please” and “thank you” to ChatGPT — while nice — is burning tens of millions of dollars in unnecessary compute costs and driving a bigger impact on the environment. Altman believes being nice to your GPT has a positive impact on the results you get, but Ethan Mollick disagrees. How about you...are you nice to your GPT like I am? Read more (Futurism)

Anthropic Warns Virtual Employees Are Just a Year Away

At last week's Marketing AI Pulse event in Atlanta, reps from Salesforce spoke about how AI agents are already changing their workplace, and there will soon be AI agents communicating with each other to accomplish work tasks. This is further proof that the workplace of the near future will look very different. And now, Anthropic’s security team predicts fully autonomous “virtual employees” — AI agents with real job accounts and responsibilities — could arrive within a year. Corporate security teams aren’t ready, and the risks are real. Read more (Axios). (Image via Business Weekly)

Marc Benioff: AI Agents Will Supplement, Not Replace, Human Workforces

Meanwhile, Salesforce CEO Marc Benioff predicts that while AI agents will radically augment how companies operate, they will not replace the need for human employees. Speaking at the World Economic Forum, Benioff emphasized that successful companies will blend human workers with AI "co-pilots" — creating more productivity without eliminating the uniquely human aspects of work like creativity, trust, and collaboration. His comments reinforce a growing theme: the future isn't about humans vs. AI — it's about humans and AI working together.

Microsoft Launches AI Agents to Transform the Workday — and Challenge Google

Later this week and then again next week, I'm doing presentations and demos of AI agent technologies at the HMPS and Mirren conferences, including Microsoft Copilot. Microsoft continues to double down on agent technologies, last week introducing powerful new AI agents within its Office and 365 ecosystems, aimed at automating complex workplace tasks across multiple applications. Rather than simply responding to prompts, these agents can now proactively carry out multi-step tasks like coordinating meetings, sending follow-ups, and drafting reports. It’s a bold move to reframe Microsoft as the leader in workplace AI — and to challenge Google's growing push with Gemini AI for business. Read more (VentureBeat)

AI Could Be Coming to K-12 Classrooms — By Executive Order

In further evidence that the Trump administration is taking AI seriously, news broke last week that they are drafting an executive order aimed at boosting AI literacy in U.S. schools. The focus will be on early education and teacher training. Whether you see this as a good idea or a cause for caution, it signals how fast AI is moving into everyday life. Read more (Washington Post) (Image via MIT)

California’s AI Privacy Fight Is Heating Up

One of the challenges governments face is how much to regulate AI; the tension is around finding the balance between innovation and caution. With a recent statement, California Governor Gavin Newsom pushed back against new privacy agency rules that would curb AI use in hiring, healthcare, and beyond.

Newsom said,"...enacting these regulations could create significant unintended consequences and impose substantial costs that threaten California’s enduring dominance in technological innovation.” The battle between innovation and regulation is just getting started. Read more (Politico)

DeepMind: It's Time to Plan for Artificial General Intelligence (AGI)

Google DeepMind says AGI is no longer hypothetical — it’s coming. They’re urging the industry to think about safety frameworks now, not after the fact. Read more (Axios)

The Washington Post Partners with OpenAI to Bring News into ChatGPT

The Washington Post has struck a deal with OpenAI to integrate its journalism into ChatGPT, offering users short summaries, quotes, and direct links to full articles. Given the shifting sands of people's web search habits and the ongoing copyright lawsuits filed by The New York Times and other media outlets, where lawsuits over AI use are escalating, this kind of collaboration offers a glimpse at one possible path forward for media companies and the AI frontier companies. Read more (TechCrawlr)

Upcoming Speaking Engagements

I have three upcoming speaking gigs, two at healthcare's HMPS 2025 in Orlando, and one at Mirren Live in NYC. If you're attending either event, stop by and say hi! I'd love to connect and hear how things are going with AI implementation at your organization.