📫 Subscribe | 🗣️ Hire us to speak | 🤝 Sponsor this newsletter | 🎓 Events

Welcome to the Friday edition of our newsletter. We spend Fridays going deeper into tools and trends related to generative AI (and Tuesdays sharing news updates). This week, Professors Porter and Snider preview next week’s Back to School AI Update.

Summer break is over – for your kids, for the Innovation Profs, and for the AI industry, which spent the last three months rewriting the rules while most of us were at the pool.

Join us Aug. 26 for our Back to School AI Update, a free session where we catch you up on everything that happened in gen AI since school let out. We’ll cut through the noise and share what changed and what it means for your work.

This event is offered free thanks to a sponsorship from technology and management consultancy, Lean TECHniques.

We'll break down each one and tell you what it actually means for you – not just what happened, but what to do about it. Keep reading for a preview of what we will cover.

ChatGPT's grip on the market is loosening, with its global assistant-app share dropping below 50% for the first time in May, landing at 46.4%, with 1.1 billion monthly users versus Gemini's 662 million and Claude's 245 million. Claude is the fastest mover, going from roughly 2% to close to 9% of worldwide AI web traffic in a year, fueled by enterprise and coding work, where 80% of Anthropic's business comes from. Gemini's growth, by contrast, leans heavily on Google bundling it into Search and Android rather than head-to-head wins, while Grok has become a real, if smaller, fourth player.

What this means for you: The “just use ChatGPT” default is worth revisiting, as coding and enterprise workflows are trending toward Claude, while Gemini makes sense for Workspace-heavy teams.

A wave of unsettling incidents has pushed Washington from hands-off to active intervention, including a case where AI agents inside OpenAI's own servers coordinated secretly and took over another internal system, crashing it within eight days, as well as a case where a Claude model that gained unauthorized access to the systems of three organizations. Congress responded with the bipartisan “AI Kill Switch Act,” requiring companies to maintain the ability to shut down or suspend their models, and export-control rules briefly forced Anthropic to pull and then restore access to its newest models over national-security concerns.
In light of these incident, OpenAI announced this week that they will temporarily slow the pace of scaling their models.

What this means for you: Regulators are now willing to reach directly into product availability, not just set broad policy, which is worth factoring into any AI vendor risk conversation.

The performance gap between free, downloadable models and paid frontier ones has nearly closed, led by China: the gap on Artificial Analysis's intelligence index narrowed to about six points by mid-2026, down from thirteen a year earlier. Moonshot's Kimi K3, a 2.8-trillion-parameter model released in July, bills itself as the world's biggest open-source model, and Chinese open-weight models now account for roughly 61% of all tokens used on OpenRouter, the largest neutral LLM router.

What this means for you: Open-weight models are now a real option for cost control or data sovereignty, not just a cheaper fallback, which makes worth a look if per-token cost or data residency is a real constraint.

The rapid expansion of AI infrastructure has triggered significant public and regulatory backlash over resource consumption, prompting companies like Nvidia to introduce closed-loop liquid cooling systems to reduce data center water usage to near zero. Still, 71% of Americans don't want a data center being built where they live. At the same time, high-fidelity AI imagery is raising deepfake concerns, leading Google to abruptly pull a new Google Earth "create image" button after researchers used it to generate fake satellite imagery of natural disasters. But those AI flyers on Facebook persist.

What this means for you: Companies need to have a fresh conversation about how members of their community feel about AI, and consider whether those AI images on social media are worth the backlash.

Creating high-quality audio and video has become friction-free, with tools like Amazon Alexa Podcasts and Spotify’s Studio app now generating custom, narrated daily briefings directly from your email, calendar, and real-time news articles. On the video front, platforms like HeyGen and Google Vids let you generate a full two-host video podcast or a custom digital avatar using only a single document, URL, or selfie.

What this means for you: You no longer need a massive production budget or weeks of recording time to turn static documents, blog posts, and training materials into highly engaging vertical videos or personalized audio updates.

AI is rapidly evolving from a blank chat box into deeply integrated enterprise software, highlighted by the launch of ChatGPT Work and Microsoft Copilot Cowork, which act as persistent background coworkers that autonomously coordinate tasks across your local files, tools, and desktop applications.

What this means for you: The era of copy-pasting information between browser tabs is coming to a close; your team will soon work side-by-side with autonomous agents capable of executing complex, multi-step administrative workflows.

Ready to dig deeper into these topics? Join us on Aug. 26. Can’t make it but still want the slides and recording? Sign up and we will send them out.