Each week we’ll gather headlines and tips to keep you current with how generative AI affects PR and the world at large. If you have ideas on how to improve the newsletter, let us know!
What You Should Know
All Eyes on Nvidia at GTC 2025
This week, the tech world converges on San Jose, California, for Nvidia’s GPU Technology Conference (GTC) 2025, the chipmaker’s premier event for AI and accelerated computing. Tech pundits expect this event to highlight what’s next in AI, but it will be easy to get lost in the noise. Nvidia says the event will feature more than 1,000 sessions, 2,000 speakers, and nearly 400 exhibitors before it concludes on Friday.
NVIDIA Founder and CEO Jensen Huang will deliver the keynote today at 1 p.m. EDT. He is expected to unveil the next-generation Blackwell Ultra GPUs and the forthcoming Rubin chip, which will be the company’s flagship offering. Expect several hundred press releases from other tech companies at the conference, too, which can make it hard to determine what is real news and what is just trying to ride the event’s coattails.
For communicators, the real story is the broader AI trajectory all this news signals. Will the industry double down on high-cost, high-power AI models requiring expensive infrastructure, or will we see a shift toward more efficient, cost-effective alternatives like DeepSeek’s low-compute models? How will new technology support even more powerful reasoning models?
The answers will shape how AI tools evolve and what kind of infrastructure will be needed to support them. GTC’s announcements will help communicators gauge where AI is headed and what that means for the tools they use and the work they do.
Elsewhere …
- PODCAST: AI Is Reshaping Business Faster Than Leaders Expected
- Perplexity, Kalshi Prediction Market Partner on March Madness
- Google Plans to Release New ‘Open’ AI Models for Drug Discovery
- Local Newsrooms Are Using AI to Listen in on Public Meetings
- Over Half of American Adults Have Used an AI Chatbot, Survey Finds
- OpenAI and Google Ask for Government Exemption to Train AI Models on Copyrighted Material
Tips and Tricks
How to get a more compelling ending
What’s happening: One of the questions our team is asked the most is how to make AI writing more human (we built a list!). Recently, one particular AI tell keeps popping up: boring, repetitive, predictable conclusions. AI tools like to use the end of a piece of content to summarize key points and highlight the “first-mover advantage.”
If you’ve read something like, “By doing X, Y, and Z, companies will thrive in the future, while those that don’t will struggle to survive,” congratulations, you’ve definitely found some AI writing.
Why it matters: We want writing to stand out, not fade into the background with a generic, forgettable ending. A strong closing should leave the reader with something to think about, not regurgitate what they just read.
It’s not surprising that AI tools default to this structure considering how many blog posts suggest a proper conclusion includes some variation of “restating the thesis” and “summarizing key points.” But this is where human intuition comes in. You don’t want your conclusion to blend in with all the others that read the same way.
Try this: We’ve found success in including instructions like “bring the story full circle without making it repetitive,” “keep the conclusion forward-looking without urgency or high stakes,” “avoid framing success as dependent on immediate action,” and “use a sharp takeaway to keep the reader engaged until the end.”
If you have a fix you want to share, let us know and we’ll include it next time.
Quote of the Week
“They have a hammer, and they’re just making bigger hammers. They own the (training) market. And so every new chip they come out with has a lot of training baggage.”
— Bob Beachler, Vice President at Untether AI, to Reuters on Nvidia’s lead in inference computing, the process by which AI tools develop answers for user
How Was This Newsletter?
