Scaling AI, Editing Copy, and Pouring Cocktails in Vegas
No Time for a Thinkpiece—But I Still Want to See You at AI4
I don’t have time for a full essay this week. I’m waist-deep in editing copy for the next issue of Worth, which means fighting the good fight against jargon, chasing down missing data points, and rewriting ledes that start with “In today’s rapidly evolving landscape.” You get the idea.
But if you’re heading to Las Vegas for AI4, I hope you’ll make time for two things:
First, I will be moderating a panel on Scaling AI in the Enterprise on Wednesday, August 13. This isn’t a buzzword bingo session. We’re tackling the real-world challenges of transitioning AI from pilot to production—dealing with messy data, legacy systems, organizational charts, and the cultural inertia that stalls even the best ideas. If you’ve ever tried to push an AI product past a risk committee, this conversation is for you. If you will be there, here’s the link to add it to your schedule.
Second, on—Monday, August 11—Worth is teaming up with IBM for a private reception at the MGM Grand. It’s called Sip, Savor & Discover the Future of AI. The format is simple: smart people, strong opinions, good food, and better cocktails. We’re keeping this event off the record and invite-only, but if you’re receiving this newsletter, you’re invited. Space is limited, so RSVP here.
That’s it for now. I’ll be back next week with something more polished—assuming I survive this editorial deadline. In the meantime, let’s connect in Vegas.
Machined Readings
OpenAI preps GPT‑5 for August launch
OpenAI’s next‑gen model, GPT‑5, is slated for an early‑August debut with “mini” and “nano” variants. Sam Altman told a podcast host that GPT‑5 answered questions that made him feel “useless,” and the model unifies OpenAI’s o3 reasoning into a single system. If GPT‑5 arrives on schedule, expect a fresh wave of AI one‑upmanship from rivals.
Google’s AI Overviews slash news traffic by 79 percent
A study by Authoritas found that when Google’s AI-generated summaries appear, websites previously ranked first can lose 79 percent of their clicks. A Pew survey showed users click the link under an AI summary only once per 100 searches. News executives complain that AI snippets trap users in Google’s walled garden and threaten publishers’ traffic. As a publisher, let me tell you, this story checks out.
Trump orders “neutral” AI models and deregulates datacenter construction
Donald Trump’s July 23 orders direct federal agencies to ensure AI models are politically neutral and free from “woke” ideology. The orders expedite data‑center construction and call for deregulation, with Trump jokingly suggesting renaming “Artificial Intelligence” as “Genius”. Critics worry the neutrality mandate could become a pretext for policing content.
AI firms score low on existential safety
A Future of Life Institute report graded major AI labs on their readiness to avoid catastrophes; none scored above a D. Anthropic got a C+, OpenAI a C, and DeepMind a C‑. The institute stated that companies lack coherent safety plans, comparing the situation to building a nuclear plant without a safety protocol. And not for nothing, but some of these companies are really looking at building nuclear plants…
The high‑stakes talent war for AI researchers
The Guardian likens the scramble for AI talent to sports trades, noting that a high‑profile OpenAI researcher joined Meta’s Superintelligence lab. Mark Zuckerberg has pledged to spend “hundreds of billions” on AI, including a data center nicknamed Hyperion. Nvidia CEO Jensen Huang told The New York Times that half of the world’s top AI researchers are in China, underscoring the global stakes.
Hallucinatory bug reports swamp security bounty programs
TechCrunch reports that large language models are generating fake vulnerability reports that flood bug‑bounty programs. Projects like Curl and CycloneDX have shut down their bounties, while companies like HackerOne have built AI triage tools to filter bogus submissions. It’s a reminder that AI can be both a valuable cybersecurity tool and the most prolific source of spam the world has ever known.
Trump’s AI Action Plan favors growth over guardrails
TechCrunch notes that the Trump administration’s AI Action Plan emphasizes building data centers, scrapping environmental reviews, and preempting state regulations. The plan aims to upskill workers and partner with states, but treats regulation as a barrier. It mandates the use of neutral AI for federal procurement and calls for the FCC to evaluate state AI regulations. We are nailing the “move fast” part, what comes next again?
Samsung backs Memories.ai to tackle long‑form video analysis
The startup Memories.ai raised $8 million from investors, including Samsung Next, to analyse up to ten million hours of video. Co-founder Shawn Shen says existing models struggle with understanding long-term context. By processing videos on the device, the company claims to preserve privacy. Added benefit to Samsung: more phone sales!
AI to run on gas: Trump and energy industry champion fossil‑fuel‑powered datacenters
At a Pittsburgh summit, Trump, energy executives, and tech leaders promoted AI as a boon for fossil fuels. Trump said data centers will need “double the electric” and that AI demand justifies more gas generation. The summit announced $92 billion in energy and AI investments, with Meta building a gas‑powered data center called Prometheus in Ohio.
Just a reminder for those of you who slept through that mythology class, Prometheus stole fire from the gods and gave it to mankind. A solid, by any measure. But in return, Zeus chained him to a rock, had an eagle eat his liver every day, and then healed him overnight so he could endure a life of endless torment.
Happy Monday!
Vladimir Lukic: Seamless Tech, Smart Autonomy & Reframing the Future | Milken 2025
At the Milken Institute Global Conference, Worth Media’s Dan Costa sits down with Vladimir Lukic: Managing Director & Senior Partner at BCG - to explore how the most powerful uses of AI are often the ones we don’t even see. From consumer experiences to enterprise tools, AI is quietly transforming the way we work, decide, and interact—with just the right balance of autonomy and control. In this practical and insightful conversation, you’ll learn:
🤖 Why the best AI experiences are invisible but impactful
🧩 How companies are giving AI just enough autonomy to handle low-risk, high-speed tasks
🛍️ What AI-powered review summarization, video buffering, and chatbots reveal about seamless user experience
🛠️ How the real value of AI is found in pattern recognition—spotting the biggest questions and reframing them
🔁 Why reframing problem statements is just as important as solving them
Chapters:
0:00 – Giving AI the Right Level of Autonomy
2:00 – Best-in-Class AI Applications You Don’t Even Notice
3:40 – Video Buffering, Sentiment Summarization & Seamless Design
5:30 – What Makes AI Valuable for Pattern Recognition
6:30 – The Power of Reframing Problems in the AI Space
⚙️ “The best AI isn’t flashy—it’s just quietly making everything work better.” Subscribe to Worth Media for more conversations from the forefront of innovation, strategy, and real-world AI deployment. #Milken2025 #ArtificialIntelligence #InvisibleAI #TechStrategy #UserExperience #Automation #WorthBeyondWealth #BCG