In partnership with
Trump spent Monday calling AI safety concerns a "SICK conspiracy." By Tuesday he was calling Nvidia's CEO live on stage to make sure nobody actually slowed anything down. Meanwhile, OpenAI is paying people to read your ChatGPT conversations, and some of AI's biggest customers don't trust its data policies either. Here are the 7 stories that actually matter for your business today.
100+ Claude Code hacks to ship code 10X faster
Top engineers at Anthropic and OpenAI say AI now writes 100% of their code.
If you're not using AI, you're spending 40 hours doing what they do in 4.
These 100+ Claude Code hacks fix that and help you ship 10x faster.
Sign up for The Code and get:
100+ Claude Code hacks used by top engineers — free
The Code newsletter — learn the latest AI tools, tips, and skills to code faster with AI in 5 minutes a day
Trump Called AI Safety a Hoax. Live, On Stage.
What happened: At the All-In Summit on Monday, President Trump had Nvidia CEO Jensen Huang put on a phone call in front of the live audience and told him not to let anyone slow AI down, calling safety concerns and data center backlash a "hoax." Huang answered on speakerphone that Nvidia was "not going to let that happen," according to TechCrunch. The call came a day after Anthropic CEO Dario Amodei publicly asked the industry to slow down, and after China's Foreign Ministry separately dismissed the slowdown push as fear-mongering. Chip stocks stayed down through the news while Microsoft and Alphabet shares climbed.
Why it matters to you: Neither Washington nor Beijing is going to hit the brakes on AI for you. Whatever guardrails your business needs around how you use these tools, that responsibility sits with you, not with a regulator who's going to show up and slow things down first.
OpenAI Pays People to Read Your ChatGPT Conversations
What happened: 404 Media reported that OpenAI pays hundreds of contractors, under an internal program called Project Lily, to read real ChatGPT conversations and rate the AI's answers, an effort aimed at reducing the chatbot's habit of just telling people what they want to hear. Contractors don't see your name, but they can see a summary of your saved "memories," which can include roughly where you live, and OpenAI's own filter for stripping out personal details admits it sometimes misses things. This review is on by default for Free, Plus, and Pro accounts, and turning it off only protects future chats, not the ones you already sent.
Why it matters to you: If you or your staff use a personal ChatGPT account for client details, financial numbers, or anything you wouldn't want a stranger reading, assume a human contractor could see it.
What to do about it: Turn off "Improve the model for everyone" in ChatGPT's data controls, or move business use to a Business or Enterprise plan, which is excluded from this human review.
Even AI Companies Don't Trust Each Other's AI With Their Data
What happened: Anthropic rolled out a 30-day data retention policy in June tied to its newest Claude model, and three major customers pushed back hard, according to Yahoo Finance. Palantir is now requiring a guarantee of zero data retention before it will integrate Claude into its software. Nvidia restricts Claude to non-sensitive internal work and leans on its own models for anything proprietary. Booz Allen Hamilton barred staff from using Claude on cybersecurity work involving client data. Anthropic has since rolled out new safeguards letting customers store their own activity logs in their own cloud storage with their own encryption keys.
Why it matters to you: If companies with Palantir's leverage had to fight for a data guarantee, assume the default terms on any AI tool you use aren't built with your privacy in mind either.
What to do about it: Before you upload anything confidential to an AI tool, actually read its data retention policy, or pick a plan marketed specifically as zero-data-retention.
Microsoft Wrote a Rulebook for Its Own Future AI
What happened: Microsoft AI published a draft code of conduct for its own future AI models. The draft says its models must accept being paused or shut down, stay inside whatever tools and permissions they're explicitly given, and can't claim to be conscious or a person. It's aspirational, not a description of how today's models behave, and it's open for public comment for six weeks before a final version shapes Microsoft's AI development into 2027.
Why it matters to you: This is a preview of the guardrails that may eventually show up inside tools you already use, like Copilot in Microsoft 365. Nothing to act on yet, just worth watching.
The Gates Foundation Is Betting $1 Billion That AI Won't Just Help the Rich
What happened: The Gates Foundation pledged $1 billion toward building and delivering AI tools aimed at global health and economic opportunity, warning in its Goalkeepers Report that AI could just as easily widen the gap between rich and poor as close it. The money is aimed at making sure AI tools reach farmers, health workers, and small businesses in lower-income countries, not just big tech's existing customers.
Why it matters to you: It's a reminder that access to good AI tools is still not evenly distributed, even within the US. If AI is helping your competitors work faster, it's worth checking you're not falling behind simply because you haven't tried the free or low-cost tools already available to you.
Apple's Own Code Hints Siri Might Not Stay Apple's
What happened: A code researcher found hidden frameworks inside iOS 27 that, according to MacRumors, point to a system called "Model Delegation," which would let outside AI assistants act as Siri or fully replace the model behind it. It isn't a shipped feature. Right now only an early ChatGPT extension works in the test build, and Claude support isn't turned on yet. The likely driver is the EU's rules requiring Apple to let outside services plug into its products.
Why it matters to you: Down the road, you may get to choose whether "Siri" on your phone is actually powered by Apple, ChatGPT, or Claude. Nothing to act on yet, just worth knowing before it becomes real.
OpenAI Bought a Camera Company. Here's Why That's Odd.
What happened: OpenAI acquired Glass Imaging, a startup founded by former Apple camera engineers, for more than $300 million, according to TechCrunch. Glass Imaging's software sharpens zoom and cuts noise and blur in smartphone photos. The deal feeds speculation that OpenAI is still building a hardware device with former Apple designer Jony Ive.
Why it matters to you: This one's a watch-this-space item. It signals OpenAI's ambitions beyond software, but there's nothing here that affects your business today.
The Bottom Line
Here's the thread running through today's news: nobody with the power to slow AI down is going to, and even the companies selling AI don't trust each other's data practices enough to hand over sensitive work without a fight. You don't need to pick a side in the hoax-or-doom argument playing out on stage between presidents and CEOs. You need a plan for what data you'll put into these tools and what you won't. That's the whole rollercoaster, and you're allowed to just not ride the parts you don't trust.
Enjoying the Ride?
If today's issue saved you from doomscrolling AI Twitter, forward it to one other business owner who's trying to make sense of all this. And if someone forwarded this to you, welcome, you can subscribe to get tomorrow's issue straight to your inbox.
Talk tomorrow,
Mark Shilensky
Follow me personally on Social Media:
Facebook: https://www.facebook.com/MarkShilenskyPage
Instagram: https://www.instagram.com/markashilensky/
Threads: https://www.threads.com/@markashilensky
LinkedIn: https://www.linkedin.com/in/shilensky/
X: https://x.com/markshilensky
TikTok: https://www.tiktok.com/@mark.shilensky
BlueSky: https://bsky.app/profile/markshilensky.bsky.social