Good morning. Influencers have always been a little… fake. Filters, Photoshop, sponsored “authentic” posts. But this year, being fake is officially a competitive sport. A new contest called “AI Personality of the Year” has kicked off, courtesy of OpenArt x Fanvue and backed by AI voice startup ElevenLabs. Judges will score virtual influencers on authentic storytelling, follower engagement, brand appeal, and even correct finger counts. Organizers call it the “Oscars for AI personalities,” which is either peak internet culture or proof we’ve officially run out of good ideas.
OpenAI’s latest energy deal is straight up science fiction. They’re negotiating with Sam Altman’s startup, Helion Energy, to buy 12.5% of its fusion power – about 5 gigawatts by 2030, scaling up to 50 gigawatts by 2035. To hit these targets, Helion must quickly build hundreds (and eventually thousands) of commercial fusion reactors. There’s just one tiny problem. Helion hasn’t achieved “scientific breakeven,” meaning its reactors currently consume more energy than they produce. Altman recently stepped down from Helion’s board to avoid conflicts of interest. Now all Helion needs to do is make fusion work. No pressure. [Axios]
Liminal wants to become the corporate condom for AI. It sits between employees and models like OpenAI and Anthropic, masking sensitive data before it leaves the building, then restores it on the way back. The idea is simple: let banks, hospitals, and compliance-obsessed giants use frontier AI models without leaking trade secrets, customer data, or their last shred of legal sanity. Way smarter than hosting and updating your own private AI. But Liminal’s bigger play is observing how your team actually works, then automatically spinning up agents to handle repetitive tasks. First it protects your prompts. Then it watches your workflow and finds exactly what to automate next. How cool is that? [Liminal]
SOC 2 isn’t paperwork. It’s revenue. Every day without compliance equals deals you’ll never close.
Comp AI helps companies get audit-ready in hours, not months. No spreadsheets. No consultants. Comp AI agents automate the busywork for easy compliance in record time. 100% guaranteed.
Win more deals. Book a demo today.
Anthropic thinks it can beat OpenClaw at its own game. OpenClaw went viral by letting people control powerful AI agents from their phones. Anthropic’s betting it can simplify the idea, add security, and take it mainstream. This week it launched Channels, a feature that connects a running Claude Code session directly with messaging apps like Telegram so developers can work anywhere. A bot receives a message, executes tasks on a desktop, then replies back in the chat. All that’s missing is the cult following. [Claude]
The Pentagon used to worry AI would make war messier. Now it’s all in. Project Maven started as a tool for scanning drone footage, but it’s grown into full-blown battlefield software powered by Palantir. It helps the US military quickly identify targets and move from detection to strike in minutes. What once triggered protests at Google is now deployed in real combat missions. The pitch is speed. The problem is what comes with it. Faster decisions with fewer safeguards and shaky oversight. Automating warfare is easy. Figuring out who’s responsible when something goes wrong is another story. [Wired]
Reach your audience
Get in front of 70,000+ tech decision makers with budget and technical influence. 95% US based. 66% open rate. Read daily by AI leaders and builders at high growth companies.
Nobody’s buying tech companies anymore. Bankers had big expectations for tech M&A this year, but the deals never happened. CEOs froze, terrified AI might make billion dollar acquisitions obsolete. Deal volume plunged, propped up mostly by Musk dropping $250 billion on his own AI startup. AI panic isn’t the whole story, though. The stock market is shaky, borrowing costs are high, and private equity can’t raise funds. Still, there’s a bright spot. Companies building data centers, power plants, cooling systems, and real estate are suddenly hot commodities. At least someone’s still cashing in on AI hype. [The Information]
fun stats
🏆 17.5%. Guaranteed minimum return OpenAI is offering private equity investors. They’re also promising early access to new models, hoping to raise enough cash to beat Anthropic in the enterprise AI race.
🚨 1 in 3. Enterprises were hit by a security incident involving AI agents in 2025. Despite running an average of 13 security tools, most CISOs admit they can’t spot the difference between human and bot activity – and have little idea what data AI is actually accessing.
🛍️ 3x. How much less likely shoppers were to complete a purchase through ChatGPT’s in-chat checkout versus Walmart’s own site. Next, Walmart’s going to try putting their chatbot Sparky inside ChatGPT to boost sales. So a chatbot inside a chatbot, trying to beat a plain old website. Progress.