In partnership with
Hi AI Futurists,
Anthropic built an AI so good at finding security flaws that they won't release it to the public. Meanwhile, last week's accidental code leak is still revealing what's coming next. Here's what both stories tell us.
Our agenda.
Top AI news
Anthropic's accidental reveal
3 AI tools to boost your workflow
AI in the wild
AI Investment Report
Best,
Lex Sokolin
P.S. Enjoying the newsletter? Have a suggestion? Hit reply and let us know!
Manage your settings: Share | Unsubscribe | Upgrade
Ecosystem: AI Venture Fund | Fintech Research | Lex Linkedin / Twitter
88% resolved. 22% stayed loyal. What went wrong?
That's the AI paradox hiding in your CX stack. Tickets close. Customers leave. And most teams don't see it coming because they're measuring the wrong things.
Efficiency metrics look great on paper. Handle time down. Containment rate up. But customer loyalty? That's a different story — and it's one your current dashboards probably aren't telling you.
Gladly's 2026 Customer Expectations Report surveyed thousands of real consumers to find out exactly where AI-powered service breaks trust, and what separates the platforms that drive retention from the ones that quietly erode it.
If you're architecting the CX stack, this is the data you need to build it right. Not just fast. Not just cheap. Built to last.
Top AI news
$122 billion at an $852 billion valuation. Amazon put in $50B, Nvidia and SoftBank each dropped $30B. That's not a fundraise. That's a GDP.
Budget video generation at $0.05 per second. Launched the same week Sora died. Google brought flowers to the funeral and a business card.
Apple is running its redesigned Siri on Google's 1.2 trillion parameter Gemini model. Nothing says "we'll do it ourselves" like borrowing your rival's brain.
A user says his private financial queries were shared with Meta and Google, even in "Incognito" mode. Quotation marks doing a lot of heavy lifting there.
Disney found out Sora was being killed less than an hour before the public did. Talk about getting ghosted.
AI on every phone, not just the $1,200 ones. Samsung thinks mass adoption starts in the mid-tier, and the math agrees.
The MTIA 300 through 500 are headed to Meta data centers by end of 2027. Announcing four chip generations at once is a flex and a deadline.
Silicon photonics and next-gen AI networking. Nvidia's strategy remains: if it touches AI, own a piece of it.
The demos are impressive. The FDA approvals? Still loading.
New compression tech that squeezes more performance out of less hardware. The AI diet everyone's been waiting for.
Anthropic's Mythos: When AI Gets Too Good to Release
Two weeks of Anthropic headlines, and they tell a connected story.
Last week was an accident. A routine update included Anthropic's entire Claude Code codebase, over 500,000 lines of proprietary code fully visible to anyone who looked. No customer data was exposed, but 44 unreleased features were. The standout was KAIROS, an always-on background agent that organizes your notes and context while you're away. There was also Buddy, a Tamagotchi-style virtual pet built into a coding tool, and hidden defenses designed to break anyone trying to copy the product.
This week was intentional. Anthropic announced Project Glasswing and previewed Claude Mythos, their most powerful model. Mythos has found thousands of previously unknown security flaws across every major operating system and browser, some hiding for decades. The oldest was 27 years old. In one test, it chained four vulnerabilities together to break out of a secured environment entirely on its own.
Anthropic's response was to lock it down. AWS, Apple, Google, Microsoft, JPMorgan, CrowdStrike, and others got early access. $100 million in usage credits committed. No plans to make it publicly available.
The code leak showed where AI tools are heading: background agents, persistent memory, systems that work for you when you're not asking. Mythos shows what happens when that capability reaches high-stakes domains. Anthropic is betting that getting these tools to defenders first beats waiting for the same capability to show up somewhere without guardrails.
Takeaways at a Glance:
Claude Mythos has found thousands of critical security flaws, some hidden for decades
Project Glasswing restricts access to partners including Apple, Google, Microsoft, and JPMorgan
$100M in credits and $4M to open-source security committed
No plans to make the model publicly available
Last week's code leak revealed 44 unreleased features including KAIROS and Buddy
What We Think About It:
AI tools are getting good enough that the companies building them are starting to pump the brakes on release. That's new. If you work with AI in any capacity, pay attention to how Anthropic handles Mythos access over the next few months. It will likely become the template other labs follow.
What You Can Do Right Now:
Watch how Project Glasswing's partner list expands. How Anthropic manages access to Mythos will set the playbook for how powerful AI gets distributed from here.
Get more done with these AI tools
Phygital+ - All-in-One AI Design Platform
50+ AI models on one infinite canvas for images, video, 3D, and text.
Set your brand guidelines once and every asset stays consistent.
Chain models together by dragging outputs between them. No tab switching.
Works with Zoom, Figma, and Canva workflows out of the box.
ElevenCreative - Creative Studio for Voice, Video, and Music
Voice, music, sound effects, image, and video generation in one workspace.
Localize content into 70+ languages from a single dashboard.
Edit and remix everything in the browser. No exporting to other tools.
Built for creators and marketers who are done juggling a dozen AI apps.
Raydian - AI-First Full-Stack Product Builder
Describe what you want in plain language. It builds the full stack.
Frontend, backend, database, auth, and hosting handled automatically.
Visual editor gives you design control without writing code.
Ships real software, not prototypes. Idea to live product in hours.
AI In the Wild
AI Investment Report
This 158-page research report provides the first comprehensive taxonomy of public companies, private ventures, and tokenized protocols building the infrastructure for autonomous AI systems. Compiled by Lex Sokolin, former Chief Economist at ConsenSys, fintech strategist at Autonomous Research, and current Managing Partner at Generative Ventures, this report delivers institutional-grade analysis of 100+ companies across 14 critical infrastructure layers. Learn more here.
AI & Robotics Industry Encyclopedia
In this 158 page report, we collate all the meaningful companies across public equities, private equity, and digital assets related to the machine economy. There are 13 sections of company types, a...
What type of coverage would you like to see most?
That’s all for today, folks!
Reach out to our audience by becoming a sponsor here.
If you’re enjoying the newsletter, share with a friend by sending them this link: 👉 https://www.futureblueprint.xyz/subscribe
Looking for past newsletters? You can find them all here.
Working on a cool A.I. project that you would like us to write about? Reply to this email with details, we’d love to hear from you!