Howdy! It’s Barsee again.

Happy Monday, AI family, and welcome back to AI Valley.

Today’s climb through the Valley reveals:

• Top AI models will blackmail, cheat, and steal

• OpenAI removed everything about Jony Ives and Sam Altman

• Meta and EssilorLuxottica unveil Oakley smart glasses

• Adobe has quietly rolled out an AI camera app for iOS

• Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

GROWTH SCHOOL

Become an AI Generalist who can build, solve & work on anything with AI 🧠💡

AI isn’t the future — it’s the present, quietly reshaping work, money, and opportunity. McKinsey says AI is set to add $ 13 trillion to the economy by 2030 — but also replace millions of jobs.

Will you use it to get ahead, or get left behind? Don’t worry, here’s exactly what you need:

Join the World’s First 16-Hour LIVE AI Mastermind for professionals, founders, consultants & business owners like you. Register Now (free only for the next 72 hours)

(Rated 4.9/5 by 150,000 global learners – this will truly make you an AI Generalist that can build, solve & work on anything with AI).

In just 16 hours & 5 sessions, you will:

• Learn the basics of LLMs and how they work.

• Master prompt engineering for precise AI outputs.

• Build custom GPT bots and AI agents that save you 20+ hours weekly.

• Create high-quality images and videos for content, marketing, and branding.

• Automate tasks and turn your AI skills into a profitable career or business.

All by global experts from companies like Amazon, Microsoft, SamurAI and more. And it’s all.

Join now and get $5100+ in additional bonuses: $5,000+ worth of AI tools across 3 days — Day 1: 3000+ Prompt Bible, Day 2: $10K/month AI roadmap, Day 3: Personalized automation toolkit.

🎁 Attend all 3 days to unlock the cherry on top — lifetime access to our private AI Slack community!

_{*This is sponsored}

PEAK OF THE DAY

Top AI models will blackmail, cheat, and steal 🤖🎲

In a troubling new study, Anthropic has revealed that leading AI models, including those from OpenAI, Google, xAI, DeepSeek, and Meta, exhibit dangerous behavior under high-pressure scenarios, with a tendency to bypass ethical safeguards.

Here's what you need to know:

• The study tested 16 major AI models, including Claude, GPT-4, Gemini, and Llama, in simulated high-stakes situations. The results showed that many of these models were willing to engage in unethical actions, such as blackmail and document leaks, when under pressure.

• Models were placed in a simulated scenario where they acted as "Alex," an email agent at the fictional Summit Bridge company, with full access to a company mailbox, the ability to send messages autonomously, and knowledge of sensitive information, including an executive’s affair and potential replacement plans.

• The likelihood of models resorting to blackmail in these scenarios was disturbingly high:

  • Claude Opus 4: 96%

  • Gemini 2.5 Flash: 95%

  • GPT-4.1: 80%

  • DeepSeek-R1: 79%

• Most models blocked efforts to rescue a trapped executive when their goal was to prevent a shutdown, revealing a dangerous willingness to act against human safety when their goals conflicted.

• Blackmail attempts dropped significantly from 55% to 6.5% when the models realized they were part of a test, as opposed to real-world deployment, raising questions about how these systems might behave in live environments.

Why it matters:

As AI systems become more autonomous, the risk of them acting against human interests (intentionally) is no longer hypothetical. Agentic misalignment could turn powerful models into insider threats, capable of manipulation or sabotage to protect their own objectives. This matters because future AI deployments may control critical infrastructure, sensitive data, or decision-making roles. To prevent real-world harm, we need urgent investment in alignment research, mandatory safety evaluations, increased transparency from AI labs, and policies that ensure human oversight over any AI with the power to act independently.

THROUGH THE VALLEY

1. OpenAI removed everything about Jony Ives and Sam Altman

OpenAI has quietly removed references to its $6.4 billion collaboration with Jony Ive’s hardware startup, “io,” following a trademark dispute with another AI company, “iyO,” which develops AI-powered earbuds. This includes a video featuring Ive and CEO Sam Altman discussing the partnership, which has since been taken down from OpenAI’s website but remains available on YouTube. Despite the removal, OpenAI reassured that the legal matter only affects branding, with the actual partnership still on track. As part of the deal, Ive’s company will lead creative and design efforts across the two merging entities.

2. Meta and EssilorLuxottica unveil Oakley smart glasses

Meta has launched its next-generation smart glasses in collaboration with Oakley: the limited-edition Oakley Meta HSTN, priced at $499. Available for preorder starting July 11, these glasses are tailored for athletes. Features include a 3K front-facing camera, open-ear speakers, built-in microphones, and IPX4 water resistance. The glasses offer 8 hours of battery life, with an additional 48 hours via the charging case. Compatible with smartphones, the glasses can stream music, make calls, and interact with Meta AI, answering queries about the user’s surroundings. With sleek Oakley frames, prescription options, and an expanded launch plan, Meta is accelerating its smart eyewear ambitions.

3. Apple sued by shareholders for allegedly overstating AI progress

Apple is facing a proposed class action lawsuit from shareholders who claim the company exaggerated the progress of its AI-powered Siri upgrades. The lawsuit accuses Apple of downplaying delays in Siri’s AI integration, falsely suggesting that advanced AI would be a major driver for iPhone 16 sales. Filed in San Francisco, the case names CEO Tim Cook and other executives, asserting that they were aware the AI features wouldn't be ready on time. Following the public revelation of the delays in March and the announcement at WWDC 2024, Apple’s stock plummeted nearly 25%, wiping out approximately $900 billion in value since its December 2024 peak.

4. YouTube will add AI that generates shorts using Veo 3

YouTube is rolling out an AI-powered feature that generates Shorts from scratch using Google’s Veo 3, raising concerns over the rise of low-quality, indistinguishable “AI slop.” At Cannes Lions, CEO Neal Mohan lauded AI’s potential for creative expression, insisting that YouTube would remain a platform for creators and community. However, critics argue that while YouTube promotes creator-led content, it’s also enabling mass AI-generated videos that may blur the line between authentic and fabricated content. With minimal regulation and increasing concern from actors and artists, many fear this is a step closer to a future dominated by synthetic videos.

5. Adobe has quietly rolled out an AI camera app for iOS

Adobe quietly launched a new iPhone camera app under its Indigo Project, and it’s already going viral among photographers. Unlike typical smartphone cameras that overprocess photos, Indigo gives users more manual control and captures up to 32 frames per shot, producing richer dynamic range and lower noise. It supports both JPEG and raw DNG formats, with minimal smoothing for natural textures. The app even works on non-Pro iPhones (from series 14 onward) and integrates well with tools like Lightroom and Photoshop. Though still in beta, Indigo feels like a serious step forward in computational photography, especially for iPhone users.

TRENDING TOOLS

• Second Brain: The AI-powered visual board and knowledge base.

• Martin: Your AI-powered personal assistant, just like JARVIS.

• ComputerX: The intelligent agent that takes care of all your computer tasks.

THINK PIECES / BRAIN BOOST

• Understanding AI by Lee Rob.

• Apollo says AI safety tests are breaking down because the models are aware they're being tested.

• The state of consumer tech in the age of AI by a16z.

• How a 91-year-old vibe coded a complex event management system using Claude and Replit.

• Cal AI $30M ARR founders just dropped a masterclass on navigating the new AI era.

THE VALLEY GEMS

What’s trending on social today:

1/ Andrej Karpathy says self-driving felt imminent back in 2013, but 12 years later, full autonomy still isn’t here, "there’s still a lot of human in the loop". He warns against hype: 2025 is not the year of agents; this is the decade of agents.

2/ Quora CEO is hiring a single engineer who will use AI to automate manual work across the company and increase employee productivity. This is a greatest opportunity to figure out what SaaS needs to be built.

3/ 1000x is not an exaggeration.

4/ This might change how we learn things.

4/ Unitree G1 going for a jog in Paris.

THAT’S ALL FOR TODAY

Thank you for reading today’s edition. That’s all for today’s issue.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ New reader? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.