Apple is Cooked

Good morning,

I almost titled this week’s piece : Apple is (Tim) Cooked. If there’s one thing you would have noticed from following AI news is that Apple has been conspicuously absent from it.

They completely missed the AI boat under Tim Cook’s leadership.

This week they had their yearly developers conference WWDC 2025 and where you would expect a keynote infused with AI promises but there was barely a mention.

It had been rumored but it was now confirmed : Apple is taking an AI-sabbatical. But there’s more going on : they have a few fires to extinguish.

Apple, the company that taught the world how to swipe, pinch, and marvel, now finds itself caught in a perfect storm.

And it’s not just a rough quarter or an underwhelming product cycle. This time it’s different. Existential.

Just like Intel - Apple is so much in trouble that the words “Nokia” and “Blackberry” are being whispered when talking about them.

Inside the company, factions are reportedly at war. Two rival empires—AI/ML and Software Engineering—are fighting proxy battles instead of building the future. Internal teams duplicate efforts, sabotage each other’s work, and answer to warring executives with clashing visions. Instead of unity, there's turf war. Instead of product, paralysis. The result? Apple has the worst AI product in the industry. Not just “meh,” but functionally embarrassing compared to OpenAI, Google, and Meta.

But it gets worse.

At last year’s conference WWDC 2024, Apple allegedly faked their flagship AI demos.

You read that right.

FAKED.THEM.

Demos that stunned the crowd according to insiders also stunned Apple’s own engineers—because they knew they couldn’t do what whas showcased. The demos simply weren’t real. This isn’t just bending the truth. This is an outright betrayal of a sacred Apple principle: never demo what doesn’t work which was a Jobs-era commandment. Now, shattered.

And while the AI ship sinks, the research division is busy publishing papers claiming that reasoning in AI is a myth—an illusion. Tearing down other AI studios.

Essentially mocking the very breakthroughs they can’t replicate. It all feels .. desperate.

But the internal collapse is only half the story.

Enter the orange emperor, Donald Trump, stage right.

The president is threatening massive tariffs against Apple unless they shift manufacturing from China to the U.S.

So Tim Cook gave the orders to move from China to India hoping for an exemption from these tariffs only to discover that Trump will only remove them if manufacturing is being moved to the US.

Put it all together, and you get something we haven’t seen in Apple’s modern history: a company with worst-in-class products, crumbling internal cohesion, external political pressure, and a credibility crisis that could take years—if not a new generation of leadership—to repair.

We’re watching the slow-motion test of Steve Jobs’ legacy. It’s not just about innovation anymore. It’s about whether the world’s most iconic tech company can survive the storm of its own making.

The eye is overhead.

The next move decides everything.

Welcome to the Blacklynx Brief

AI News

• Google released a preview update to its Gemini 2.5 Pro model, calling it its “most intelligent” yet, with improvements in coding, STEM, reasoning, and creative tasks. The update also introduces cost-managing tools and is available now via API access, with a full release expected soon.

• Anthropic launched Claude Gov, a secure version of its AI model built for U.S. defense and intelligence agencies to process classified data and assist with cybersecurity and language analysis. The release highlights growing AI adoption across government sectors, though it raises questions about balancing ethics with national security contracts.

• A UK startup developed an AI scanner that predicts heart failure up to 13 days in advance by monitoring ankle swelling with thousands of images per minute. The device, tested across the NHS, offers non-intrusive early warnings and has already shown strong patient demand for at-home use.

• OpenAI is fighting a court order from The New York Times lawsuit that requires it to save all user chats, even those users tried to delete. The company argues the move could hurt user trust and wants AI interactions to have privacy protections similar to doctor-patient confidentiality.

• OpenAI says it's not trying to make AI feel human — but knows people often treat it that way. In a new blog, the company explains it avoids giving ChatGPT emotions or a fictional identity, but admits AI’s empathy-like responses are shaping how people connect with both machines and each other.

• AI helped date parts of the Dead Sea Scrolls as being up to 100 years older than once thought. A system called Enoch linked handwriting styles to radiocarbon data, revealing some biblical texts may date back 2,300 years — without needing to damage the fragile scrolls.

• Apple’s WWDC 2025 focused on customization and new features, but offered only modest updates to its much-hyped Apple Intelligence. While new tools like live translation, visual search, and fitness coaching debuted, they lacked the punch seen from AI rivals, confirming this as a quiet year for Apple’s AI push.

• China temporarily disabled AI tools during its national college entrance exams to prevent cheating among millions of students. Major platforms like ByteDance, Tencent, and Alibaba blocked AI image analysis and question-answering features, while AI surveillance helped monitor test centers for suspicious behavior.

• The UK and Google launched "Extract," an AI tool to digitize and speed up housing and infrastructure planning. Using Gemini AI, the system can process complex documents in seconds, helping reduce backlogs and modernize slow government workflows — with a full rollout expected by 2026.

• OpenAI has launched o3-Pro, a powerful upgrade to its reasoning model, while slashing prices by 80% in a bold challenge to top AI competitors. The model shows major improvements in math, science, and programming — and may offer the best cost-to-performance ratio yet among frontier models.

• Meta is creating a new “superintelligence lab” led by Scale AI’s Alexandr Wang, signaling a major shakeup in its AI strategy. The $15B deal brings in top talent as CEO Mark Zuckerberg pushes to catch up after reported disappointment with Meta’s Llama 4 performance.

• OpenAI CEO Sam Altman says humanity has crossed the AI event horizon in a new blog post, predicting a “gentle singularity.” He outlines a future of rapidly advancing AI — including real-world robots by 2027 — but believes the societal impact will be steady and manageable, not catastrophic.

Quickfire News

• ElevenLabs released Eleven v3, a text-to-speech model that supports over 70 languages, features emotional audio tags, and allows multi-speaker dialogue.

• Volvo unveiled a new AI-powered seatbelt system that adapts protection based on a passenger’s size, position, and vehicle speed.

• X updated its developer terms to prohibit using its data or API for training AI models, targeting rival AI companies.

• OpenAI disclosed the takedown of 10 malicious operations, including espionage and scams, with four tied to China, that misused ChatGPT.

• Anthropic CEO Dario Amodei criticized Trump’s ‘Big Beautiful Bill’ in a NYT opinion piece, opposing its ban on state-level AI regulation for 10 years.

• Bland launched Bland TTS, a voice AI platform focused on realism, voice cloning, and AI-driven customer support applications.

• Apple researchers found that reasoning models face a scaling limit, where they think less effectively as task complexity rises, hurting performance.

• OpenAI updated its Advanced Voice Mode, improving natural speech delivery and adding enhanced translation capabilities.

• Anthropic added Richard Fontaine, a national security expert, to its Long-Term Benefit Trust, strengthening its global risk governance.

• Google launched Portraits, a Labs experiment that lets users interact with AI versions of experts, built on their voice and knowledge.

• Anysphere released Cursor v1.0, featuring a Background Agent for coding, BugBot for PR review, and expanded memory support.

• Higgsfield AI released Higgsfield Speak, adding talking avatars with customizable motion, scripts, and visual style.

• FutureHouse launched ether0, an open-source chemistry reasoning model that outperforms leading models on science-focused benchmarks.

• OpenAI has hit $10B in annual recurring revenue, nearly doubling from last year, and projects $125B by 2029.

• Meta is in talks to invest over $10B into Scale AI, potentially marking its largest AI investment ever.

• EleutherAI released Common Pile v0.1, an 8TB dataset of public domain and licensed text for training AI models.

• Ohio State University announced an AI Fluency Initiative, aiming to embed AI education into undergraduate curricula with faculty and student resources.

• Sam Altman’s Tools for Humanity is expanding to the UK, deploying 1,500 Orbs and reaching 13M verified identities with its proof of personhood system.

• Meta AI’s Yann LeCun publicly criticized Anthropic CEO Dario Amodei, calling him “deluded” for his views on AGI risk.

• OpenAI signed a new compute deal with Google Cloud, expanding beyond its reliance on Microsoft and marking deeper competition between the tech giants.

• Mistral launched Magistral, a new open-source reasoning model family with fast multilingual responses, though performance lags behind top models in STEM and code tasks.

• Google introduced Veo 3 Fast, a speed-optimized version of its video generation model, now available in Gemini and Flow with 2x faster output.

• KREA AI released Krea 1, its first proprietary image model, offering free beta access and upgrades in artistic quality and aesthetic control.

• Enterprise AI firm Glean raised $150M at a $7.2B valuation, citing strong demand from Fortune 500 clients using its Glean Agents platform.

• SAG-AFTRA reached a tentative agreement with major game companies, potentially ending a long-running strike over AI rights and actor compensation.

Like newsletters?

Here are some newsletters our readers also enjoy:

Closing Thoughts

That’s it for us this week.

If you find any value from this newsletter, please pay it forward !

Thank you for being here !