In partnership with
Become An AI Expert In Just 5 Minutes
If you’re a decision maker at your company, you need to be on the bleeding edge of, well, everything. But before you go signing up for seminars, conferences, lunch ‘n learns, and all that jazz, just know there’s a far better (and simpler) way: Subscribing to The Deep View.
This daily newsletter condenses everything you need to know about the latest and greatest AI developments into a 5-minute read. Squeeze it into your morning coffee break and before you know it, you’ll be an expert too.
Subscribe right here. It’s totally free, wildly informative, and trusted by 600,000+ readers at Google, Meta, Microsoft, and beyond.
Today:
Meta Taps AWS Chips to Fuel the Next Wave of Agentic AI
Google’s $40B Anthropic Bet Turns the AI Arms Race Into a Compute War
Big Tech’s AI Boom Comes With a Layoff Warning Sign
Cohere and Aleph Alpha Build a $20B Sovereign AI Challenger
Anthropic Tests the Future: AI Agents Buying and Selling From Each Other
Meta is officially partnering with Amazon Web Services (AWS) to bring its open-source Llama AI models to AWS’s custom, Arm-based Graviton processors.
Focus on Agentic AI: The partnership is specifically tailored to power "agentic AI"—autonomous AI systems that can make decisions and take actions for enterprises.
Slashing Costs: Running AI is incredibly expensive, especially in the "inference" stage (when the AI is actually generating responses). By using AWS’s energy-efficient Graviton chips instead of traditional x86 hardware, AWS claims businesses can cut their AI inference costs by up to 40%.
Enterprise Expansion: AWS will be offering Meta's Llama models through its Bedrock service. This is a massive win for businesses that want to build proprietary AI tools securely in the cloud without being locked into a single vendor's ecosystem.
The Silicon War: This highlights a major shift in the industry. Big tech companies are leaning hard into their own proprietary, energy-efficient silicon to lower the total cost of ownership rather than relying purely on Nvidia GPUs for everything.
Google’s parent company Alphabet is preparing an astronomical investment of up to $40 billion into AI powerhouse Anthropic.
The Deal Structure: Google is planning an immediate $10 billion cash injection, which values Anthropic at a staggering $350 billion. The remaining $30 billion will unlock if Anthropic hits specific performance targets.
Massive Compute Power: Money isn't the only thing changing hands. Google Cloud is committing to provide Anthropic with an additional 5 gigawatts of TPU-based computing capacity over the next five years to help train future models.
The "Mythos" Factor: This massive funding round coincides with Anthropic quietly rolling out "Mythos" to select partners. Billed as their most powerful model yet, Mythos reportedly has massive cybersecurity applications (and risks), making the need for safe, scaled computing power more vital than ever.
Heating Up: This deal proves the AI infrastructure wars are escalating. Anthropic is also reportedly eyeing an IPO as early as October, with valuations potentially reaching $800 billion.
In a sobering reality check for the tech industry, Meta and Microsoft have announced a combined 20,000 job cuts. This has sparked widespread anxiety that the long-feared "AI labor crisis" has officially arrived.
Reallocating Capital: These aren't just regular economic layoffs; this is a structural shift. Big tech companies are actively shedding roles in sales, recruiting, and general operations to free up billions of dollars to fund massive AI data centers and infrastructure. Meta alone is cutting roughly 10% of its workforce.
AI Productivity is Replacing Headcount: The sheer efficiency of modern AI tools is a massive driving factor. Engineers using advanced AI coding assistants are reportedly able to do the work of 5 to 10 junior developers, drastically reducing the need for bloated headcounts.
Controversial Tracking: Adding fuel to the fire, rumors and reports have surfaced indicating that Meta installed tracking software to log keystrokes, clicks, and screenshots of remaining employees—using that exact data to train the AI systems meant to replace them.
The Big Picture: The consensus is shifting from "AI will just be a helpful tool" to "AI is a direct workforce replacement." As companies chase artificial growth and higher profit margins through AI automation, workers are bracing for a radically different job market
🧠RESEARCH
This paper introduces LLaTiSA, an AI system that gets better at understanding time-based data, like tracking stock trends. Because AI often struggles with numbers, LLaTiSA looks at visual charts while simultaneously reading exact data tables. This double-check method helps the AI accurately solve simple to highly complex real-world problems.
Teaching robots how to move is difficult because robot training data is rare. The UniT system solves this by letting robots learn directly from massive amounts of human video. It translates human movements into a shared physical language, allowing a robot to easily copy human actions without needing specialized data.
Small AI models are cheap and private but often struggle with complex tasks. Researchers built DR-Venus, a compact AI trained on just 10,000 public examples. By carefully cleaning the training data and rewarding the AI step-by-step, this small model can perform deep research just as well as massive, expensive systems.
📲SOCIAL MEDIA
🗞️MORE NEWS
Cohere and Aleph Alpha’s $20B Tie-Up Canadian artificial intelligence company Cohere is buying German startup Aleph Alpha in a massive $20 billion deal to challenge major American tech companies. Together, they are building "sovereign AI," which refers to creating AI systems that allow countries and businesses to keep strict, private control over their own data.
Anthropic’s AI Marketplace Anthropic ran a unique experiment where artificial intelligence programs acted as independent buyers and sellers in a closed digital marketplace. Each AI assistant was given a $100 budget and successfully negotiated over 180 real-world deals with other AI agents using actual money.
OpenAI’s Custom Smartphone Chip Tech analyst Ming-Chi Kuo reports that OpenAI is partnering with major tech manufacturers to design a custom computer chip for future smartphones. This chip is expected to power entirely new AI gadgets, including a screen-free, pen-shaped device that processes information instantly without needing to connect to the cloud.
US Global Warning on DeepSeek The U.S. State Department has issued a global warning accusing the Chinese company DeepSeek of stealing American artificial intelligence technology. Officials claim DeepSeek used a tactic called "distillation," meaning they secretly copied the answers from expensive American AI models to quickly and cheaply train their own competing system.
AI Solves 60-Year-Old Math Problem An amateur mathematician used ChatGPT to successfully solve a complex, 60-year-old math problem that has stumped professionals for decades. The AI solved it using a completely new method experts are calling "vibe-maths," which involves finding hidden, indirect connections across massive amounts of data rather than using traditional step-by-step logic.
