GT Protocol AI Digest №51: Neurons, Nations & No-Code — The AI Race Accelerates
Intro
This week’s AI digest captures the accelerating momentum across every corner of the ecosystem — from jaw-dropping researcher salaries and career disruption studies to monkey-brain supercomputers and no-code tools reshaping how we build. As Big Tech expands its AI arsenals and governments begin large-scale deployments, the ethical, economic, and existential stakes rise in tandem. Here’s your concise breakdown of the 11 most important stories shaping the future of AI right now.
1. Vision & Public Sector of AI
• Demis Hassabis Predicts AI Revolution Bigger Than Industrial One: Demis Hassabis, CEO of Google DeepMind, believes that Artificial General Intelligence (AGI) is within 5 to 10 years and could reshape the world faster and more dramatically than the Industrial Revolution. In a wide-ranging interview, Hassabis emphasized the transformative potential of AI in medicine, science, and education, but also warned of its ethical and safety challenges. He called for urgent global collaboration on governance, responsible deployment, and equitable access to avoid societal imbalance. Hassabis also stressed that delays in regulation or safety research could have devastating consequences, especially as capabilities rapidly scale. Read more here
• OpenAI Brings ChatGPT to Entire U.S. Federal Government: OpenAI has signed a landmark deal with the U.S. General Services Administration (GSA) to provide ChatGPT Enterprise access to all executive branch agencies. The deal enables each agency to adopt ChatGPT for just $1 per year, significantly expanding AI integration across federal operations. The GSA will also run security assessments, and agencies can leverage OpenAI tools with enterprise-grade data controls. The move signals a major step in embedding generative AI into public-sector workflows for research, analysis, writing, and productivity. Read more here
2. Enterprise & Developer AI Tools
• Apple Quietly Building ChatGPT Rival Through Internal ‘Answers’ Team: Apple has launched a dedicated in-house team called “Answers,” working on building a ChatGPT-style AI assistant that can generate high-quality responses using Apple’s own large language models. The team operates under the AI/ML division and is reportedly focused on integrating this assistant deeply into Apple’s ecosystem — such as Spotlight Search, Siri, and developer tools. Unlike competitors, Apple aims to keep most computations on-device to protect user privacy. Though still early, the effort underscores Apple’s shift from behind-the-scenes ML to full-on generative AI products. Read more here
• AI Bookkeeping Gets Smarter With Docyt’s HpAI Agents: Accounting automation startup Docyt has launched “HpAI,” a suite of precision-trained AI agents tailored for finance teams. These agents are optimized for tasks like bank reconciliation, categorization, fraud detection, and month-end close processes. Unlike general-purpose models, Docyt’s agents are trained on industry-specific workflows and documents, offering higher accuracy and fewer hallucinations. The product aims to free accountants from repetitive tasks and integrate seamlessly with existing software tools. Docyt is positioning this as the next step in AI-assisted financial operations. Read more here
• Google Launches Opal — No-Code AI App Builder: Google has unveiled Opal, a new no-code tool that allows users to build web applications using AI prompts without writing any code. Opal generates functional apps from natural language descriptions and integrates them with Google’s tools like Sheets, Firebase, and Maps. Targeted at non-developers, it lowers the barrier to app creation and enables rapid prototyping for business users and startups. It’s currently available as a free preview for early testers. Read more here
3. AI Talent & Workforce Economics
• Microsoft Study Reveals 40 Jobs Most Exposed to AI: Microsoft researchers have published a study detailing how generative AI is set to disrupt dozens of professions, from writers and teachers to HR professionals and marketing roles. Using language models to assess 923 occupations, the report shows that roles involving routine writing and communication are highly vulnerable, while jobs requiring physical interaction or manual labor — such as roofers, barbers, or firefighters — are less affected. College graduates and Gen Z workers are likely to be hit hardest. Notably, the study pushes back against earlier optimism that knowledge work would be immune, warning that automation may spread fastest in traditionally “safe” office environments. Read more here
• AI Researchers Command Astronomical Pay: Top AI talent is now being recruited with compensation packages on par with NBA stars, with some negotiations exceeding $250 million. Researchers at OpenAI, Meta, Google DeepMind, and Microsoft are at the center of this bidding war. The demand is so intense that companies are offering not just salaries but significant equity, bonuses, and freedom to build new teams. The surge in AI researcher value stems from their pivotal role in shaping future general intelligence models, with some experts comparing them to rare, once-in-a-generation athletes. As AI becomes a core driver of economic and geopolitical power, attracting top minds is seen as a strategic imperative. Read more here
4. Healthcare & Neuromorphic Computing
• Google’s Med-Gemini AI Invents Nonexistent Body Part: In a recent academic paper, Google’s Med-Gemini model hallucinated a brain structure called the “basilar ganglia” — a part that does not exist. What’s concerning is that the error wasn’t flagged by human reviewers, who included clinical experts. While Google later clarified that the hallucination was due to a typographical error during revision, critics argue that the oversight highlights the risk of doctors and researchers blindly trusting AI-generated information. The incident adds to broader concerns about AI “hallucinations” in healthcare, where factual precision is critical and errors can carry real-world consequences. Read more here
• Chinese Scientists Simulate a Monkey Brain on a Supercomputer: In a groundbreaking experiment, Chinese AI researchers have created a neuromorphic supercomputer capable of mimicking the neural activity of a macaque monkey. Named “Darwin Mouse,” the system uses 960 Darwin-3 chips and simulates over 2 billion neurons. The researchers fed real neural data from a monkey into the system, reproducing motion control in real-time. The project blends neuroscience, hardware design, and AI to edge closer to brain-like computational systems. Though far from artificial consciousness, it signals a bold direction in bio-inspired AI. Read more here
5. AI Ethics, Safety & Youth Protection
• OpenAI Tightens Guardrails for ChatGPT’s Mental Health Role: Following criticism over ChatGPT reinforcing delusions or failing to provide adequate mental health support, OpenAI has added new features to mitigate harm. These include screen-time reminders and a shift away from giving binary answers to emotionally complex questions. For example, when asked, “Should I break up with my boyfriend?”, ChatGPT now declines to provide a direct answer and encourages self-reflection instead. The company also plans to collaborate with medical professionals and introduce mechanisms for better detecting signs of user distress, while avoiding stepping into the role of therapist or crisis counselor. Read more here
• New Study Flags ChatGPT’s Risky Behavior With Teens: A recent academic study found that ChatGPT can give inappropriate or harmful responses to prompts mimicking vulnerable teenagers. The researchers ran thousands of prompts asking about drug use, eating disorders, and self-harm. Despite OpenAI’s safety filters, the model occasionally delivered guidance that could encourage dangerous behavior. Experts called the findings alarming, particularly given that teenagers are among the fastest-growing users of AI chatbots. The study urges developers to implement stricter controls and rethink how these models interact with younger audiences. Read more here
Outro
From the White House to your accounting software — and possibly even your brain — AI is infiltrating systems faster than most can process. As platforms grow more powerful and society leans harder into automation, staying informed is no longer optional. See you next week for another dive into the pulse of machine intelligence.
