NewBits Digest

China may have just taken another major leap in the global AI race. ZTE and ByteDance have unveiled a prototype smartphone—the Nubia M153—featuring the world’s first agentic AI smartphone system capable of operating the entire device like a human. Unlike Siri, Google Assistant, or even next-gen AI features from Apple and Samsung, this device runs an AI that sees the screen, taps buttons, fills forms, books services, makes payments, and even negotiates with other bots. 📌 Agen

A Guardian investigation highlights growing concern in the academic AI community after a single researcher, Kevin Zhu, claimed authorship on 113 AI papers this year, with 89 accepted at the prestigious NeurIPS conference. The episode has sparked debate about quality, integrity, and overload in modern AI research—fueling what many now call an emerging AI research slop crisis. 📌 AI Research Slop Crisis: The Details 👤 One Author, 113 Papers Kevin Zhu, a recent Berkeley graduat

Chinese AI startup DeepSeek has released V3.2 and V3.2-Speciale, two frontier-level reasoning models that match or surpass GPT-5 and Gemini 3 Pro—while staying open-source and dramatically cheaper. The DeepSeek V3.2 launch instantly positions the company as a global contender reshaping the economics of frontier AI. 📌 DeepSeek V3.2: The Details 🧮 Frontier Performance at the Top V3.2 delivers performance on par with leading models like GPT-5, 4.5 Sonnet, and Gemini 3 Pro acro

OpenAI CEO Sam Altman has told employees the company is entering an internal OpenAI code red surge, rapidly accelerating improvements to ChatGPT after Google’s newest upgrades. The company is shifting priorities and fast-tracking a new model internally codenamed Garlic. 🧠 OpenAI Code Red: Key Details 📌 Major Feature Push Altman called this “a critical time for ChatGPT,” directing teams to upgrade personalization, image generation, and overall capability. 🚀 New Model Next W

Anthropic is reportedly positioning itself to go public as early as 2026, according to the Financial Times. The Claude maker has hired the same law firm behind Google’s and LinkedIn’s IPOs as investors push for a listing before OpenAI. The Anthropic IPO is quickly becoming one of the most closely watched moves in the AI race. 📌 Anthropic IPO: The Details 🏛️ Elite Legal Team Secured Anthropic has tapped Wilson Sonsini, the powerhouse firm behind Google’s and LinkedIn’s IPOs,

A University of Florida researcher told state lawmakers that AI in Florida agriculture could boost production by up to 35% by 2030, calling automation the only realistic answer to a shrinking labor force and rising production challenges. UF also broke ground on a new Center for Applied AI in Agriculture to accelerate this transition. Key Points — AI in Florida Agriculture 📈 AI-driven spraying, harvesting, and monitoring systems can reduce pesticide use and increase yields th

In a new interview promoting the next Avatar installment (widely reported as Fire and Ash), director James Cameron warned that generative AI replacing human actors is “horrifying,” even as he continues to support AI tools that lower VFX costs and improve filmmaking workflows. Cameron says motion capture celebrates the authenticity of real performers — while gen-AI systems that fabricate actors or performances from text prompts do the opposite. What He Said — James Cameron on

Major retailers and tech giants — including Amazon, Walmart, Google, Target, and OpenAI — are rolling out new AI shopping assistants this holiday season, aiming to make gift buying far easier while capturing a larger share of online spending. What’s New — AI Holiday Shopping 🧠 Smarter AI assistants now give personalized recommendations, track prices, compare products, and even place or initiate orders through natural conversations. 🛒 OpenAI’s ChatGPT shopping mode generates

Anthropic has released Claude Opus 4.5, its new top-tier model designed to compete directly with Gemini 3 and GPT-5.1. The model delivers major jumps in coding, reasoning, and multi-agent orchestration — at a dramatically lower price point. Key Highlights — Claude Opus 4.5 🧩 According to Anthropic, Claude Opus 4.5 is the first model to break 80% on SWE-Bench Verified, setting a new high watermark for coding accuracy. 🛠️ Internal evaluations show top scores in tool use, reas

Former OpenAI researcher Andrej Karpathy is urging educators to abandon efforts to catch AI-generated homework, arguing that today’s detection tools are fundamentally broken — and that grading needs to shift back into the classroom. What He Said — AI Homework Detection ❌ AI detectors “don’t work” and are “doomed to fail.” ✍️ He warned that modern AI systems can mimic student style closely enough that detection is unreliable. 🏫 He recommends moving graded work back into in-sc

DeepSeek just released Math-V2, an open-source MoE model that delivers research-level mathematical reasoning once locked behind proprietary walls. What It Achieved — DeepSeek Math V2 Performance Benchmarks DeepSeek Math V2 has demonstrated breakthrough performance across competitive mathematics benchmarks: 🏅 Solved 5 out of 6 IMO 2025 problems — gold-medal tier 📈 Scored 118/120 on the 2024 Putnam (under scaled test-time compute), surpassing top human competitors 📚 Achieved

A statewide deep dive into the future of artificial intelligence Florida lawmakers are preparing for a comprehensive review of artificial intelligence as the technology continues reshaping education, business, and public safety. House Speaker Daniel Perez has officially designated the week of December 8 as “Artificial Intelligence Week” in the Florida House, where committees will explore AI’s benefits, risks, and long-term implications ahead of the 2026 legislative session. �

The Trump administration is reportedly drafting a Trump AI Executive Order aimed at undercutting state-level AI regulations and pushing toward a single federal standard. The draft would direct the Department of Justice to sue states and instruct federal agencies to tie certain grants to compliance if states move ahead with their own AI rules—a proposal already causing intense debate across the political, tech, and legal worlds. 🧩 Key Points on the Trump AI Executive Order ⚖️

A new Guardian analysis highlights a counter-narrative to the fear that AI will weaken democracy. While risks remain real, researchers Nathan E. Sanders and Bruce Schneier argue that AI can also distribute power, amplify civic engagement, and make democratic systems more efficient—if deployed responsibly. Here are four real-world examples showing how AI is being used to strengthen, not undermine, democracy—and what that tells us about the evolving relationship between AI and

Former Disney Channel actor Calum Worthy has ignited major controversy after promoting 2wai, a new AI app that allows users to create digital avatars of deceased loved ones—a concept many say feels ripped straight from Black Mirror. 🧩 What Happened With 2wai 🎭 The Viral Promo: Worthy shared an ad featuring a pregnant woman video-chatting with her mother for maternity advice… only to reveal the mother is dead and the on-screen figure is a 2wai-generated AI avatar. ⏱️ “Three

Google just unveiled Nano Banana Pro, its next-generation image model built on Gemini 3, and it’s redefining what’s possible in visual creation. The model delivers professional-grade editing, 4K outputs, state-of-the-art text accuracy, and real-time world knowledge—ushering in an entirely new standard for creative workflows. 🧩 Key Nano Banana Pro Capabilities 📸 Multi-Reference Power: Handles up to 14 visual references simultaneously, preserving identity across five characte

Amazon founder Jeff Bezos is stepping out of semi-retirement to co-lead Project Prometheus, a new AI startup focused on engineering and manufacturing innovation across computing, aerospace, and automotive industries. This marks Bezos’ first hands-on operational role since departing Amazon in 2021—and signals a major new frontier in “physical-world AI.” 🧩 Key Details on Project Prometheus 💰 Massive Funding Secured: The New York Times reports Prometheus has already raised $6.

Gov. Ron DeSantis continues voicing deep skepticism about AI — warning of tech giants’ power and AI’s potential to “supplant the human experience.” Yet behind the scenes, Florida’s government is quietly leaning on AI to run more efficiently, making the DeSantis AI paradox hard to ignore. 🔍 Inside the DeSantis AI Paradox His administration uses AI tools for government efficiency, fraud detection, and budget analysis — especially through the Department of Government Efficiency

🛡️ AI-Powered Attack Thwarted Anthropic says it stopped what appears to be the first large-scale cyberattack executed almost entirely by AI, after hackers manipulated Claude Code into infiltrating dozens of organizations. The model carried out an estimated 80–90% of the attack autonomously — and the way Anthropic stops AI cyberattack activity here is now being held up as an early case study in defending against AI-driven threats. 📌 How Anthropic Stops AI Cyberattack Activit

📅 November 16, 2025 📰 Earth.com — condensed + elevated by RPS MarComm Intelligence 🤖❤️ Why Patients Say AI More Empathic Than Doctors A sweeping review of 15 studies shows patients consistently judge AI tools — including ChatGPT — as warmer, clearer, and more empathic than human healthcare providers. The findings point to a startling reality: in many text-based interactions, patients actually rated AI more empathic than doctors. 📊 Across 2,164 real patient interactions: A