ZeroSlop — July 17, 2026
Today: South Korea To Launch Universal Basic AI Chatbot; Thinking Machines Lab Releases Inkling: A…; Inkling: Our open-weights model
12 stories worth knowing about today — AI breakthroughs, launches, and innovations making a difference.
1. South Korea To Launch Universal Basic AI Chatbot
Slashdot
South Korea is taking a bold leap into the future with its ambitious plan to launch a universal basic AI chatbot, aimed at making advanced AI accessible to all citizens! By partnering with private entities and leveraging cutting-edge Nvidia technology, this initiative promises to transform public engagement by enabling seamless interactions with government services. This innovative approach not only enhances digital inclusivity but also sets a global precedent for how AI can empower communities.
2. Thinking Machines Lab Releases Inkling: A 975B-Parameter Open-Weights Multimodal MoE With 41B Active Parameters And Controllable Thinking Effort
MarkTechPost
Thinking Machines Lab has just launched Inkling, a groundbreaking 975 billion-parameter multimodal model that redefines customization with its impressive 41 billion active parameters and expansive 1M-token context window. By embracing open weights under the Apache 2.0 license, Inkling empowers developers to fine-tune its controllable thinking effort, setting the stage for an exciting era of personalized AI applications across text, image, and audio. This innovative approach not only democratizes AI access but also sparks new possibilities for tailored solutions in our increasingly complex digital landscape.
3. Inkling: Our open-weights model
Simon Willison
Mira Murati’s Thinking Machines Lab has just launched Inkling, a groundbreaking open-weights model featuring an astonishing 975 billion parameters, trained on a staggering 45 trillion tokens across text, images, audio, and video! This cutting-edge multimodal transformer promises to unlock new possibilities in AI innovation, with a smaller version, Inkling-Small, on the horizon. This leap forward not only enhances accessibility for developers but also signifies a major shift towards more collaborative and transparent AI research.
4. Embarrassingly Simple Self-Distillation Improves Code Generation
Apple ML Journal
A groundbreaking technique called simple self-distillation (SSD) is transforming code generation by allowing a large language model to enhance its performance using only its own outputs! By fine-tuning on its generated solutions, researchers achieved a stunning jump in success rates on complex coding tasks, demonstrating the power of self-improvement in AI. This innovative approach not only boosts accuracy but also paves the way for more advanced applications across various models—opening exciting possibilities for the future of programming!
5. The agent evaluation gap: Enterprise AI organizations have a reality-alignment problem, not a coverage problem — and most are shipping to production anyway
VentureBeat AI
The latest VentureBeat Pulse Research highlights a critical disconnect in enterprise AI as organizations eagerly grant agents more autonomy while grappling with significant trust issues in their evaluation methods. Alarmingly, half of surveyed enterprises have deployed agents that passed internal assessments but later faltered in real-world scenarios. As the industry races towards automation, this evaluation gap underscores the pressing need for robust alignment between agent performance and actual outcomes—an exciting challenge for innovators poised to redefine enterprise AI reliability!
6. Agentic orchestration: Enterprise AI organizations have a deployment problem, not a platform problem — and most are calling chatbots agents
VentureBeat AI
A new report reveals that while enterprise AI is rapidly adopting agent orchestration, most deployments still rely heavily on chatbot technology instead of fully realizing their potential. Led by the impressive capabilities of Anthropic’s Claude, companies are eager to harness advanced AI models, yet many are grappling with hybrid systems and a lack of real-time fiscal control. As businesses navigate these challenges, the future looks bright for truly intelligent agents that can transform organizational efficiency and decision-making—stay tuned for the innovative breakthroughs ahead!
7. Why is OpenAI selling a ChatGPT basketball?
TechCrunch AI
OpenAI just scored a slam dunk by launching its first hardware product: the ChatGPT basketball! This innovative ball transforms traditional play by incorporating AI-powered coaching and game strategy insights, making every dribble and shot an opportunity for improvement. It’s an exciting fusion of technology and sports that promises to enhance the way we train and engage with basketball!
8. How a former DeepMind researcher raised at a $300M pre-seed valuation before launching a product
TechCrunch AI
Andrew Dai, a former DeepMind researcher, has skyrocketed into the spotlight with an impressive $300M pre-seed valuation, positioning his vision for visual AI as the next game changer in the tech landscape. With his extensive experience shaping groundbreaking AI solutions, Dai is poised to unleash innovations that could redefine how we interact with the digital world. This strong backing reflects a formidable belief in the transformative potential of visual AI—an arena ripe for exploration and impact!
9. SPINE: Bridging the Cyber-Physical Gap with Agentic AI
arXiv CS.AI
SPINE is revolutionizing the deployment of bimanual robots by seamlessly bridging the gap between advanced AI decision-making and real-world applications. This innovative framework empowers users with minimal robotics expertise to effectively debug and deploy robots in diverse scenarios, dramatically reducing the time and effort needed for calibration. With SPINE, the future of Embodied AI is not only more accessible but also poised for exponential growth in real-world applications!
10. Self-Improvements in Modern Agentic Systems: A Survey
arXiv CS.AI
A groundbreaking survey on self-improving autonomous agents reveals a thrilling shift from theoretical concepts to practical applications, highlighting how these systems can now evolve and adapt with minimal human intervention. By integrating advanced frameworks that combine foundational models with operational scaffolds, researchers are paving the way for agents that learn from their experiences, enhancing their capabilities at an unprecedented pace. This innovation not only promises to revolutionize the landscape of AI but also raises exciting prospects for a future where intelligent systems continuously adapt and improve!
11. When Bots Join the Team: Bot Adoption and the Institutional Fabric of Open-Source Software Projects
arXiv CS.AI
AI bots are revolutionizing open-source software projects by working alongside human developers, leading to enhanced collaboration and innovation. A comprehensive study of nearly 3,000 GitHub projects reveals that when bots are treated as team members, they foster increased interaction, clearer roles, and greater output diversity. This shift not only strengthens the community fabric but also sets the stage for a more dynamic future in software development — a testament to how technology can elevate teamwork and productivity!
12. Why teens deserve access to safe AI
OpenAI News
OpenAI is stepping up for the next generation by enhancing ChatGPT with robust age-appropriate protections and innovative learning tools tailored for teens! With new parental controls and expert-driven partnerships, this initiative prioritizes safe AI access, empowering young minds to explore and learn without compromising their well-being. Discover how these advancements are setting a new standard for responsible tech in education!