Salesforce Monitors AI Adoption While Nvidia Launches Vera Rubin

Artificial intelligence continues to reshape various sectors, from education to industry, with significant developments highlighted at CES 2026. The event re-energized the tech market, boosting investor confidence in AI, robotics, and autonomous driving. Experts believe AI adoption is still in its early stages, suggesting ongoing growth for "second derivative" companies, such as those producing storage chips. Investors are closely watching AI adoption and usage metrics from software companies like Salesforce and ServiceNow in upcoming earnings reports. Nvidia remains a central player, with CEO Jensen Huang announcing the Vera Rubin AI platform is now in full production. This powerful platform features 72 GPUs and 36 CPUs per server, significantly enhancing AI computing power. The intense demand for specialized AI chips is fueling an ongoing computer chip shortage, impacting industries like automotive, with Honda delaying a plant reopening. This demand also drives Bitcoin miners to pivot, offering data center services to AI companies for more stable income, though it raises costs for smaller miners. As AI systems become more autonomous, securing AI agents is a growing concern. The National Institute of Standards and Technology (NIST) and its Center for AI Standards and Innovation (CAISI) are actively seeking public input to develop security guidelines. These agents, while efficient, pose risks like hijacking and cyberattacks that could affect public safety and trust. In related developments, Reolink unveiled its AI Box and OMVI Series security cameras at CES 2026, with the AI Box processing data locally using Qualcomm Dragonwing Q8 Series for enhanced privacy and features like Prompt-Based Alerts. AI is profoundly transforming education, shifting learning from traditional classrooms to virtual environments. By August 2024, over 86 percent of college students utilized AI in their studies, with more than half using it weekly. Tools like Khanmigo offer personalized, on-demand learning. However, countries like Kazakhstan face a significant digital divide, where rural schools lack internet and devices, hindering equitable access to AI benefits. To address this, CSU's TILT is offering workshops for instructors on integrating generative AI, including topics like AI for accessibility and designing writing assignments with AI using platforms like Microsoft Teams. Advancements in AI model development continue at a rapid pace. Portola developed Tolan, a voice-first AI companion utilizing OpenAI's GPT-5.1 models, designed for personalized, ongoing conversations with low latency and consistent personalities. Tolan uses OpenAI text-embedding-3-large and Turbopuffer for fast memory recall. Meanwhile, researchers from Carnegie Mellon and UC Berkeley introduced "Absolute Zero," a system where AI models, including LLaMA 2, learn by generating their own training data through self-questioning, significantly improving coding and reasoning skills and addressing the challenge of finding new training data.

Key Takeaways

• Nvidia's Vera Rubin AI platform is now in full production, featuring 72 GPUs and 36 CPUs per server, enhancing AI computing power.
• The booming AI industry is causing a significant computer chip shortage, leading to production delays and higher prices for electronics.
• Bitcoin miners are shifting their business to offer data center services to AI companies, seeking more stable revenue streams.
• NIST and CAISI are actively seeking public input to develop security guidelines for AI agents, addressing risks like cyberattacks and public safety concerns.
• Reolink introduced its AI Box, powered by Qualcomm Dragonwing Q8 Series for local AI processing, and the 24 MP OMVI X16 PoE camera at CES 2026.
• AI is rapidly transforming education, with over 86% of college students using AI by August 2024, though a digital divide persists in regions like Kazakhstan.
• CSU is offering expanded workshops for instructors on integrating generative AI into classrooms, covering topics such as accessibility and assignment design using Microsoft Teams.
• Portola developed Tolan, a voice-first AI companion utilizing OpenAI's GPT-5.1 models, designed for personalized, low-latency conversations with memory recall.
• Researchers developed "Absolute Zero," a system enabling LLaMA 2 models to learn by generating their own training data through self-questioning, improving coding and reasoning.
• CES 2026 re-energized the tech market, boosting investor confidence in AI and highlighting growth for companies like Salesforce and ServiceNow.

AI Transforms Education from Classrooms to Virtual Learning

Artificial intelligence is changing education, moving it from traditional classrooms to virtual learning. By August 2024, over 86 percent of college students used AI in their studies, with more than half using it weekly. This shift means schools and governments must adapt quickly to new expectations. AI tools like Khanmigo offer personalized, on-demand learning, helping students at all levels. Governments can modernize education by using dynamic content, data analytics for personalized learning, and modular credential systems.

Kazakhstan Faces Digital Divide in AI Education Push

Artificial intelligence is changing education globally, and Kazakhstan is working to integrate AI into its schools. AI offers customized learning, adaptive systems, and tutors for subjects like language and math, also reducing teacher workloads. However, a major digital gap exists between urban and rural areas in Kazakhstan. Rural schools often lack reliable internet, devices, and teacher training, meaning many students cannot access AI benefits. To ensure fairness, Kazakhstan's national education strategy must focus on equity, investing in digital infrastructure for all schools and training teachers.

NIST Seeks Public Help to Secure AI Agents

The National Institute of Standards and Technology, or NIST, is asking the public for ideas to help secure AI agents. NIST's Cyber Security Division is seeking feedback on how to protect these agents from hackers. Many companies use AI agents without fully understanding their security flaws, which creates new risks for cyberattacks, especially in critical infrastructure. Unchecked risks could harm public safety and consumer trust in AI. The public has 60 days to provide examples and recommendations to help NIST create security guidelines and best practices for AI systems.

CAISI Asks for Public Ideas on AI Agent Security

The Center for AI Standards and Innovation, or CAISI, is asking the public for ideas on how to secure AI agents. CAISI, part of NIST, wants to develop guidance for these systems that can complete tasks on their own. While AI agents offer efficiency, they also carry risks like hijacking and cyberattacks, which could affect public safety and trust. CAISI has already done experiments and now seeks input on threats, best practices, and how to assess agent security. The center also wants to know how to monitor environments to reduce risks.

Reolink Unveils AI Box and New 24 MP Camera at CES 2026

Reolink introduced its new AI Box and OMVI Series security cameras at CES 2026 in Las Vegas. The Reolink AI Box is an independent hub powered by Qualcomm Dragonwing Q8 Series, processing AI data locally for faster and more private security. It adds ReoNeura AI features like Prompt-Based Alerts and Event Description to many Reolink devices. The flagship OMVI X16 PoE camera, a CES 2026 Innovation Awards Honoree, features a 24 MP triple-lens system with 16x optical zoom and Synchronized Smart Tracking. Reolink also showed new power-efficient battery cameras and a solar floodlight camera.

Nvidia Rubin Platform in Production Bitcoin Miners Pivot to AI

Nvidia CEO Jensen Huang announced at CES 2026 that the new Vera Rubin AI platform is now in full production. This powerful platform will feature 72 GPUs and 36 CPUs per server, greatly improving AI computing power and efficiency. The AI boom is causing Bitcoin miners to change their business, offering data center services like power and cooling to AI companies. This shift can provide more stable income than just mining Bitcoin. However, the high demand for data center space also raises costs for smaller miners.

Tolan Creates Voice AI Companion Using OpenAI GPT-5.1

Portola developed Tolan, a voice-first AI companion that uses OpenAI's GPT-5.1 models. Tolan offers a personalized, animated character designed for ongoing conversations, learning from interactions over time. The app focuses on low latency, accurate context, and consistent personalities. Tolan's system rebuilds its context for each turn, allowing it to adapt quickly to changing topics. It also has a memory system that stores facts, preferences, and emotional cues, using OpenAI text-embedding-3-large and Turbopuffer for fast recall.

CES 2026 Boosts AI Tech Market and Chip Stocks

Chris Versace from Schwab Network discussed how CES 2026 re-energized the tech market, especially for AI. He noted that many updates in AI, robotics, and autonomous driving at the event boosted investor confidence. Versace believes AI adoption is still in its early stages, suggesting that "second derivative" companies, like those making storage chips, will continue to see growth. Investors should watch for AI adoption and usage metrics in upcoming earnings reports from software companies like Salesforce and ServiceNow. The market's next move will depend on future earnings guidance.

AI Investment Fuels Ongoing Computer Chip Shortage

A significant shortage of computer chips continues, largely driven by massive investments in artificial intelligence. This shortage has already caused Honda to delay reopening a plant in China, and electronics makers warn of higher prices for items like TVs and smartphones. Experts like Renzo DeMeo from RCD Advisors say that specialized AI chips are in extremely high demand. Unlike the pandemic-era shortage caused by consumer demand, this current issue is directly linked to the booming AI industry. There is currently no clear end in sight for this chip scarcity.

AI Models Learn by Asking Themselves Questions

Researchers from Carnegie Mellon and UC Berkeley developed "Absolute Zero," a system where AI models learn by asking and answering their own questions. This method allows AI to generate its own training data, greatly improving the coding and reasoning skills of LLaMA 2 models. The approach is similar to how humans learn beyond just memorizing information. While the idea of AI "self-play" is not new, Absolute Zero is the first to show it works for large language models. This innovation could help AI systems become more capable as traditional training data becomes harder to find.

CSU Offers Expanded AI Workshops for Instructors

The Institute for Learning and Teaching, or TILT, at CSU announced its Spring Generative AI Workshop Series is returning with more topics. These workshops help instructors understand how to use artificial intelligence in their classes. TILT staff and experienced CSU instructors will lead interactive sessions with hands-on strategies. Instructors can attend in person at TILT Room 104 or online through Microsoft Teams. The series runs from January 27 to April 21, covering topics like AI for accessibility and designing writing assignments with AI.

Sources

• Opinion: From Lecture Halls to Virtual Classes, AI Is Rewriting the Rules
• How Artificial Intelligence Transforms Education in Kazakhstan
• NIST asks public for help securing AI agents
• NIST AI center looks for input on agentic AI security, best practices
• Reolink Reveals New AI Box and Flagship 24 MP Triple-Lens Security Camera at CES 2026
• Bitcoin miners chase AI demand as Nvidia says Rubin is already in production
• How Tolan builds voice-first AI with GPT-5.1
• VIDEO: CES Re-Energizes AI Trade, Puts 'Second Derivative' Tech Plays in Focus
• AI investment continues to drive chip shortage
• AI Models Are Starting to Learn by Asking Themselves Questions
• Spring generative AI workshop series returns with expanded offerings