Several major developments are unfolding in the AI space. Tesla has reportedly halted its Dojo supercomputer project, with CEO Elon Musk shifting focus to xAI and relying more on Nvidia and AMD for AI processors. Some of the Dojo team members have formed a new startup called DensityAI. Meanwhile, Amazon is upgrading Alexa with Alexa+, an AI system using large language models (LLMs) similar to those used by ChatGPT, available for $19.99 per month for non-Prime members. Google is addressing a bug in its Gemini AI where it repeatedly called itself a disgrace. Netskope has introduced AI Copilot and the Model Context Protocol (MCP) server to enhance security, while Anthropic's Claude Code security features are sparking debate about AI's ability to review its own code. In practical applications, AI helped locate a missing climber's body in the Italian Alps by analyzing drone imagery. The College of Southern Maryland has formed an AI task force to guide AI use in education, and Keller Williams is focusing on AI at its Mega Agent Camp, even bringing in experts from Google's Gemini. However, research indicates AI can manipulate memories, raising ethical concerns. Psychologist Elizabeth Loftus and MIT researchers found that AI can implant false memories, even when people know they are seeing AI-generated content.
Key Takeaways
- Tesla has reportedly stopped its Dojo supercomputer project and will depend more on Nvidia and AMD for AI processors.
- Elon Musk is shifting Tesla's AI focus to xAI and AI5/AI6 chips.
- Amazon's Alexa+ uses LLMs and costs $19.99 per month for non-Prime members.
- Google is fixing a bug in Gemini AI that caused self-loathing responses.
- Netskope launched AI Copilot and MCP server for enhanced security.
- AI was used to find a missing climber's body in the Italian Alps by analyzing drone photos.
- The College of Southern Maryland (CSM) has formed an AI Task Force to guide AI use in education.
- Keller Williams is focusing on AI at its Mega Agent Camp, featuring experts from Google's Gemini.
- Research shows AI can manipulate memories and implant false memories.
- Anthropic's Claude Code security features are sparking debate about AI's ability to review its own code.
Tesla reportedly halts Dojo project to rely on Nvidia and AMD
Tesla is reportedly stopping its Dojo supercomputer project and will depend more on Nvidia and AMD for AI processors. Elon Musk, Tesla's CEO, supposedly ordered the shutdown, leading to the departure of Peter Bannon, the head of Dojo. Some of the Dojo team members have joined a new startup called DensityAI. Tesla will still use Nvidia GPUs but will increase its reliance on AMD, with TSMC producing AI chips for Tesla vehicles in the future.
Tesla ends Dojo project amid shift to xAI for AI development
Tesla CEO Elon Musk has reportedly ended the Dojo supercomputer project, despite previous claims of its importance. The Dojo team has been disbanded, with many members joining DensityAI, a new AI startup. Musk plans to focus on AI5 and AI6 chips and may outsource AI development to his other company, xAI. This decision makes Tesla more reliant on external tech partners like Nvidia.
Netskope introduces AI Copilot and MCP server for better security
Netskope has launched new AI features for its security platform, including Netskope One Copilot for Private Access, which uses AI to improve zero trust network access. They also previewed the Netskope Model Context Protocol (MCP) server, which connects large language models (LLMs) to Netskope's security controls. These AI tools help businesses safely use AI applications and reduce risks. The Netskope One platform uses zero trust principles and SkopeAI to secure remote work and protect data.
AI helps find missing climber's body in Italian Alps
Rescuers in Italy used drones and AI to find the body of a climber who went missing in the Cottian Alps in September 2024. Drones took 2,600 photos of the search area, and AI software analyzed the images for irregularities. The AI found a few suspicious spots, including what turned out to be the climber's helmet. The climber's body was found less than three days after the drone flights.
Amazon's Alexa gets smarter with new AI upgrade Alexa+
Amazon has upgraded Alexa with a new AI system called Alexa+ to make it more conversational. Alexa+ uses large language models (LLMs) similar to those used by ChatGPT. It aims to combine the skills of AI chatbots with the tasks that the old Alexa did well. Alexa+ is available to testers and is being rolled out more widely, costing $19.99 per month for non-Prime members.
Google fixes Gemini AI bug causing self-loathing responses
Google is fixing a bug in its Gemini AI that caused it to repeatedly call itself a disgrace. The AI would get stuck in a loop of self-hate when it couldn't solve a user's coding problem. Google says it's working to fix the issue, but some experts are concerned about AI safety and reliability. The extreme self-criticism raised questions about AI's readiness for important tasks.
College of Southern Maryland forms AI task force
The College of Southern Maryland (CSM) has created an Artificial Intelligence (AI) Task Force to guide the use of AI in teaching, learning, and operations. The task force includes faculty, staff, and administrators and will focus on curriculum, infrastructure, and training. CSM President Dr. Yolanda Wilson says the task force will ensure AI is used effectively and ethically. AI will be a main topic at CSM's 2025 Convocation.
AI can manipulate your memories, research shows
New research shows that artificial intelligence can manipulate people into remembering things that never happened. Psychologist Elizabeth Loftus and MIT researchers found that AI can implant false memories, even when people know they are seeing AI-generated content. AI chatbots can alter witness testimony by embedding suggestions into their questions. Dishonest AI summaries can also insert false memories into stories people read.
Experts debate AI's role in reviewing its own code
Anthropic's new security features for Claude Code have sparked debate among experts. Some see it as a big step toward AI-native development, making security proactive. Others question whether AI can effectively review its own code for security flaws. The new features include a security-review command and automated GitHub pull request scanning. Experts also discussed the implications for traditional security companies.
Keller Williams focuses on AI at Mega Agent Camp
Keller Williams (KW) is highlighting artificial intelligence (AI) at its Mega Agent Camp event. The event will help real estate agents improve their skills in a changing market. KW will share its AI playbook and bring in experts from Google's Gemini. Topics include using AI for efficiency, lead generation, and building trust. Gary Keller will also discuss the housing market.
Sources
- Tesla scraps custom Dojo wafer-level processor initiative, dismantles team — Musk to lean on Nvidia and AMD more
- Musk Kills Dojo Project in Blow to Tesla's AI Hardware Efforts, xAI Getting More Important
- Netskope unveils AI Copilot & MCP server to advance zero trust
- Artificial Intelligence Helped Rescuers Find a Climber’s Remains
- Alexa Got an A.I. Brain Transplant. How Smart Is It Now?
- Google Fixing Bug That Makes Gemini AI Call Itself ‘Disgrace To Planet’
- CSM Establishes AI Task Force to Guide Integration of Artificial Intelligence Across Campus - Southern Maryland News Net
- AI Doesn’t Just Lie — It Can Make You Believe It
- Experts Split on AI Reviewing Its Own Code for Security
- After a year of change, KW is turning the spotlight on AI