microsoft, openai and google Updates

Developments in the AI space continue across multiple sectors. In government, Colorado's AI law is facing revisions and delays, with the start date pushed to May and a reshuffling of a key committee to ensure its passage. The bill's estimated cost is $4.4 million. Meanwhile, Microsoft's AI Economy Institute has named fellows from the University of New Mexico, granting them $80,000 to develop AI courses. In the tech world, one writer had a surprisingly positive experience with OpenAI's GPT-5, viewing it as a helpful assistant. Google is reportedly working on AI glasses, potentially manufactured by HTC in Taiwan, building on HTC's recent launch of its own AI glasses. Micron Technology is seeing record sales due to the demand for AI memory chips, and the company is collaborating with TSMC. Plastno has launched 'Plastnofy,' an AI-powered app for eco-friendly house cleaning. However, there are concerns about the potential downsides of AI, including the possibility of an "AI winter" due to unclear business benefits and high costs. Palantir's AI tools are also raising concerns about civil rights due to their use in surveillance and targeting. Furthermore, AI's flaws can mirror human thinking, leading to mistakes and biases, as highlighted by a fintech startup's experience with its AI loan approval model and an MIT study showing radiologists making less accurate diagnoses when using an AI tool.

Key Takeaways

• Colorado's AI law is delayed until May and faces committee changes, with an estimated cost of $4.4 million.
• Microsoft's AI Economy Institute is funding AI course development at the University of New Mexico with an $80,000 grant.
• One writer found OpenAI's GPT-5 to be a surprisingly helpful and advanced AI assistant.
• Google's upcoming AI glasses may be manufactured by HTC in Taiwan.
• Micron Technology is experiencing record sales due to AI-driven demand for memory chips and is partnering with TSMC.
• Plastno launched 'Plastnofy,' an AI-powered app for eco-friendly house cleaning.
• Concerns are rising about a potential "AI winter" due to unclear business benefits and high costs.
• Palantir's AI tools raise civil rights concerns due to their use in surveillance and targeting.
• AI systems can mirror human biases and decision-making flaws.
• An MIT study suggests AI should be used for collaboration with experts rather than full automation, as radiologists made less accurate diagnoses when using an AI tool.

AI Bill Advances in Senate Committee Amid Cost Concerns

The Senate Appropriations Committee passed Senate Bill 4, an artificial intelligence bill, with a 4-3 vote. The bill now moves to the full Senate, but its future is uncertain due to cost concerns. The bill's estimated cost is $4.4 million, mainly for the Office of Information Technology and the judicial department. An amendment to shift disclosure requirements increased the cost by $400,000. The committee approved an amendment to move the implementation date to May.

Colorado Senate Reshuffles Committee to Push Forward AI Law Changes

Senate Majority Leader Robert Rodriguez changed a key committee's members to help pass his bill adjusting Colorado's AI law. The change gives Democrats a 5-2 advantage. The goal is to shift the regulatory load onto AI developers instead of companies and schools. The bill was stuck because it lacked enough votes to move to the Senate floor. The new committee members are expected to vote in favor of the bill.

Colorado AI Law's Start Date Delayed Until May After Legislative Debate

Colorado's AI law will not start until at least May, after the next legislative session. Senate Majority Leader Robert Rodriguez amended his bill to delay the start date from February. This change follows intense talks on the AI law during a special session. Rodriguez said the delay will allow time to determine which government groups use AI and how. Another bill, HB25-1001, suggests delaying the start date to October 2026.

One Writer's Surprising Experience with OpenAI's GPT-5

One writer shares a positive experience with GPT-5, the latest AI model from OpenAI. Unlike others who find it underwhelming, the writer sees it as a leap forward, comparing it to setting up thumbprint recognition on an iPhone. GPT-5 feels like a helpful assistant, finding a camp for the writer's children and offering insights on various topics. While it has flaws, like inventing a music album, it feels like a step toward having an always-available AI companion.

A Better Way to Use AI Collaboration Over Automation

The article suggests that instead of fully automating tasks with AI, we should use AI to help people. An MIT study showed that radiologists using an AI tool made less accurate diagnoses. This is because the AI couldn't use other important information that doctors use. The article suggests building AI that works with experts, like doctors and teachers, instead of trying to replace them. Collaboration between humans and AI can lead to better results.

Plastno Launches AI-Powered App for Eco-Friendly House Cleaning

Plastno, known for its sustainable home products, introduced 'Plastnofy,' an AI-powered house cleaning app. The app helps users manage tasks, track supplies, and get eco-friendly cleaning tips. Plastno also expanded its product line with items like sponge towels and dissolvable refills. The company aims to make sustainable cleaning easier while reducing waste. This launch positions Plastno as a leader in smart, eco-friendly home solutions.

Google's AI Glasses May Be Made in Taiwan by HTC

Google is reportedly done with its first AI glasses and may have them made in Taiwan. HTC and Quanta are possible manufacturers. HTC recently launched its own AI glasses, the Vive Eagle, made entirely in Taiwan. HTC also transferred some AR and VR staff and patents to Google for $250 million. If HTC gets the Google contract, it would help its factory in Taoyuan.

Micron Sees Record Sales Thanks to AI Boom

Micron Technology is experiencing record sales due to the growing demand for artificial intelligence. The company makes memory chips needed for AI, cell phones, and other tech. Micron's stock price has increased by over 40% in 2025. The company is also preparing to start building a large chipmaking factory in Central New York. Micron has begun working with TSMC, the world’s most valuable company and a leader in AI.

Microsoft's AI Economy Institute Names UNM Faculty as Fellows

Stephanie Moore and Victor Law from the University of New Mexico were chosen as fellows of Microsoft's AI Economy Institute. This fellowship recognizes their work in AI research and innovation. They will receive $80,000 to support their project, which aims to create AI courses for all students and focus on AI in the workplace. The goal is to promote AI knowledge, collaboration, and ethical thinking. Microsoft's institute aims to connect industry, policymakers, and education to boost AI innovation.

Is an 'AI Winter' Coming Soon?

The article suggests that the AI boom may be coming to an end. It argues that the business benefits of AI are not clear, with many AI investments failing to yield returns. The author also points out that AI technology is easily copied, making it hard for companies to maintain a competitive edge. The high costs of data centers and AI training, combined with the social and economic costs of AI, could lead to an "AI winter."

Palantir's AI Tools Pose Hidden Dangers to Civil Rights

Palantir's AI tools are used for surveillance and targeting, posing a threat to human rights. These systems gather data to find patterns and deliver targets to operators. They are used in mass surveillance, forced migration, and urban warfare. The invisible nature of these tools makes it hard for the public to understand their impact. These systems can violate privacy and limit freedom of speech by tracking people's movements and connections.

AI's Flaws Mirror Human Thinking

AI systems can make mistakes because they learn from human data and decisions. A fintech startup found that its AI loan approval model reacted to wording, just like humans. AI can fall into stereotype traps, like favoring male candidates for coding jobs. AI systems can also be influenced by vivid examples, first numbers, and the way information is presented. Like humans, AI can be overconfident and resist updates.

Sources

• Senate committee gives green (well, maybe yellow) light to artificial intelligence bill
• Top Senate Democrat shuffles committee to overcome opposition to measure tweaking Colorado’s first-in-the-nation AI law
• Colorado’s AI law won’t take effect until at least May — after next legislative session — under latest special session bills
• Opinion | How ChatGPT Surprised Me
• A Better Way to Think About AI
• AI-Powered House Cleaning Apps
• Google’s AI glasses rumoured to be made in Taiwan, possibly by HTC
• AI boom means record sales for Micron (Weekly top stories)
• OILS faculty named fellows of Microsoft’s inaugural AI Economy Institute
• The next ‘AI winter’ is coming
• Palantir’s tools pose an invisible danger we are just beginning to comprehend
• Humans And AI Share The Same Flawed Ways Of Thinking