Gov. Ron DeSantis warned at the University of Alabama that the massive wave of investment flowing into AI companies and hyperscale data centers is forming a bubble that will likely crash. He said some AI firms may try to become systemically important to the economy so they can receive government help, and taxpayers could ultimately be asked to bail out struggling companies. He made clear he is not predicting AI technology itself will fail, only that many companies will not survive. He also urged local communities to have more say over large data center projects.
More than 900 million people use ChatGPT every week, and some have accidentally shared sensitive information with the chatbot. An Engadget article warns that people should never use ChatGPT for legal advice, therapy, medical decisions or handling confidential work information. ChatGPT conversations are not private and have been subpoenaed or exposed in some cases. The article recommends relying on real lawyers, doctors and therapists for important personal matters instead of AI chatbots.
The New York Times is setting clear standards for AI licensing partnerships. Adam Greenberg, VP of strategic partnerships, spoke at the Digiday Publishing Summit in Miami and said AI marketplaces remain underdeveloped and deals are difficult to structure. The Times focuses on three principles: fair value exchange, long-term partnership and control over content use. He praised the Times' AI licensing deal with Amazon for meeting those conditions, and said the partnership team has grown to 10 people.
On the consumer side, a new masterclass bundle offers six self-paced video courses on prompt engineering for $16. The courses cover tools including Perplexity, Gemini, DeepSeek, Copilot, GPT and Tabnine, teaching users how to write clearer requests so AI returns more accurate and useful answers. The training focuses on research, multimodal tasks, reasoning, productivity and coding workflows.
Alexandr Wang is using a humorous social media strategy called the tweet cannon to support Meta's AI promotion efforts. His online activity has become a key part of how Meta markets its AI work, blending a tongue-in-cheek tone with a serious promotional purpose. The strategy was highlighted by Bloomberg and reported by Business Insider.
In the AI infrastructure stock market, Super Micro Computer and Avnet are competing for investor attention. SMCI has posted 20 percent revenue growth but saw a 15 percent drop in net income recently. AVT trades at a lower price-to-earnings ratio of 20 compared to SMCI's 25. Analysts suggest AVT's broader AI exposure and improving operating leverage may make it a stronger pick.
Strong demand for DRAM and high-bandwidth memory is boosting prices and profits for companies like Micron, SanDisk and SK hynix. Allen Bond from Jensen Investment Management says AI demand is real but warns that memory businesses are highly cyclical. He cautions that current earnings may not reflect long-term normalized economics and that investors should consider whether today's strong pricing can last as supply catches up.
The US Patent and Trademark Office has launched more than 20 AI capabilities and plans to add more. Its large language model Scout has over 7,000 daily and weekly users, and an AI plugin has already summarized more than 100,000 patent applications. The office aims to be fully cloud-based using Amazon, Google and Microsoft by June 2027 while maintaining strict quality standards.
Researchers have developed a new framework to evaluate conversational AI using both offline and online tests. The method was tested across 27 experiments on a deployed multi-turn assistant with 489 paired offline-online contrasts. The composite approach achieved 81.1 percent F1 compared to 34.3 percent for a raw classifier score, helping teams prioritize AI improvements before allocating limited experiment traffic.
Key Takeaways
- <ul><li>DeSantis warns AI investment boom will crash and taxpayers may bail out failing companies</li><li>ChatGPT should not be used for legal, medical, therapeutic or confidential work advice, says Engadget</li><li>New York Times sets three AI licensing principles and has a deal with Amazon</li><li>$16 masterclass bundle teaches prompt engineering for Perplexity, Gemini, DeepSeek, Copilot, GPT and Tabnine</li><li>Alexandr Wang uses tweet cannon strategy to promote Meta's AI efforts on social media</li><li>Avnet trades at lower PE ratio of 20 vs SMCI's 25 despite SMCI's 20% revenue growth</li><li>Memory stocks like Micron, SanDisk and SK hynix benefit from AI demand but face cyclical risks</li><li>USPTO has 20+ AI tools, Scout LLM has 7,000+ users, and plans cloud migration via Amazon, Google and Microsoft by 2027</li><li>New AI evaluation framework achieves 81.1% F1 using offline-online testing across 27 experiments</li><li>Over 900 million people use ChatGPT weekly, raising privacy and misuse concerns</li></ul>
DeSantis warns of AI investment crash and possible taxpayer bailout
Gov. Ron DeSantis spoke at the University of Alabama and warned that the huge amount of money flowing into AI companies and hyperscale data centers is creating a bubble. He said a financial crash is likely and that some AI companies may become too big to fail. He warned that taxpayers could be asked to bail out struggling tech companies. He stressed he is not predicting AI technology will fail, only that some companies will not survive. He also said local communities should have a say over large data center projects.
DeSantis predicts AI financial crash and bailout risk for tech firms
Gov. Ron DeSantis said AI has strong potential but the current investment boom will likely crash. He warned that some AI companies may try to become systemically important to the economy so they can receive government help. He compared the situation to a bubble and said not all AI companies will last. He also raised concerns about hyperscale data centers and their impact on local communities. He argued that government should not accelerate AI through subsidies or bailouts.
Five important tasks you should never use ChatGPT for
An Engadget article warns that ChatGPT should not be used for legal advice, therapy, handling confidential work information, medical decisions or replacing licensed professionals. The article notes that more than 900 million people use ChatGPT weekly and some have shared sensitive information by mistake. It says ChatGPT chats are not private and have been subpoenaed or exposed. The article recommends using real lawyers, doctors or therapists for important personal matters instead of AI chatbots.
New York Times shares principles for AI licensing partnerships
Adam Greenberg, VP of strategic partnerships at The New York Times, spoke at the Digiday Publishing Summit in Miami about AI content licensing. He said AI marketplaces are still underdeveloped and deals are hard to structure. The Times focuses on three principles: fair value exchange, long term partnership and control over content use. He praised the Times AI licensing deal with Amazon for meeting those conditions. He said the Times partnership team has grown to 10 people.
AI assistant masterclass offers prompt engineering courses for $16
A new masterclass bundle includes six self-paced video courses to help users get better results from AI tools. The courses cover prompt engineering for tools like Perplexity, Gemini, DeepSeek, Copilot, GPT and Tabnine. They teach how to write clearer requests so AI returns accurate and useful answers. The bundle focuses on research, multimodal tasks, reasoning, productivity and coding workflows.
Alexandr Wang's social media strategy fuels Meta's AI promotion
Alexandr Wang uses a humorous social media approach called the tweet cannon to support Meta's AI promotion efforts. His online activity has become a key part of how Meta markets its AI work. The strategy was highlighted by Bloomberg and reported by Business Insider. Wang's posts play a tongue-in-cheek tone while serving a serious promotional purpose.
SMCI vs AVT: Which AI infrastructure stock deserves attention
Super Micro Computer and Avnet are two companies competing in the AI infrastructure space. SMCI has seen 20 percent revenue growth but faced a 15 percent drop in net income recently. AVT trades at a lower price-to-earnings ratio of 20 compared to SMCI's 25. Analysts suggest AVT's broader AI exposure and improving operating leverage may make it a stronger pick for investors.
AI demand boosts memory stocks but earnings may mislead investors
Strong demand for DRAM and high-bandwidth memory is driving up prices and profits for companies like Micron, SanDisk and SK hynix. Allen Bond from Jensen Investment Management says AI demand is real but warns that memory businesses are highly cyclical. He cautions that current earnings may not reflect long-term normalized economics. Investors should consider whether today's strong pricing and margins can last as supply catches up.
USPTO expands AI tools with over 20 capabilities in development
The US Patent and Trademark Office has launched more than 20 AI capabilities and plans to add more. Its large language model Scout has over 7,000 daily and weekly users. An AI plugin has already summarized more than 100,000 patent applications. The office aims to be fully cloud-based using Amazon, Google and Microsoft by June 2027 while maintaining strict quality standards for AI tools.
New framework evaluates conversational AI using offline and online tests
Researchers developed a framework to test conversational AI using offline signals that match online experiment results. The method was tested across 27 experiments on a deployed multi-turn assistant with 489 paired offline-online contrasts. The composite approach achieved 81.1 percent F1 compared to 34.3 percent for a raw classifier score. The framework helps prioritize AI improvements before allocating limited experiment traffic.
Sources
- DeSantis predicts AI investment 'crash,' warns taxpayers could be asked to bail out tech
- DeSantis predicts AI investment 'crash,' warns taxpayers could be asked to bail out tech
- Never use ChatGPT for these five tasks
- What The New York Times looks for in AI licensing partnerships
- Avoid common mistakes with popular AI tools by studying this new masterclass, just $16
- Alexandr Wang's is Meta's not-so-secret weapon in the AI promo war
- SMCI vs. AVT: Which AI Infrastructure Stock is a Better Buy?
- AI Is Supercharging Memory Stocks. Are Earnings Lying?
- USPTO has over 20 AI capabilities with more coming, CIO says
- From Offline Proxies to Online Decisions: A Layered Engagement Evaluation Framework for Conversational AI
Comments
Please log in to post a comment.