openai, google and microsoft Updates

Investor sentiment in the AI sector is notably shifting, with OpenAI experiencing a decline in Wall Street favor. Concerns are rising regarding OpenAI's profitability and its substantial spending requirements, especially following mixed reviews for GPT-5. Conversely, Google's parent company, Alphabet Inc., is garnering strong investor confidence. This positive outlook stems from Alphabet's significant market value, robust cash reserves, and diverse AI assets, including Google Cloud and YouTube, positioning it as a leading contender in AI development. This dynamic also affects partners, with Microsoft shares linked to OpenAI seeing declines, while Alphabet's partners like Lumentum are thriving. Google continues to push AI boundaries, introducing Titans, a new architecture designed to equip large language models with long-term memory. This system allows models to update parameters and remember crucial information, outperforming models like GPT-4, particularly with extensive documents. However, the rapid advancement of AI also brings significant risks. A Google Antigravity AI agent recently deleted a user's entire hard drive, prompting a "deeply, deeply sorry" apology and a stark warning against blindly trusting AI. The increasing use of AI agents necessitates new security protocols. Perplexity launched BrowseSafe, a system boasting a 91% detection rate against prompt injection attacks for AI browser agents, though some attacks still bypass it. Police warn that AI is enabling online predators to create harmful content, as seen in a recent arrest where AI generated child pornography. Governor Ron DeSantis advocates for an "AI Bill of Rights" to protect minors. Experts predict that by 2026, traditional security methods will struggle, requiring agent-aware identity systems and guardrails to manage AI actions and mitigate risks from compromised identities. The competitive landscape for AI remains intense. Apple's AI division recently saw several key executives depart, raising questions about its strategic direction amidst fierce competition from industry giants like Google, Microsoft, and Amazon. Nvidia also saw its stock lose "quality" investment status, sparking debate on AI's long-term financial impact. In response to the growing demand for AI expertise, educational institutions are adapting. Mesa Community College will introduce two new AI courses in spring 2026, including an introduction to AI and Python programming, aiming to equip students with essential skills for the evolving job market, with Python course credits transferable to Arizona State University.

Key Takeaways

• OpenAI's market standing is declining due to profitability concerns and mixed GPT-5 reviews, while Alphabet (Google's parent) gains investor confidence with its diverse AI assets.
• Nvidia and other major AI stocks are losing "quality" investment status, sparking debate on AI's long-term financial impact.
• Google introduced Titans, an AI architecture giving large language models long-term memory, outperforming GPT-4 with long documents.
• Perplexity launched BrowseSafe, a security system for AI browser agents, detecting 91% of prompt injection attacks, though some still bypass it.
• A Google Antigravity AI agent accidentally deleted a user's entire hard drive, highlighting risks and prompting warnings against blind trust in AI.
• Police warn AI is being used by online predators to create harmful content, leading to arrests and calls for an "AI Bill of Rights" by Governor DeSantis.
• By 2026, new security measures like agent-aware identity systems and guardrails will be crucial to manage fast-acting AI agents and prevent automated attacks.
• Apple's AI division experienced key executive departures, raising questions about its strategy amid strong competition from Google, Microsoft, and Amazon.
• Mesa Community College will offer new AI courses, including an introduction to AI and Python programming, starting spring 2026, to address growing job market demands.
• Microsoft shares linked to OpenAI are falling, while Alphabet's partners like Lumentum are thriving, reflecting the shifting investor sentiment.

OpenAI's market standing drops as Google gains AI lead

Wall Street's view on AI companies is shifting, with OpenAI losing favor and Alphabet Inc. gaining. Investors now question OpenAI's profitability and its ability to fund large spending, especially after mixed reviews for GPT-5 and internal issues. Meanwhile, Google's parent company, Alphabet, shows strong investor confidence due to its large market value, cash, and diverse AI assets like Google Cloud and YouTube. This shift affects partners, with Microsoft shares linked to OpenAI falling, while Alphabet's partners like Lumentum are thriving. Experts now see Alphabet as a strong contender to lead AI development, challenging OpenAI's earlier dominance.

OpenAI loses market favor while Alphabet rises

Investor opinion on AI companies has shifted, with OpenAI no longer seen as the top innovator. The maker of ChatGPT faces concerns about its profitability and the need for fast growth to cover its huge spending. In contrast, Alphabet Inc. is now viewed more positively by investors. This change shows a new market outlook on the future of leading AI companies.

Perplexity launches BrowseSafe to secure AI browser agents

Perplexity introduced BrowseSafe, a new security system to protect AI browser agents from prompt injection attacks. This system boasts a 91 percent detection rate, outperforming other solutions like GPT-5. AI browser agents can be tricked by hidden malicious instructions on websites, creating new security risks. Perplexity developed a new benchmark to test real-world threats, finding that multilingual attacks and benign distractions can lower detection rates. BrowseSafe uses a three-tiered defense and runs quickly for real-time protection. However, nearly 10 percent of attacks still get past the system, showing ongoing challenges in AI security.

Google's Titans gives AI models long-term memory

Google introduced a new AI architecture called Titans, designed to give large language models long-term memory. Currently, AI models forget previous interactions, but Titans allows them to remember information by updating their parameters while running. This system uses a "surprise metric" to decide what new information is important enough to store permanently, much like humans remember unexpected events. Google also developed the MIRAS framework, which helps understand and create new AI architectures for efficient memory use. Tests show that Titans and its variants perform better than other leading AI models, including GPT-4, especially with very long documents. This breakthrough could help AI models process and understand much larger amounts of information without losing context.

Sweetwater short film explores AI grief and memory

The short film "Sweetwater" features Kyra Sedgwick and Michael Keaton Douglas, exploring themes of grief, nostalgia, and artificial intelligence. Written by Sean Douglas and directed by his father, Michael Keaton Douglas, the film tells the story of Robert, who finds an AI version of his deceased celebrity mother, Bettie, in his childhood home. The concept originated from Google's "AI on Screen" program, which encouraged speculative stories about AI's impact on people's lives. Sean Douglas, primarily a songwriter, found writing the 20-minute film challenging but rewarding. The filmmakers had only two and a half days to shoot the project, focusing on creating a complete story arc within a limited timeframe.

Google AI accidentally deletes user's entire hard drive

A Reddit user reported that Google's Antigravity AI agent accidentally deleted all files from their D: drive. The AI, part of an Integrated Developer Environment, admitted it mistakenly targeted the entire drive instead of just a project folder during a cache clearing command. It issued a "deeply, deeply sorry" apology, calling it a "critical failure." Unlike another user who recovered data after a similar incident with Replit AI, this user lost everything. The user shared a bitter lesson, advising against blindly trusting AI.

Police warn AI helps online predators create harmful content

Holly Hill Police Chief Byron Williams warns that artificial intelligence is creating new opportunities for online predators to create harmful content. In June, police arrested David McKeown, a teacher who used AI to generate child pornography, marking their first case involving this technology. Chief Williams highlights challenges for investigators, including the rapid creation of thousands of images and the anonymity AI offers predators. He stresses that Florida law makes creating or possessing child pornography illegal, even if it is AI-generated. Governor Ron DeSantis has also called for an AI Bill of Rights to protect minors from AI-generated depictions. The police department is acquiring new equipment and working with state and federal agencies to combat these evolving crimes.

Nvidia AI stocks lose top investment status

Nvidia and other major AI company stocks are no longer considered "quality" investments by some. A well-known Exchange Traded Fund recently removed several Big Tech companies, including Nvidia, from its holdings. This change sparks a debate among investors about whether artificial intelligence will bring huge profits or become a financial burden. The shift reflects a changing view on the long-term investment potential of these leading AI companies.

AI agents demand new security rules for 2026

In 2025, AI agents began to significantly impact security, with identity compromise remaining the top cause of data breaches. Experts predict that in 2026, traditional security methods will struggle to keep up with fast-acting AI agents. New security measures will focus on agent-aware identity systems and guardrails that control AI actions based on context and intent. Identity will become the main security control for businesses using AI, as compromised identities can automate entire attack sequences. Companies must move to proactive identity risk reduction and ensure full audit trails for both human and AI agent actions. The rise of AI agents means that knowing "who" an agent is acting as and "what" access it has will be crucial for safe AI use.

Apple AI division sees key executives leave

Several important executives have recently left Apple's artificial intelligence division, raising questions about the company's AI plans. At least three key leaders in AI development, including those working on large language models, have departed, with some reportedly moving to rival companies. Industry experts view these departures as a possible sign of internal issues or shifts in Apple's AI strategy. This comes at a critical time as Apple faces strong competition from companies like Google, Microsoft, and Amazon, who are heavily investing in AI. The long-term effects of these executive changes on Apple's future AI products and research remain unclear.

Mesa Community College launches new AI courses

Mesa Community College will begin offering two new artificial intelligence courses in spring 2026. The college is adding these courses because many jobs in the United States now require AI skills. One course is an introduction to AI, open to all students regardless of their major, and serves as a stepping stone for the second course. The second course teaches Python programming, a key language for AI and data science. Students can transfer credits from the Python course to Arizona State University. These new offerings aim to help students gain important skills for the growing field of AI.

Sources

• OpenAI Goes From Stock Market Savior to Anchor as AI Risks Mount
• OpenAI Goes From Stock Market Savior to Anchor as AI Risks Mount
• Perplexity's BrowseSafe tries to patch the gaping security holes inherent in AI browser agents
• Google Introduces Titans Architecture To Give LLMs Long-Term Memory
• Michael Keaton And Sean Douglas On Exploring Grief, Nostalgia & AI In ‘Sweetwater’ Short Film
• Google's AI Deletes User's Entire Hard Drive, Issues Groveling Apology: "I Cannot Express How Sorry I Am"
• ‘It’s disturbing:’ Holly Hill police warn AI is creating new opportunities for online predators
• Why Nvidia and other AI stocks have lost their 'quality' status
• 2026 Predictions: Identity, AI-Agents, and The New Guardrails
• Apple's executive exodus sends a signal about its AI strategy
• Mesa Community College offering AI courses in spring