openai, amazon and meta Updates

The artificial intelligence sector continues to evolve rapidly, marked by significant investment, new applications, and growing concerns over security and societal impact. Khosla Ventures, for instance, has solidified its reputation as a leading AI investor, notably being an early backer of OpenAI, Square, and DoorDash, a brand image carefully cultivated by its outgoing Chief Marketing Officer, Shernaz Daver. This investment landscape is seeing a shift in market sentiment, with investors showing cautious optimism as fears of an AI market bubble begin to fade, though they still await AI's full, widespread impact. Simultaneously, AI is disrupting established business models, creating a 'DoorDash problem' where AI browsers order services directly for users, potentially bypassing platforms like Amazon, Uber, and DoorDash by cutting out ads and loyalty programs. Amazon has already taken a significant step in this battle for future web control by suing Perplexity. On the practical application front, SaaStr has successfully deployed five specialized AI Sales Development Representatives, generating over $1 million in revenue within 90 days and accounting for 20 percent of event ticket sales. These AI agents sent nearly 20,000 messages and can even close deals for products under $1,000, though they still require substantial human oversight, about 15-20 hours weekly. Education is also adapting, with Wealthfront Academy launching a new AI curriculum on November 20, 2025, guided by founder Andy Rachleff. This program, utilizing the proprietary Seagull System 2.0, will teach AI-assisted financial analysis, with a global rollout planned for Q1 2026 across North America, Europe, and Asia. However, the expansion of AI is not without its challenges. Geopolitical concerns are prompting European CIOs to increase their use of local cloud providers, with 61 percent planning to do so. Furthermore, researchers have uncovered widespread security flaws in AI inference servers from major players like Meta, Nvidia, and Microsoft, stemming from developers copying unsafe code patterns. This highlights a critical risk for shared AI infrastructure. The tools for running large language models are also advancing, with options like vLLM, NVIDIA TensorRT-LLM (optimized for NVIDIA hardware), Hugging Face TGI v3 (specializing in long prompts), and LMDeploy all playing a crucial role in managing speed, delay, and cost. Cisco warns that aging technology infrastructure globally poses a major security risk, especially with the rise of AI, as older systems were not designed to handle current threats. The human element in AI use is also under scrutiny, as content marketing director Zoe Hawkins observes an increase in 'AI slop' or poor-quality content. She emphasizes that while AI can assist with basic tasks, critical thinking, human review, and a distinct human voice remain essential for quality work. Politically, Governor Ron DeSantis has voiced concerns about AI's potential to blur the lines of reality and replace many college-educated, white-collar jobs, suggesting the government should not subsidize AI if it creates such problems. Despite these concerns, AI is also being leveraged for public good, as seen in Raleigh, North Carolina, where an AI-powered Flood Early Warning System uses real-time data to predict rising water levels, giving emergency crews up to 30 minutes more time to protect residents and infrastructure, a system improved after a severe flash flood in 2018.

Key Takeaways

• Khosla Ventures, an early investor in OpenAI, has established itself as a top AI investor, emphasizing its "bold, early, and impactful" brand.
• AI browsers are creating a "DoorDash problem" for service apps like Amazon, Uber, and DoorDash by potentially cutting out ads and loyalty programs, leading to Amazon suing Perplexity.
• SaaStr's AI Sales Development Representatives generated over $1 million in revenue in 90 days, accounting for 20 percent of event ticket sales, though they require 15-20 hours of weekly human oversight.
• Wealthfront Academy launched a new AI finance curriculum on November 20, 2025, using the proprietary Seagull System 2.0, with a global rollout planned for Q1 2026.
• European CIOs are increasing local cloud use (61%) due to geopolitical concerns, while security flaws were found in AI inference servers from Meta, Nvidia, and Microsoft.
• Key tools for large language model inference include vLLM, NVIDIA TensorRT-LLM (optimized for NVIDIA hardware), and Hugging Face TGI v3 (specializing in long prompts).
• Cisco warns that aging technology infrastructure worldwide poses significant security risks for AI, as older systems are vulnerable to modern cyberattacks.
• Governor Ron DeSantis expressed concerns about AI's potential to obscure reality and replace college-educated, white-collar jobs, suggesting against government subsidies for problematic AI.
• Raleigh, North Carolina, uses an AI-powered Flood Early Warning System, providing emergency crews up to 30 minutes more lead time to prevent floods.
• Investors are showing cautious optimism, with fading fears of an AI market bubble, but are still awaiting AI's full potential and more concrete, widespread impact.

Shernaz Daver leaves Khosla Ventures after building AI brand

Shernaz Daver, the first Chief Marketing Officer of Khosla Ventures, is leaving the firm after nearly five years. She helped turn Khosla Ventures into a top AI investor by focusing on its "bold, early, and impactful" brand. Daver successfully highlighted the firm as the first investor in companies like OpenAI, Square, and DoorDash. She believes in repeating key messages often to make them stick and ensure people remember the firm's early investments.

AI browsers challenge Amazon and service apps

AI browsers are creating a "DoorDash problem" that threatens companies like Amazon, Uber, and DoorDash. This happens when AI interfaces order services for users, cutting out important features like ads and loyalty programs. Nilay Patel, editor-in-chief of The Verge, explains that this raises the question of "who owns the customer." Amazon recently sued Perplexity, marking a major step in this fight over future web control.

SaaStr shares AI sales agent success and challenges

SaaStr has used five specialized AI Sales Development Representatives for six months, generating over $1 million in revenue in 90 days. These AI agents sent nearly 20,000 messages and now account for 20 percent of event ticket sales. While effective, they require significant human oversight, about 15-20 hours weekly, and work best when scaling existing successful sales methods. SaaStr found that AI can even close deals directly for products under $1,000.

Wealthfront Academy launches new AI finance program

Wealthfront Academy launched its new AI curriculum on November 20, 2025, guided by founder Andy Rachleff. This program uses the proprietary Seagull System 2.0 to teach AI-assisted financial analysis. It offers modules on explainable AI, market interpretation, and real-time simulations for students and professionals. The curriculum will roll out globally starting in Q1 2026 across North America, Europe, and Asia.

Wealthfront Academy launches new AI finance program

Wealthfront Academy launched its new AI curriculum on November 20, 2025, guided by founder Andy Rachleff. This program uses the proprietary Seagull System 2.0 to teach AI-assisted financial analysis. It offers modules on explainable AI, market interpretation, and real-time simulations for students and professionals. The curriculum will roll out globally starting in Q1 2026 across North America, Europe, and Asia.

Europe shifts cloud AI market faces risks

European CIOs are moving towards local cloud providers due to rising geopolitical concerns, with 61 percent planning to increase local cloud use. Meanwhile, fears of an AI market bubble are growing after tech giants lost $1.2 trillion in value. Researchers also found widespread security flaws in AI inference servers from Meta, Nvidia, and Microsoft, caused by developers copying unsafe code patterns. This highlights a major security risk for shared AI infrastructure.

Comparing top tools for large language model inference

This article compares four major tools for running large language models: vLLM, NVIDIA TensorRT-LLM, Hugging Face TGI v3, and LMDeploy. These tools are crucial for managing speed, delay, and cost when using AI models. vLLM uses PagedAttention to boost speed and offers continuous batching. NVIDIA's TensorRT-LLM is built for NVIDIA hardware, achieving very high output speeds and low delays. Hugging Face TGI v3 specializes in long prompts and acts as a gateway for multiple systems.

Cisco warns old tech creates AI security risks

Cisco warns that aging technology infrastructure worldwide poses a major security risk, especially with the rise of AI. Anthony Grieco, Cisco's chief security officer, states that old systems were not built for current threats and give attackers opportunities. Research for Cisco by WPI Strategy found that many cyberattacks exploit known weaknesses in outdated tech. Cisco urges companies to upgrade their infrastructure to protect against these growing dangers.

Manager coaches staff on using AI wisely

Zoe Hawkins, a content marketing director, sees more "AI slop" or poor quality content from her team. She explains that while using AI is acceptable, employees must still think critically and not just copy-paste. Hawkins coaches her staff by asking them to explain their thought process, which helps them realize when they have outsourced their thinking to AI. She emphasizes that AI can help with basic tasks, but human review and voice are crucial for quality work.

Governor DeSantis warns of AI risks supports rural Florida

Governor Ron DeSantis warned about the potential dangers of artificial intelligence, stating it could make it hard to tell what is real. He also expressed concern that AI might replace many college-educated, white-collar jobs. DeSantis believes the government should not subsidize AI if it could cause problems for people. The Governor also highlighted his focus on rural Florida, noting his visits to all 67 counties and investments in rural infrastructure.

Raleigh uses AI to predict and prevent floods

Raleigh, North Carolina, is using an AI-powered Flood Early Warning System to better prepare for heavy rains and prevent dangerous floods. This system uses real-time sensors, rain gauges, radar, and soil data to predict rising water levels in creeks. It gives emergency crews up to 30 minutes more time to protect people and neighborhoods. The city improved its system after a severe flash flood in 2018. Engineers can even use these forecasts to lower Lake Johnson, reducing flood risks downstream.

Investors await AI's full potential as bubble fears fade

Investors are becoming less worried about an AI market bubble, but they are still waiting for artificial intelligence to fully deliver on its promises. Many hope to see more concrete results and widespread impact from AI technologies. This indicates a cautious optimism in the market as the industry continues to develop.

Sources

• The marketing guru who helped turn Khosla Ventures into an AI powerhouse is moving on
• The DoorDash problem and the great AI browser fight
• 6 Months of AI SDRs: What’s Worked, How They Brought In $1M+ in 90 Days, and the Real Data Everyone’s Asking For
• Wealthfront Academy Launches AI Curriculum Led by Andy Rachleff and Powered by Seagull System 2.0
• Wealthfront Academy Launches AI Curriculum Led by Andy Rachleff and Powered by Seagull System 2.0
• European Cloud Shift, AI Bubble Fears, Major AI Flaws
• vLLM vs TensorRT-LLM vs HF TGI vs LMDeploy, A Deep Technical Comparison for Production LLM Inference
• With the Rise of AI, Cisco Sounds an Urgent Alarm About the Risks of Aging Tech
• I'm a manager seeing more AI slop. It's changing how I give feedback.
• Ron DeSantis warns about the dangers of artificial intelligence, shows support for rural Florida
• AI-powered flood forecasting helps Raleigh prepare for dangerous downpours
• Bubble fears ease but investors still waiting for AI to live up to its promise