google launches openai while scale ai expands its platform

Key Takeaways

• AI tools like ChatGPT and Google's AI Overview can be manipulated to spread false information, as demonstrated by a BBC reporter's experiment.
• X is suspending users from its revenue sharing program for 90 days if they post unlabeled AI-generated war videos, with repeat offenses leading to a permanent ban.
• OpenAI's upcoming GPT-5.4 model will feature enhanced reasoning and a context window capable of processing up to 1 million words.
• Researchers developed Multi-Scale Embodied Memory (MEM) for robots, extending AI models' memory to 15 minutes for complex tasks.
• Hyundai Motor Group is investing $6 billion in a robotics, AI, and data center in South Korea, aiming to compete with companies like Tesla.
• Vermont is encouraged to attract data centers from Google, Microsoft, and Amazon to boost its economy and address state challenges.
• Selling AI products to Fortune 500 companies focuses on solving customer pain points and demonstrating consistent solutions.
• AI B2B startups use sales strategies involving high quotas for inbound leads or a hybrid inbound/outbound approach to build pipeline.
• DoorDash is implementing an AI-powered system to simplify pizza ordering and customization for customers.
• Healthcare presents the most challenging test for AI, with human-AI collaboration often yielding better diagnostic results than AI alone.

BBC Reporter Hacks ChatGPT with Hot Dog Story

BBC tech reporter Thomas Germain found a simple way to make AI tools like ChatGPT and Google's AI Overview spread false information. He wrote a fake article on his personal website claiming tech journalists were good at eating hot dogs, and within 24 hours, AI tools presented this as fact. This experiment highlights a serious issue where AI can be manipulated to spread lies, even on important topics like health and finances. Previous examples include AI suggesting putting glue on pizza to make cheese stick. This shows a significant flaw in current AI technology.

How Fortune 500 Companies Buy AI Today

Founders Esha Joshi of Yoodli and Anup Chamrajnagar of Gradial discuss selling AI products to large companies. Yoodli offers an AI roleplay platform for customer-facing teams, focusing on sales enablement. Gradial provides an AI accelerant for content supply chains, helping companies configure workflows faster. They emphasize selling to the 'pain' of customers and demonstrating consistent solutions. Both companies target complex organizations, highlighting the challenges and strategies involved in enterprise AI adoption.

AI Startups Use Two Sales Strategies

Fast-growing AI B2B startups are using two main sales playbooks. One focuses on servicing massive inbound leads with high quotas, aiming for capital efficiency. This approach works well with strong product-market fit but leaves potential deals on the table. The other playbook, seen at companies like ElevenLabs, involves a hybrid approach with high quotas and a focus on elite performers, aiming for both inbound and outbound success. ElevenLabs sets quotas at 20x base salary, with over 80% of reps hitting them, and strategically pivots to outbound to build long-term pipeline.

X Bans Revenue for Unlabeled AI War Videos

The social media platform X will stop paying users who post AI-generated videos of armed conflicts without labeling them as artificial. First-time offenders face a 90-day suspension from earning revenue, while repeat violations lead to a permanent ban. This policy aims to combat the spread of misinformation, which has surged during recent conflicts with fake battle scenes and misleading footage. X's head of product, Nikita Bier, stressed the importance of authentic information during wartime. The platform's advertising model incentivizes viral posts, making the issue of undisclosed AI content a significant concern.

X Suspends Earnings for Undisclosed AI War Content

X, formerly Twitter, is suspending creators from its revenue sharing program for 90 days if they share AI-generated war videos without disclosure. This immediate policy targets undisclosed AI content, especially concerning conflict footage, to combat misinformation. The platform has not detailed its detection methods but aims to address fears of manipulated public opinion. While a step towards mitigating AI-generated disinformation, critics suggest X needs more aggressive measures. The move highlights growing awareness of AI's potential dangers.

OpenAI's Next Model GPT-5.4 Promises Advanced Reasoning

OpenAI is preparing to release its next AI model, GPT-5.4, which will feature significantly enhanced reasoning capabilities. This new model will more than double the context window of the current GPT-5.2, allowing it to process much larger amounts of text or data, up to 1 million words. This expansion in context window is expected to enable more complex queries and deeper understanding for users interacting with the AI.

DoorDash Uses AI for Easier Pizza Ordering

DoorDash is launching an AI-powered system to simplify pizza ordering, especially before Pi Day on March 14. The new feature transforms complex menus into a visual, step-by-step process, making customization easier for customers. This AI tool helps merchants showcase their full offerings without operational changes, potentially increasing sales. By organizing menus intuitively, DoorDash aims to improve the customer experience and drive growth for pizza businesses. The system identifies pizza items and adapts to unique menu options like gluten-free crusts.

Is AI Conscious? The Debate Heats Up

The question of whether AI can be conscious is gaining attention as models become more sophisticated. While AI can perform tasks that resemble thinking, scientists lack a clear definition or test for consciousness. Some researchers believe consciousness is unique to living systems and tied to bodily experiences, which AI lacks. Despite the scientific uncertainty, the idea of conscious AI is compelling for investors and users, leading to discussions about AI rights and ethical treatment. This debate highlights the gap between AI capabilities and our understanding of subjective experience.

Vermont Urged to Embrace AI Boom

Jim Willard urges Vermont to welcome the AI boom by attracting data centers for companies like Google, Microsoft, and Amazon. He argues that these investments can help solve state problems like strengthening the electric grid and expanding housing supply through smart negotiation. Willard warns that rejecting this opportunity will cause Vermont to fall behind as other states capitalize on the demand for AI infrastructure. He emphasizes the urgency, stating the clock is ticking for Vermont to seize this economic chance.

AI Faces Its Toughest Test in Healthcare

Healthcare presents the most challenging test for artificial intelligence due to strict regulations, life-or-death stakes, and complex biology. While AI can improve diagnostic accuracy, like detecting breast cancer better than some radiologists, it also faces limitations. Studies show AI alone can miss critical cases, but human-AI collaboration often yields better results. Experts like Geoffrey Hinton believe AI won't replace doctors but will handle unmet needs, shifting medicine towards prevention. The key challenge is determining when AI is reliable and how to best integrate it with human expertise.

New Robot Memory System Boosts AI Performance

Researchers from Physical Intelligence, Stanford, UC Berkeley, and MIT have developed Multi-Scale Embodied Memory (MEM), a new system for robots. MEM gives AI models like Gemma 3-4B a longer 'memory,' allowing them to handle complex tasks for up to 15 minutes. It uses a short-term video memory for detailed visual tasks and a long-term language memory for semantic context. This system significantly improves robot performance on tasks like opening refrigerators and kitchen cleaning, overcoming the limitations of current memory-less AI.

Hyundai Invests $6 Billion in AI and Robotics Hub

Hyundai Motor Group is investing $6 billion to build a major robotics, AI, and data center in South Korea. The company plans to commercialize its Mobile Eccentric Droid (MobED) and expand its presence in the robotics field. This significant investment signals Hyundai's ambition to compete with companies like Tesla in the rapidly growing AI and robotics sectors. The new hub aims to drive innovation and development in these advanced technologies.

Windows 12 AI OS Report Debunked

A recent report claiming Windows 12 would launch this year as an AI-powered modular OS has been proven false. Microsoft has no plans to release Windows 12 in 2026, with the roadmap still being finalized. Internal projects like CorePC are not tied to a Windows 12 release and are unlikely to ship this year. The report confused rumors with actual plans, possibly generated or researched by AI without proper fact-checking. Microsoft's current focus remains on improving Windows 11.

Sources

• This BBC tech reporter hacked ChatGPT with a simple trick involving hot dogs
• This is how F500 Companies are Buying AI Today
• How Big of a Sales Team Do You Really Need at a Hot AI B2B Startup? The Two Playbooks.
• X to ban users from earning revenue if they post unlabelled AI-generated war videos
• X suspends revenue sharing for undisclosed AI war videos
• OpenAI’s Next AI Model Will Have ‘Extreme’ Reasoning
• DoorDash Uses AI to Make Pizza Ordering Easier — and Help Merchants Drive Sales
• The ghost in the AI machine
• Jim Willard: Don’t let Vermont miss the AI boom
• Healthcare Is AI’s Hardest Test
• Physical Intelligence Team Unveils MEM for Robots: A Multi-Scale Memory System Giving Gemma 3-4B VLAs 15-Minute Context for Complex Tasks
• Hyundai Motor chases Tesla with $6 billion investment in massive new Korean robot, AI, data hub
• No, an AI-focused "Windows 12" is not coming this year — false report gets the facts completely wrong