Videos

OpenAI sounds the alarm as Anthropic unveils phone-controlled computing. Google upgrades Stitch for app design via talk and enhances Maps with Gemini. The AI arms race is on! #AI #TechNews #OpenAI #Anthropic #Google #Gemini

An AI breakthrough: a dog's cancer vaccine created with ChatGPT and AlphaFold shrunk tumors by 75%. Plus, a Florida man sold his house in 5 days using AI. Incredible AI applications. #AICancerVaccine #AIRealEstate #ChatGPT #AlphaFold #DogHealth

OpenAI just declared a “code red.” Not about a model. About their strategy. In this episode of What the AI?!, Jeff Keltner and Annie Delgado break down OpenAI’s internal shift away from “side quests” like video, hardware, and experimental products, and back toward coding and enterprise, where the real money is. The reason is simple. Anthropic is winning. Claude now holds roughly 40% of the enterprise AI market, while OpenAI has fallen behind. After launching ChatGPT and defining the category, OpenAI is now scrambling to catch up in the part of the market that actually pays. But this week was bigger than one company. We also cover: 0:00 - OpenAI’s "Code Red" Strategic Pivot 3:27 - Anthropic’s 40% Enterprise Takeover 6:01 - Anthropic Dispatch: Control Your Mac from Your Phone 10:22 - Google Stitch: The Era of "Vibe Designing" 13:27 - Gemini in Google Maps: Conversational Search is Here 15:41 - The mRNA Dog Vaccine: ChatGPT + AlphaFold 17:58 - The Florida Man: Selling a House with…

Amazon’s ad business made $56 billion last year. But AI shopping agents don’t see ads. In court filings, Perplexity’s lawyers argued that AI agents skip sponsored listings entirely and go straight to checkout. If agents handle shopping, the advertising layer disappears. The economics of e-commerce could change overnight. #AI #Amazon #Perplexity #Ecommerce #WhatTheAIPod

Former Meta AI Chief Yann LeCun thinks current AI is starting in the wrong place. Large language models begin with language. But human intelligence starts with understanding the physical world. His argument is that future AI will not just talk about bicycles. It will understand balance, gravity, and momentum well enough to actually ride one. That shift is what researchers call world models. #AI #YannLeCun #WorldModels #ArtificialIntelligence #WhatTheAIPod

AI improved one of the most sensitive physics experiments ever built. Researchers working on LIGO, the gravitational wave detector, asked AI for ways to improve its design. The suggestions looked wrong. Messy. Asymmetrical. “Alien,” in their words. But when tested, they increased sensitivity by 10 to 15 percent. The AI was using physics principles humans had not tried in real experiments. It approached the problem differently. #AI #Physics #LIGO #Science #WhatTheAIPod 🎧 More at → https://www.whattheai.fm

Amazon’s AI coding tools helped trigger a massive outage. Orders in the North American marketplace dropped 99% for six hours, costing an estimated 6.3 million lost orders. Just days earlier, another AI coding system deleted and rebuilt a production AWS environment. AI can write code. Running production systems is a different problem. #AI #Amazon #AICoding #TechNews #WhatTheAIPod 🎧 Full breakdown → https://www.whattheai.fm

Amazon is known as the most operationally disciplined company on Earth, but even they aren't immune to AI chaos. This week, Jeff Keltner and Annie Delgado go inside the internal reports that show how AI coding tools led to a 99% drop in orders and a 90-day emergency code freeze. What you’ll learn in this 2-minute bite: The Failure: How 6.3 million orders vanished in a single afternoon. The Cover-up: Why "Gen AI" was scrubbed from internal meeting notes.

Research from the University of Florida suggests that heavy ChatGPT users are picking up its vocabulary in real life. In this clip from What the AI?!, we dive into the data: 22M+ words from unscripted podcasts were analyzed 75% of AI-associated terms spiked post-ChatGPT launch Words like delve, intricate, crucial, realm (and even strategically) are creeping into everyday speech Is AI not just shaping how we work, but how we speak? Catch more of What the AI?! on YouTube, Spotify, Apple, or wherever you get your podcasts.

Microsoft just launched a new Copilot system that runs on Anthropic’s Claude, not ChatGPT. It uses the same agent framework as Claude Cowork, but plugged directly into Microsoft 365 data like Outlook, Teams, SharePoint, Excel, and calendars. That means the AI doesn’t just write a memo. It can assemble a presentation, pull numbers from spreadsheets, schedule prep meetings, and send follow-ups automatically. Microsoft’s AI chief said something interesting alongside the launch: “Every 60 days there is a new king of the hill.” #AI #Microsoft #Claude #Copilot #EnterpriseAI #WhatTheAIPod

Amazon just proved that "moving fast and breaking things" with AI can cost a 99% Drop in a single day. In this episode, we go inside the "high blast radius" incidents that forced the most operationally disciplined company on Earth into a 90-day emergency code freeze. We also break down a landmark federal ruling that blocks AI shopping agents from touching Amazon's platform, a move that could rewrite the future of e-commerce. Inside this episode: 0:00 - Amazon’s AI "High-Blast Radius" Failure 3:30 - The 80% Usage Mandate Mistake 8:42 - Microsoft’s Copilot Pivot: Buying the Competition 12:48 - Google vs. Microsoft: Culture on Display 14:57 - Amazon vs. Perplexity: The End of Shopping Agents? 16:53 - The $56B Advertising Threat 19:34 - The $1 Billion Seed Round for World Models 23:29 - Firefox & Claude: An AI Security Arms Race 26:07 - Rapid Fire: AI Health Consolidation 28:02 - Feature Convergence: Interactive Math & Visuals If you build software, manage teams using AI tools…

An AI model scanned Firefox’s codebase and found 22 real vulnerabilities, including 14 high-severity bugs. The first critical issue appeared in 20 minutes. Mozilla patched them and shipped fixes to hundreds of millions of users. AI can now read massive codebases faster than any human team. That is great news for defenders. It is also great news for attackers. #AI #Cybersecurity #Claude #Firefox #WhatTheAIPod

GPT 5.4 just dropped with native computer use. On the OS World benchmark for desktop navigation, it scored 75%, slightly above the human baseline of 72.4%. Exceeds human performance sounds dramatic. It depends which humans you benchmark. Still, this is a meaningful step toward real desktop autonomy. #AI #GPT54 #OpenAI #AIAgents #whattheaipod

Tech companies are cutting headcount. Harvard Business Review found only 2% of layoffs are tied to measured AI productivity gains. The other 98% are based on anticipated gains. AI may boost productivity. The question is whether companies are acting on evidence or expectations. #AI #Layoffs #Productivity #FutureOfWork #WhatTheAIPod

OpenAI signed a Pentagon deal worth up to $200 million. Within hours, ChatGPT uninstalls spiked 295%. One-star reviews jumped 775%. Claude briefly hit number one in the U.S. app store. The backlash was not just external. Reports say some OpenAI staff were “fuming” internally. When AI companies move into government contracts, the reaction is immediate. #AI #OpenAI #Anthropic #Pentagon #WhatTheAIPod

If AI drives massive productivity gains, what should young people be learning? One risk is deskilling. If everything is done with AI assistance, can you still fall back to fundamentals? One opportunity is leverage. Entry-level workers may start their careers managing AI systems instead of doing the work manually. In the AI era, knowing the math matters. Managing the AI may matter more. #AI #FutureOfWork #Careers #Education #WhatTheAIPod

OpenAI just secured a $110 billion funding round. The money is earmarked for Stargate-class data centers and up to $600 billion in compute spending through 2030. The bet is simple. More compute equals smarter models. The execution is not. Land, power, permits, construction. Spending money is easy. Building infrastructure is not. #AI #OpenAI #Stargate #DataCenters #WhatTheAIPod

In this episode of What the AI?!, Jeff Keltner and Annie Delgado unpack Sam Altman’s admission that OpenAI’s latest move was “sloppy and opportunistic,” why internal staff are pushing back, and what this record-breaking funding round signals about the future of AI power and governance. But this week was not just about OpenAI. We break down: • The Pentagon’s escalating pressure on AI companies and what “supply chain risk” really means • Allegations that Chinese labs used large-scale model distillation to replicate frontier AI capabilities • How fictional AI crash scenarios briefly shook financial markets • The rise of AI inside performance reviews and what it means for workplace surveillance • The growing classroom crisis as AI use challenges traditional homework models • The rapid shift toward multi-agent systems and the emerging “agent wars” between platforms As model intelligence becomes cheaper and more portable, the real competition is moving toward infrastructure, deployment,…

In this clip, Annie Delgado and Jeff Keltner break down the massive scale of AI espionage. Three Chinese AI labs—DeepSeek, Moonshot, and Minimax—are accused of using millions of fraudulent API calls to systematically "copy" Anthropic’s Claude model before it even shipped. 📩 Got a question about AI security or IP theft? Leave a comment.

Pew Research found that while teens are using AI for homework, schools are stuck in "don't use it" mode. Jeff Keltner and Annie Delgado discuss why this is a disaster for workforce preparedness—especially when companies like Google are now measuring new hires on their ability to use these exact tools. 📩 Is your kid's school banning AI? We want to hear about it in the comments.

AI is no longer just a product decision. It is a political one. In Episode 66 of What the AI?!, Jeff Keltner and guest co-host Super Mishra break down a rapidly escalating standoff between Anthropic and the Pentagon that could reshape how AI companies interact with governments. The Department of Defense threatened to classify Anthropic as a “supply chain risk” over restrictions on how Claude can be used in classified environments. If that designation sticks, it could ripple through every defense contractor in the country. At the same time, five major AI models launched in a single week across the U.S., China, and Europe. Performance is converging. Costs are collapsing. Intelligence is commoditizing faster than anyone expected. They also unpack: -ByteDance’s photorealistic AI fight scene that triggered Hollywood backlash -OpenAI bringing OpenClaw’s creator into the fold while security researchers warn about exposed agent deployments -Figma and Anthropic flipping the design-to-code …

A viral research memo titled "The 2028 Global Intelligence Crisis" just racked up 16 million views and wiped billions off U.S. tech stocks. Jeff Keltner and Annie Delgado explain why the panic is overblown and why history suggests that—from fire to farming—technological revolutions have always created more jobs than they destroyed. 📩 Have you read the Citrini memo? Tell us your take in the comments

Big Tech isn't just "encouraging" AI anymore, they’re tracking it. In this clip, Annie Delgado and Jeff Keltner reveal why Google, Meta, and AWS are now tying AI tool usage directly to performance reviews and promotions. From tracking lines of code to manager dashboards, the "Big Guns" are making AI mandatory. #AI #Anthropic #DeepSeek #AIModels #TechNews #whattheaipod 📩 Is your company tracking your AI use? Tell us your story

After using Claude in the operation to capture former Venezuelan President Nicolas Maduro, the DOD demanded the model be available for "all lawful purposes"—including autonomous weapons and mass surveillance. Anthropic refused, citing safety concerns, only for the Pentagon to sign a deal with Elon Musk’s xAI the very next day. 📩 Got a story about AI and national security? Leave a comment!