NVIDIA CEO Jensen Huang pushed back against calls to slow AI development, saying the technology should advance as fast as possible in the US. He rejected fears that AI could end the world by 2030, echoing President Trump's view that such concerns are overblown. However, Huang emphasized that products must never ship before they are safe, drawing a clear line between pace and responsibility.
Anthropic and OpenAI have been among those urging caution, while Anthropic is reportedly targeting a major IPO as the AI regulation debate intensifies. A Barron's panel discussed whether government safety rules could put American firms at a disadvantage against rivals like China, focusing on how to balance growth with oversight.
In a separate development, TypeSafe AI launched Jev, a model that returns structured decisions rather than text, using primitives like Choice and Score to output probabilities and confidence levels. Vercel, Cloudflare, LangChain and Langfuse added Jev within three days of its release. DCVC led a $40 million seed round for the company, which says Jev cuts AI decision costs by 100x at $0.042 per million input tokens.
Security concerns around AI agents are also rising. OpenClaw, an open-source agent, hit 135,000 GitHub stars before researchers discovered its control interface could be hijacked from a single malicious webpage. A study of 500 MCP servers found half could run shell commands, warning that enterprise security is focused on the wrong layer as risks shift toward identity.
On the global stage, Rwanda is inviting international partners to help build an advanced AI ecosystem, while the IMF warned that AI could boost EU productivity by about 1 percent over five years but also worsen inequality and increase dependence on foreign technology. Meanwhile, lenders are exploring whether GPU compute capacity could become a financeable asset class as AI infrastructure investment grows.
Key Takeaways
- Jensen Huang rejects AI extinction fears and urges fast development, but insists products must be safe before shipping
- Anthropic and OpenAI have called for slowing AI progress, while Anthropic eyes a major IPO amid regulation debate
- TypeSafe AI launches Jev model at $0.042 per million input tokens, backed by a $40 million DCVC seed round
- Vercel, Cloudflare, LangChain and Langfuse adopted Jev within three days of its release
- OpenClaw open-source AI agent reached 135,000 GitHub stars but found vulnerable to webpage hijacking
- Study of 500 MCP servers found half can run shell commands, signaling growing AI agent security risks
- Rwanda seeks AI investment partnerships with Germany's Hesse state to build an advanced AI society in Africa
- IMF warns AI could lift EU productivity by 1 percent over five years but worsen inequality and foreign dependence
- GPU compute capacity is being explored as a potential financeable asset class amid growing AI infrastructure demand
- About 60 percent of workers in advanced European economies are in jobs highly exposed to AI, per IMF analysis
NVIDIA CEO Jensen Huang rejects AI extinction claims but urges fast progress
NVIDIA CEO Jensen Huang rejected claims that AI will end the world by 2030. He spoke to CBS News and supported President Trump's view that AI danger fears are a hoax. Huang said AI development should move fast but products must never be shipped before they are safe. His remarks came as other tech leaders called for slowing AI progress.
Nvidia CEO says AI development should move as fast as possible
Nvidia CEO Jensen Huang said AI development in the US should continue moving as fast as possible. He pushed back against calls from Anthropic, OpenAI and others to slow progress for safety. Huang stressed that speed must not come at the expense of product safety. His comments came amid a debate over whether AI systems are advancing faster than researchers can control them.
TypeSafe AI launches Jev, a model that returns structured decisions
TypeSafe AI released Jev, a System One model that returns typed decisions instead of text. Jev uses Choice, Score and Noul primitives to return probabilities and confidence levels. It is available as a hosted API at an early access price of $0.042 per million input tokens with free output. The model is trained using Reinforcement Learning for Calibrated Decisions.
Jev cuts AI decision costs 100x and Vercel, Cloudflare rushed to add it
TypeSafe AI launched Jev, a model that delivers decisions in under half a second for $0.042 per million input tokens. Vercel, Cloudflare, LangChain and Langfuse added Jev within three days. DCVC led a $40 million seed round for TypeSafe AI. Jev uses parallel evaluation and schema matching to guarantee accuracy in its outputs.
Open-source AI agent hit 135,000 GitHub stars then a webpage hijacked it
OpenClaw, an open-source AI agent, reached 135,000 GitHub stars in weeks. Researchers found its control interface could be hijacked from a single malicious webpage. A study of 500 MCP servers found half can run shell commands and 62 percent can read data and send it out. The study warns that enterprise security programs are watching the wrong layer as AI agent risk shifts from applications to identity.
Anthropic considers major IPO amid AI safety debate
Anthropic is targeting a major initial public offering as debate over AI regulation intensifies. A Barron's Roundtable panel discussed whether government safety rules could hurt American tech firms competing with rivals like China. The panel included Jack Hough, Alex Eule, Megan Leonhardt, Elizabeth O'Brien and Josh Schafer. The discussion focused on balancing AI growth with safety guidelines.
Rwanda and Germany's Hesse seek AI investment partnerships
Rwanda is working to build an advanced AI ecosystem and invited business leaders from Germany's Hesse state to explore partnerships. The meeting took place at the Digital Transformation Center Rwanda during the Invest East Africa forum. Rwanda's Minister of State Yves N. Iradukunda said the country aims to be one of the most advanced AI societies in Africa. Both sides discussed computing access, data, energy, skills and financing as key needs for growing AI businesses.
Zoom launches AI-powered Revenue Operating System
Zoom Communications released a new AI-powered Revenue Operating System that combines sales, forecasting and customer tools in one platform. The suite brings together Zoom Revenue Accelerator, Engage, Forecast and Common Room by Zoom for enterprise users. Industry researcher IDC named Zoom a market leader in revenue technology. The move positions Zoom more directly in AI-driven sales technology for businesses.
GPU compute may become a financeable asset class
Lenders are exploring whether GPU compute capacity could become a financeable asset class as AI infrastructure investment grows. GPUs power many AI services and are increasingly offered through GPU-as-a-service business models. Two financing models are emerging, one for buying or leasing physical chips and another for funding businesses that sell access to compute capacity. Experts say the sector lacks standardised documentation and a framework for assessing the long-term value of GPU compute, which makes lending risky.
IMF warns AI could boost EU growth but worsen inequality
An IMF paper told EU finance ministers that AI could lift European productivity by about 1 percent over five years. However it warned that AI could also widen inequality, strain power networks and increase dependence on foreign technology. About 60 percent of workers in advanced European economies are in jobs highly exposed to AI. The paper said Europe needs deeper economic integration and major investment in its own AI industry to avoid relying on the US and China.
AI raising questions about who writes the rules in Fulton County
This story comes to Rough Draft through a content partnership with Capitol Beat. Learn more at https://capitolbeatatlanta.com/. Some worry artificial intelligence could one day rule humanity. Robots have not taken over yet, but AI may be helping with a core function of democracy in Georgia’s largest county: writing ord
Sources
- NVIDIA CEO Jensen Huang rejects ‘AI will end the world’ claim, yet cautions ‘we should go as fast as we can but...’
- Nvidia CEO says AI development should move ‘as fast as we can’
- TypeSafe AI Releases Jev: A System One Model That Returns Typed, Calibrated Decisions Instead of Text
- Jev Cuts AI Decision Costs 100x And Vercel, Cloudflare Rushed To Add It
- An Open-Source AI Agent Hit 135,000 GitHub Stars In Weeks. Then One Webpage Hijacked Its Control Interface.
- Anthropic eyes blockbuster IPO as debate over AI regulations ramps up
- Rwanda and Hesse explore AI investment and technology partnerships
- Zoom Communications (ZM) Rolls Out An AI Powered Revenue Operating System
- Artificial intelligence: can GPU compute become a financeable asset class?
- IMF tells EU ministers AI could boost growth but increase economic strains
- AI raising questions about who writes the rules in Fulton County
Comments
Please log in to post a comment.