💼 Nvidia CEO says AI is a job creator, not a job killer

🤖 Anthropic co-founder thinks AI will start building itself by 2029

📱 OpenAI wants to build AI phone by next year

🌊 Peter Thiel is building AI data centers in the ocean

🏛️ Big Tech is handing Washington the keys to its AI models

+ 📊 Daily poll and results

+ 📈 Trending tools and resources

Nvidia CEO Jensen Huang told a Milken Institute audience that AI is generating jobs at an industrial scale, arguing that automating a task does not eliminate the broader role an employee fills in an organization.

Huang pointed to AI hardware factories as proof, saying they require large workforces and that the growing AI industry overall is creating more positions than it displaces.

BCG research suggests up to 15% of U.S. jobs could be eliminated by AI in coming years, which makes Huang's optimism easier to hold when your company sells the machines driving that shift.

Anthropic co-founder Jack Clark published a blog post giving 60%+ odds that AI systems will train their own successors before 2029, backing the claim with public benchmark data showing rapid AI capability gains across core R&D tasks.

METR data shows AI's independent work window grew from 30-second tasks in 2022 to 12-hour runs in 2026, while coding benchmark SWE-Bench jumped from 2% to 93.9% accuracy in under three years.

OpenAI is already targeting an automated research intern by September 2026. If that timeline holds, self-improving AI stops being a thought experiment well before Clark's 2029 deadline.

Analyst Ming-Chi Kuo says OpenAI has accelerated its AI smartphone timeline, now targeting mass production in first-half 2027, a year ahead of the previously reported 2028 schedule, with MediaTek's Dimensity 9600 as the likely sole chip.

The push is driven by OpenAI's IPO ambitions and rising competition in AI phones. Kuo projects combined 2027-2028 shipments could hit 30 million units if development stays on track.

This complicates OpenAI's own hardware story: Jony Ive's screenless AI device, once billed as the "coolest piece of technology the world has ever seen," has already slipped to early 2027 as a smart speaker with a camera.

PayPal and Palantir co-founder Peter Thiel led a $140M Series B for Panthalassa, an Oregon startup building self-steering ocean platforms that convert wave energy into power for AI chips, at a near-$1B valuation.

Each 85-meter steel node uses seawater for cooling, Starlink for connectivity, and its hull shape alone to navigate, with no engines. The funding will deploy the first nodes in the Pacific by 2027.

Land-based data centers are facing growing public backlash over energy and water use, making the ocean a more politically frictionless frontier, at least until someone has to fix one mid-Pacific.

Google, Microsoft, and xAI signed agreements with the Commerce Department's CAISI to hand over early access to new AI models, sometimes with safety guardrails reduced, for national security evaluation.

The deals follow reports that the Trump administration is weighing a review committee with power to approve AI models before public release, a notable shift from its stated hands-off regulatory stance.

Anthropic refused similar pressure and got labeled a Pentagon supply chain risk. The other three companies appear to have done the math and picked compliance over confrontation.

A) Real reason - 31%

B) Just an excuse - 69% 🏆