Overnight the Iran war jumped another rung — Brent crude cracked $101 and Treasury tripled its bond buyback to calm jittery markets — while inside AI circles a senior Anthropic safety researcher quit warning of a >10% chance AI kills everyone by 2030, even as OpenAI's celebrated Navier–Stokes proof kept unraveling under scrutiny; watch oil, the Fed's rate math, and whether the AI-doom exodus becomes a pattern today.
New Hampshire Democrats nominated Rep. Chris Pappas and Republicans picked John Sununu for the state's open Senate seat, setting up a marquee general-election fight for a chamber where the GOP holds a slim majority.
Stefany Shaheen won the Democratic primary for New Hampshire's 1st District, the seat Chris Pappas is vacating to run for Senate, despite attacks that she was riding her mother's political name; the district is genuinely competitive, having gone to Kamala Harris by just two points in 2024.
Missouri's Supreme Court scheduled a contempt hearing in the ongoing legal battle over the state's congressional redistricting map.
The RNC opened its midterm convention with Trump taking center stage, as Republicans try to shore up enthusiasm ahead of a midterm cycle where his approval ratings are a drag on down-ballot candidates.
The U.S. military says it destroyed five Iranian oil tankers in the Strait of Hormuz after Iran's Revolutionary Guard twice fired on an American Navy warship, the sharpest flare-up yet in a six-month war largely fought over control of the strait. Iran responded by claiming it struck ten vessels near Hormuz, and oil prices jumped past $100 a barrel for the first time since July.
A researcher who says he quit Anthropic in protest, Jacob Coxon, wrote that both Anthropic and OpenAI are "racing straight to self-improving superintelligence and gambling with our lives." A colleague who remained at the company followed up publicly, estimating a greater than 10% chance AI kills all humans within a decade and saying Anthropic lacks a plan to solve alignment for superintelligence.
A deep look at the voter base propelling the AfD's rise in German state elections, as the far-right party's surge unsettles Berlin's governing coalition and rattles allies abroad.
Bangladesh has confirmed more than 1,000 measles deaths and nearly 190,000 infections since March, in what health officials call the world's worst ongoing outbreak, after a vaccination campaign fell short of coverage targets.
Following Jacob Coxon's resignation over safety concerns, a remaining Anthropic researcher, Evan Hubinger, said publicly he believes there's a greater than 10% chance AI could kill all humans within a decade, adding the company does not yet have a plan to solve alignment for superintelligence even though it is trying its best.
Brent crude broke past $101 a barrel overnight, its highest level since July, as U.S.-Iran fighting around the Strait of Hormuz intensified; traders are now pricing in the risk of a sustained supply disruption through one of the world's most critical oil chokepoints.
The Treasury tripled its usual long-term bond buyback to $6 billion in an attempt to steady debt markets, but the 10-year yield still climbed to its highest level since 2023, suggesting investors weren't fully reassured by the move.
A deep explainer on the Langlands correspondences, the decades-long project connecting far-flung branches of mathematics like wormholes between fields that seemingly have nothing to do with each other, and why hundreds of mathematicians consider it one of the deepest unifying ideas in the discipline.
Speaking at Goldman Sachs' Communacopia tech conference, CoreWeave's CEO attributed the growing public backlash against AI to the sheer speed of change rather than the technology's substance, acknowledging that pace itself is unsettling people.
Google announced a record $15 billion AI infrastructure investment in Finland, a country it's nicknaming the "Texas of Europe" for its cheap, abundant power, underscoring how capital continues pouring into AI buildout even amid rising safety anxiety.
An NYU mathematician's own Navier-Stokes statement resurfaces right as OpenAI's rival proof is under fire, and HN is dissecting who actually deserves credit.
Meta's new personal AI agent launch, with the thread split between genuine interest and skepticism about yet another assistant.
Google DeepMind's open genomics release is drawing praise as real open-science infrastructure rather than another closed model demo.
A developer's blow-by-blow of using an AI coding agent for a trivial UI change turns into a broader debate about how much trust to put in agentic coding tools.
A from-scratch hardware build log that HN loves precisely because it's the opposite of an AI wrapper post.
Tailwind CSS's creators are joining Shopify, sparking the usual open-source-acquisition anxiety about roadmap and licensing.
DeepSeek's newest cheap-and-capable model release keeps the open-weight price war with US labs escalating.
A pitch for small, on-device models is landing well with a crowd increasingly tired of cloud-inference costs and latency.
A researcher's firsthand account of how easily malware gets through Google's ad review is fueling fresh distrust of ad-platform moderation.
Science magazine's own writeup of the OpenAI Navier-Stokes dispute, adding institutional weight to a story that started as a social-media fight.
A genuinely delightful bit of plant biology about wood tissue acting like a muscle, popular purely for the wow factor.
OpenAI's case study on AI-assisted quantum experiment design lands amid a week where its own credibility on scientific claims is under scrutiny.
A software-history retrospective on Lotus Notes becomes a cautionary tale HN keeps applying to every rewrite debate.
The classic Ian Frazier satirical legal brief resurfaces on HN's front page for a fresh round of appreciation.