AI News · 2026-06-20

💡

Jason Says

Today's biggest signal is the Jumper + Trump combo: a Nobel laureate defects to Anthropic while the White House quietly walks back its threat designation—Anthropic is staging a quiet double upgrade in both talent and political legitimacy.

🛠️
AI ToolsTechCrunch AI

Nobel Laureate John Jumper Leaves DeepMind for Anthropic

AlphaFold architect and 2024 Nobel Chemistry laureate John Jumper is departing Google DeepMind for Anthropic, part of a broader talent exodus from DeepMind. The move signals Anthropic's growing ambitions in scientific AI and could accelerate its work in structural biology and drug discovery.

🛠️
AI Tools36Kr / TechCrunch AI

Trump: Anthropic No Longer a National Security Threat

President Trump says he no longer views Anthropic as a national security threat, reversing the position that led to the forced removal of Fable 5 and Mythos 5 models. The policy reversal, combined with backlash from cybersecurity researchers, has inadvertently boosted Anthropic's brand visibility.

🛠️
AI ToolsInterconnects

Banning Open Source AI Would Be a Mistake, Experts Warn

An op-ed argues banning open-source AI would repeat history's failed export control playbook—from PGP to now—and damage US competitiveness without actually stopping proliferation. The same jailbreaks that triggered Mythos restrictions exist across closed models too, making blanket bans security theater.

🛠️
AI ToolsHacker News / Asia AI

Beyond DeepSeek's $7.4B Series A: China's AI Alliance Reshaping

DeepSeek's $7.4B Series A signals more than just a funding milestone—it marks a strategic realignment of Chinese AI capital toward homegrown models amid mineral tensions with the West. Eastern Asia is accelerating AI supply chain independence, posing a structural challenge to Western AI dominance.

💰
MonetizationTechCrunch Venture

YC Spring 2026 Demo Day: 11 Standout AI Startups, Valuations Over $175M

TechCrunch surveyed VCs to identify 11 standout startups from YC's Spring 2026 batch, with some commanding valuations exceeding $175M. The cohort reveals the current hottest AI application verticals and serves as a useful market signal for indie developers and founders.

🔥
SkillsHacker News

FERNme: Agent Memory Graph with Near-Zero LLM Calls

FERNme is an open-source graph-based persistent memory system for AI agents that uses fuzzy-edge Hebbian co-occurrence rules to maintain memory with near-zero additional LLM calls. It dramatically cuts token costs while building personalized user knowledge graphs over time—ideal for memory-augmented agents.

🛠️
AI ToolsGitHub Trending

Palmier Pro: Open-Source macOS Video Editor Built for AI Agents

Palmier Pro is a Swift-native open-source video editor for macOS Tahoe (Apple Silicon) where users and AI agents co-edit directly on the same timeline. With built-in generative AI, it aims to be an AI-native alternative to Premiere Pro—one of the few truly human-agent collaborative creative tools.

🛠️
AI Tools36Kr

Norway to Ban Generative AI for Primary School Students

Norway's PM announced legislation banning generative AI use for students aged 6-13, with strict supervision required for middle schoolers. It's one of the most systematic government-level AI restrictions in education yet, and may trigger a wave of similar policies across Europe and beyond.

💰AI Funding Roundup

ZuzuZoos(查无此园)

Pre-A数千万人民币

Investors: 锦秋、上海复容

AI 陪伴机器人+潮玩 IP 融合品牌,主打有「活人感」的硅基伙伴,团队来自 moody、大疆、泡泡玛特,差异化切情绪体验赛道。

Subscribe for daily AI updates + free playbook

📘 Subscribe Free