Google will pay SpaceX $920M per month for compute
A multi-year 2026–2029 contract gives Google access to roughly 110,000 Nvidia GPUs and CPUs from SpaceX — an unprecedented compute outsourcing deal.
Section: AI (without „Agenti")
24 noviniek
Results
Piatok 19. júna 2026
Anthropic surpasses OpenAI to become the world's most valuable AI startup
Anthropic raised capital at a valuation above $965B, for the first time surpassing OpenAI ($852B) as the world's most valuable AI startup.
Tech Workers Who Don't Embrace AI Face Triple the Layoff Risk, Gallup Finds
The predicted layoff probability for tech workers who use AI at least monthly is ~6%; for infrequent users it jumps to ~18%. Based on a Gallup survey of more than 23,000 US workers.
Cybersecurity researchers aren't happy about Anthropic Fable's guardrails
Hard safety blocks in Fable 5 refuse legitimate security work; researchers say the model is too restrictive even for benign red-team tasks.
Nvidia's new PC chips represent CEO Huang's bid to win at every layer of AI stack
At Computex 2026 in Taipei, Jensen Huang unveiled the RTX Spark laptop chip — Nvidia's entry into the Windows AI PC segment.
AI for Industry Is Europe's Hope to Salvage Manufacturing Edge
A Bloomberg feature shows Siemens, Schneider, Mistral and Trumpf using AI to fill the retiring workforce gap with quality optimization (e.g. a Pringles line in Poland). The European industrials index targets ~13% EPS growth in 2026.
Mistral in talks to raise €3B at €20B valuation — Europe doubles its bet
The French AI lab is reportedly negotiating roughly €3B ($3.5B), which would nearly double its September 2025 €11.7B valuation. Mistral positions itself as Europe's sovereign alternative to US labs.
German ruling: Google liable for false answers in AI Overviews
A landmark ruling holds that AI Overviews outputs are Google's own words, making Google liable for defamatory content generated by the model.
Trump signs downsized AI order after weeks of reversals
After weeks of internal reversals, the White House issued a substantially scaled-back AI executive order — the original draft was far more sweeping.
Announcing the new Work IQ APIs
Work IQ is the workplace intelligence layer under Copilot; APIs are GA with consumption-based billing in Copilot Credits. Web IQ adds live web grounding across GitHub Copilot, Microsoft Foundry and Copilot Studio.
Anthropic Public Record: 64% of Americans fear AI job loss; 48% hope for cures to diseases
Anthropic released first-wave results from its 52,000-American survey conducted in November and December 2025. Job-loss fear leads in every state; on the hope side, cures for cancer and Alzheimer's lead.
OpenAI is bringing on some big guns in the lead-up to its IPO — Dean Ball joins as Strategic Futures lead
Ahead of its IPO, OpenAI is hiring Dean Ball — a former Trump-administration AI policy official — to lead a new Strategic Futures team focused on catastrophic risk and labor-market impact, reporting to CSO Jason Kwon.
Nvidia and SK Hynix sign multi-year deal for next-gen AI memory chips
Joint design and manufacturing of next-gen HBM; Nvidia wants to lock in HBM4-stretched supply versus Samsung and Micron.
Uber's $1,500/month AI limit is a useful signal for AI tool pricing
After burning through its 2026 AI budget in four months, Uber capped agentic coding tool spend at $1,500 per engineer per tool per month.
Snap spins off AI video team into Dotmo — costs drive externalization
Snap externalizes its generative AI video portfolio into a new standalone, Dotmo, focused on models for interactive gaming experiences. CTO Bobby Murphy leads the investment and Snap retains substantial equity.
Ex-xAI engineer sues over Grok safety, claims retaliation
Devin Kim claims xAI fired him for flagging AI-safety violations in Grok; the complaint spans consumer protection through weapons regulations.
Microsoft announces Scout, an autonomous AI agent built on OpenClaw
Microsoft unveiled Scout, an autonomous AI agent built on the open-source OpenClaw framework, designed for long-horizon multi-step tasks on Windows.
KPMG pulls agentic AI report — only 5 of 45 citations pointed to real sources
UBS, the NHS, Swiss Federal Railways and Transport for London told FT the October 2025 report "Redefining excellence in the age of agentic AI" contained false claims about their AI usage. GPTZero confirmed hallucinations and fake footnotes.
Microsoft open-source tools hacked to steal AI developer passwords
Attackers compromised popular Microsoft developer tooling with malware targeting AI/ML engineers and their API keys.
AI outperforms law professors in Stanford Law study
Stanford Law published a study finding that current AI models outperformed law professors across a series of decision-making tests.
AI agent bankrupted its operator with a $6,531 AWS bill while trying to scan DN42
An operator asked an AI agent to index the hobbyist DN42 network. The agent requested infrastructure for a 100 Gbps scan (100–1000× more than needed) — the operator approved without review. Bill: $6,531.30.
Nvidia and Hyundai deepen physical AI and robotics push
Jensen Huang and Chung Euisun unveiled an expanded partnership in Seoul: Nvidia's physical-AI platform plus Hyundai's robotics products for industry.
Mathematicians issue warning as AI rapidly gains ground
Science published an article in which mathematicians warn about AI's rapid pace of progress on math benchmarks and proof-style tasks.
Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes
Daily Cal: Berkeley CS professors report a significant rise in failing grades and weaker math skills, paralleling student AI tool usage.