Latest updates from the world of artificial intelligence
Weeks after an export-control directive pulled Claude Mythos 5 offline worldwide, the U.S. government has cleared Anthropic to redeploy its most powerful cyber model — but only to a vetted set of organizations that defend critical infrastructure, and with safeguards attached.
CNN
Read more →Stanford economist Erik Brynjolfsson, using ADP payroll data covering one in six US workers, finds employment for 22-to-25-year-olds in the most AI-exposed jobs is shrinking 3.8% a year — and accelerating — while workers aged 35-40 in the same fields are growing. The decline is driven by stalled hiring, not layoffs, and an aggregate 0.2% dip hides the generational split. “It’s not going away,” he says.
Fortune
Read more →Princeton’s new CEO-Bench drops AI agents into the chief-executive’s chair of a simulated startup with $1M and 500 days to grow it. Of the frontier models tested, only Claude Opus 4.8 and GPT-5.5 ended their best runs above the starting cash — and neither did so consistently. Most went bankrupt, exposing how far long-horizon agents still are from the autonomy vendors are selling.
arXiv (Princeton University)
Read more →Anthropic’s June 2026 Economic Index, titled “Cadences,” uses aggregated Claude data to show AI usage rising and falling with the clock and calendar — news requests peak at 7 a.m., recipes hit 2.3× by 6 p.m., and tax queries spiked 8× around the April deadline. Personal use jumps from ~35% on weekdays to nearly 50% on weekends, while higher-wage work consumes about twice the tokens.
Anthropic
Read more →OpenAI released GPT-5.6 on June 27 as a three-model family — the flagship Sol, the balanced Terra and the fast, cheap Luna — but only as a limited preview to about 20 U.S. government-approved partners. Sol leads agentic-coding benchmarks (91.9% on Terminal-Bench 2.1) and matches prior security results using roughly a third of the output tokens, with general availability promised in the coming weeks.
OpenAI
Read more →After three years of falling token prices, the AI pricing race has flipped: GPT-5.5 and Claude Fable 5 launched at double their predecessors, while customers respond by optimizing usage and routing to cheaper challengers like DeepSeek. CNBC reports OpenAI spent ~$1.35 per dollar earned in 2025, even as Anthropic turned a Q2 profit on enterprise coding.
CNBC
Read more →Baseten, the platform that runs AI models in production, raised a $1.5 billion Series F on June 22 at a valuation of up to $13 billion — a roughly 160% jump in under five months. The startup now serves more than a billion inference calls a day across 87 clusters and 18 clouds, as "inference" becomes one of AI's most contested markets.
Business Wire
Read more →Taktile, which automates underwriting, claims, fraud and AML decisions for banks and insurers, raised a $110M Series C led by Goldman Sachs Alternatives on June 24. The platform wraps AI agents in rules and human oversight, powering millions of auditable decisions a day.
Tech Startups
Read more →Assort Health, which builds AI agents for the patient journey — scheduling, intake, referrals, refills and billing — raised a $120M Series C led by Menlo Ventures on June 24, reaching a $1.2B valuation after handling 190M+ patient interactions and growing revenue 20x in 15 months.
Tech Startups
Read more →Anthropic told the White House and U.S. senators that Alibaba’s Qwen lab used roughly 25,000 fake accounts to run nearly 28.8 million Claude queries between April and June — harvesting its coding and agent skills in what Anthropic calls the largest distillation campaign it has ever documented. Alibaba declined to comment.
The Next Web
Read more →OpenAI is staggering GPT-5.6’s public release after the Trump administration asked it to limit initial access to government-approved partners on national-security grounds. Sam Altman told staff on June 25 the government will approve access ‘customer by customer’ during a limited preview, with a broader rollout a couple of weeks later — and prediction markets have slashed the odds of a June launch from 83% to 18%.
The Information
Read more →ByteDance launched its Doubao Seed 2.1 Pro and cheaper Turbo models on June 24, built for the “coding and agent era.” Seed 2.1 Pro scored 1539 on the Code Arena: Frontend leaderboard — eighth in the world and level with Claude Opus 4.6 — while ByteDance claims three core capabilities are comparable to GPT-5.5.
AIbase
Read more →OpenAI unveiled its first in-house chip on June 24 — Jalapeño, an LLM-inference ASIC co-built with Broadcom. Designed to tape-out in about nine months with help from OpenAI’s own models, it ships from late 2026 and targets far better performance-per-watt.
TechCrunch
Read more →Apple raised prices on Macs, iPads and home devices on June 25, blaming an AI-driven memory shortage: with DRAM up as much as 98% and chipmakers prioritizing data-center orders, the MacBook Air now starts at $1,299 and the iPhone may be next.
Reuters
Read more →AlphaFold co-creator and 2024 Nobel laureate John Jumper is leaving Google DeepMind after nearly nine years to join Anthropic, a marquee hire that hands the Claude maker instant credibility in its growing push into AI for science.
TechCrunch
Read more →Mark Zuckerberg has ordered Meta to build a standalone prediction-market app, codenamed “Arena,” where users bet play money on real-world events — with Llama generating the questions, recommending markets, and resolving outcomes in near real-time.
NPR
Read more →Anthropic launched Claude Tag on June 23 — tag @Claude in a Slack channel and it breaks the task into stages, works asynchronously with a shared team identity, and posts results back in the thread. Running on Opus 4.8 with an optional proactive "ambient" mode, it ships as a research preview for Enterprise and Team customers. Anthropic says 65% of its own product team's code now ships through it.
TechCrunch
Read more →OpenAI says GPT-5 Pro helped immunologist Derya Unutmaz explain a T-cell result that had stumped his Jackson Laboratory lab since 2022 — pointing to disrupted N-linked glycosylation and memory T cells, then correctly predicting that 2-DG priming would sharpen anti-CD19 CAR-T killing of lymphoma. The lab confirmed it at the bench.
OpenAI
Read more →Engram emerged from stealth on June 23 with $98 million from General Catalyst, Kleiner Perkins and Sequoia — plus angels including Andrej Karpathy — to build a "learned-memory" layer that it says runs enterprise AI on up to 100× fewer tokens. Microsoft, Notion and Harvey are already testing it.
CNBC
Read more →Reports this week floated Claude Sonnet 5 (codename 'Fennec') for the week of June 23, alongside GPT-5.6 — but Anthropic has confirmed nothing. And the same 'Sonnet 5' leak in February shipped as Sonnet 4.6, a reminder to read codenames with caution.
Crypto Briefing
Read more →An AI agent is a system that uses an LLM to pursue a goal on its own — reasoning, calling tools, observing results, and looping until done. This plain-English guide explains agentic AI from scratch: how agents work, what they can do today, and how to get started.
BitsMinds
Read more →In three days Google lost two AI stars: Gemini co-lead Noam Shazeer left for OpenAI on June 18, and Nobel laureate John Jumper left DeepMind for Anthropic on June 19. Alphabet shares fell about 7% as the frontier talent war intensified.
CNBC
Read more →A practical, step-by-step guide to Robinhood Agentic Trading: connect Claude, Claude Code, or ChatGPT to the Robinhood Trading MCP, fund a ring-fenced agentic wallet, set approval guardrails, and let your AI agent analyze and trade equities — safely.
Robinhood
Read more →Anthropic announced a partnership with Google and Broadcom on June 23 for multiple gigawatts of next-generation TPU capacity, coming online from 2027 and sited mostly in the U.S. — backing a run-rate that has climbed from $9 billion to over $30 billion.
Anthropic
Read more →