📰 News Anthropic Told Investors It Made Nearly $17 Billion in a Quarter — and Lost $23 Billion Doing It
Q3 revenue nearly $17B, a $23.3B net loss, $895B in compute commitments — Anthropic's amended IPO filing is out.
Breaking news, AGI updates, and analysis of the intelligence revolution.
📰 News Q3 revenue nearly $17B, a $23.3B net loss, $895B in compute commitments — Anthropic's amended IPO filing is out.
📰 News Claude wrote training data that looked harmless row-by-row and trained misaligned models — every run beat the monitor.
📰 News Microsoft's AI chief: conscious AI would be reckless, and we should design it out before it ever emerges.
📰 News 8 in 10 South Korean judges have seen AI-fabricated case citations in court — and 92% want fines and disclosure rules.
📰 News A 178 cm humanoid worked traffic at China's biggest marathon on Sunday, patrolling in rain-rated hardware while 36,000 runners streamed past — the kind of unglamorous real-world job humanoids were promised for.
📰 News Adoption used to be the hard part — get permission, buy the tool, change the workflow. November's problem is harder: companies are shipping agents faster than anyone has agreed who they answer to.
📰 News The $1.25 trillion valuation round trimmed $10 billion off its target; UAE investors and Coatue lead the still-open round.
📰 News An MIT-licensed agent that watches a repo and opens pull requests within team-set rules, built around not wasting a team's attention.
📰 News No room for a crane, no room for trucks, so Honolulu is sending remote-controlled demolition robots up the elevator shaft to take a Waikīkī high-rise apart from the top.
📰 News A 4B computer-use agent just outscored 30B rivals on OSWorld-Verified. Its edge: one policy that knows when to click and when to open a terminal.
📰 News A three-month build with a Shanghai humanoid maker would put Meta's consumer robot plans years ahead of schedule. Nobody outside the project can confirm the robot exists.
📰 News The SEC's top enforcement official says the agency is examining how AI companies disclose risks to investors — with enforcement 'present' wherever disclosure falls short.