A for AI

News Archive

Pick a date to read that day's stories.

June 2026

Days with a blue dot have stories.

DISPATCH
Tuesday, June 30, 2026
  1. 01 of 10ANTHROPIC

    Anthropic releases Claude Sonnet 5 at mid-tier price

    Anthropic released Claude Sonnet 5, offering strong performance for its lower price. This model is ideal for companies building their own software.

    OUR TAKENothing says "we're confident in our valuation" quite like undercutting your own flagship before going public.

    Read more →
  2. 02 of 10ANTHROPIC

    Claude Science workbench opens beta for researchers

    Anthropic launched Claude Science, an AI assistant for researchers. It helps scientists with their experiments by using various tools. This beta version is a major move into academic work.

    OUR TAKEScientists who survived grad school on instant noodles will now survive on Claude, marginally more nourishing.

    Read more →
  3. 03 of 10MICROSOFT

    Microsoft's Memora project targets AI agent memory gap

    Microsoft's Memora gives AI programs a better way to remember things. It organizes information, helping them recall details for smoother conversations. This makes the AI much more consistent over time.

    OUR TAKEIt remembers everything except, one suspects, your preferred font in Word.

    Read more →
  4. 04 of 10APPS

    Startup Acti embeds AI agents inside your keyboard

    A new phone keyboard helps you do tasks in other apps. It learns from what you type to act for you.

    OUR TAKEFinally, an AI that lives where you already misspell everything.

    Read more →
  5. 05 of 10FINANCE

    Ex-DeepMind poker AI team now beats markets

    A team of past AI researchers now helps hedge funds with trading. Their methods for playing poker are now used to make money in markets.

    OUR TAKEFrom bluffing at the river card to bluffing at the yield curve, a perfectly natural career progression.

    Read more →
  6. 06 of 10NVIDIA

    Meituan's trillion-parameter model runs on Chinese chips

    Meituan's new computer model was built entirely with Chinese-made parts. This shows China's computer progress despite export limits.

    OUR TAKEExport controls: theoretically a ceiling, practically a motivational poster.

    Read more →
  7. 07 of 10EU

    EU Council formally adopts AI Act simplification package

    Europe has voted to simplify its new rules for artificial intelligence. Businesses get clearer deadlines, but some worry this lessens safety checks.

    OUR TAKEThe EU insists it hasn't watered down the rules, which is exactly what one says when watering them down.

    Read more →
  8. 08 of 10SAFETY

    Senator Warner proposes first US AI agent consumer bill

    Senator Mark Warner's new bill aims to make smart computer programs transparent and accountable. It would be the first US law focusing on these agents that act for people.

    OUR TAKECongress has decided to regulate AI agents, presumably once their own AI agents finish scheduling the hearing.

    Read more →
  9. 09 of 10SAFETY

    Congress turns up heat on AI harms to children

    Lawmakers are looking into how websites using artificial intelligence affect children's minds. There is broad agreement that big tech companies should be held responsible.

    OUR TAKEBipartisan agreement in 2026: rare, precious, and apparently only achievable when the subject is protecting children from TikTok.

    Read more →
  10. 10 of 10FINANCE

    Pie launches with $19.5M to give SMBs AI marketing

    Pie launched with $19.5 million to help small businesses with marketing. Their new product, Front Desk, uses clever computer programs to handle customer tasks.

    OUR TAKESilicon Valley has discovered the corner bakery, expect a pivot to "hyperlocal agentic sourdough intelligence" by Q3.

    Read more →