Today in AI

The day in tech, in order.

A day of reckonings: Character.AI files for bankruptcy, Microsoft bets its AI future on Maia silicon, Apple goes all-in on the smart home, and AI agents run into a bank vault.

October 14, 2026 · 10 stories

  1. MorningAI Labs

    Google's Gemini 4 Argon goes live worldwide

    Google made Gemini 4 Argon generally available on October 13 — its first Gemini 4 model, built for long-horizon ‘deep work’ in coding, finance, law, cybersecurity, and science, with a 1M-token context. Intro API pricing is $2/$10 per million tokens until November 30, then doubles. The launch follows the canceled Gemini 3.5 Pro and a DeepMind leadership overhaul. Read our full guide: museaicodes.com/guides/gemini-4-argon

    Source: Reuters
  2. EveningTraction

    Character.AI files for Chapter 11; Disney set as $20M stalking-horse bidder

    Character.AI filed for Chapter 11 bankruptcy on October 12, naming Disney as the stalking-horse bidder for some IP assets at $20 million. Two years after Google's $2.7 billion licensing deal, the chatbot platform says it could be wound down if no buyer keeps it running.

    Source: CNN
  3. EveningLaunch

    Microsoft unveils Maia 300, plans multi-gigawatt deployment starting 2027

    Microsoft announced its third-generation Maia 300 AI chip and will scale the Maia family to multi-gigawatt deployments in 2027, making it the primary inference system for its products. CEO Satya Nadella called it Microsoft's biggest bet yet on its own silicon.

    Source: Microsoft / The Decoder
  4. EveningTraction

    Fields Medalist says OpenAI 'annihilated' his research field

    Hugo Duminil-Copin told a French Mathematical Society video call that OpenAI's release of 719 AI-generated math manuscripts wiped out his area of research, with 16 problems he personally worked on among those claimed solved. Le Monde reports the 2022 medalist cried so intensely it woke his daughter.

    Source: Le Monde
  5. MorningTraction

    Zuckerberg: Meta has 'caught up' with OpenAI and Anthropic at the AI frontier

    In an FT interview timed to the Muse Wonder launch, Mark Zuckerberg said Meta has reached the AI frontier alongside OpenAI and Anthropic, and rejected calls to slow AI development. Usage data published the same day showed Muse keeping 73% of new users day after install, nearly twice ChatGPT's rate.

    Source: Financial Times (via The Inference)
  6. MorningLaunch

    Liquid AI releases open-weight LFM2.5-24B-A2B for on-device agents

    Liquid AI released LFM2.5-24B-A2B, a 24B-parameter reasoning model with only 2B active parameters per token, as open weights on Hugging Face. The company claims 184 tokens/second on an Apple M5 Max and top marks on several agentic benchmarks.

    Source: Liquid AI
  7. AfternoonSecurity

    Anthropic warns small AI models are 'months away' from self-replicating as IoT 'worms'

    Anthropic's researchers say small models may soon copy themselves across smart ovens and off-brand Android TV boxes without human help, forming persistent botnets with no off switch. The company ran distributed inference across 15 $55 TV boxes to prove the concept.

    Source: Reuters (via Economic Times)
  8. AfternoonLaunch

    OpenAI in talks to hand shuttered Sora to Katzenberg filmmaker startup

    OpenAI is reportedly negotiating to give its shelved Sora video model to an unnamed startup co-founded by Jeffrey Katzenberg, ex-Sora research lead Bill Peebles, and ex-Dropbox CFO Sujay Jaswa, which would build pro film-production software on it. OpenAI would take an equity stake. Reported by The Information; not yet confirmed by OpenAI.

    Source: The Information (via PYMNTS)
  9. EveningLaunch

    Apple goes big on the smart home: HomeView, HomePod mini 2, new Apple TV 4K

    Apple's 'Welcome Home' announcements include the $299 HomeView smart-home hub with a 7-inch display, a $149 HomePod mini 2, a $299 fourth-gen Apple TV 4K with the A19 chip, and a Home Collection accessory line built with LG and Schneider Electric.

    Source: MacRumors
  10. EveningPolicy

    JPMorgan blocks outside AI agents like Muse from customer accounts

    CEO Jamie Dimon said on the Q3 earnings call that JPMorgan won't let outside AI agents operate inside customer accounts, warning of 'rogue agents' and a 'liability transfer' for users who authorize them. Agents keep read-only access via open-banking connectors, but can't move money.

    Source: CNBC (via Banking Dive)

Timelines are composed each morning from the previous day's verified news — every story links to the source it was checked against. See the full feed on the news page.