🍦 OpenAI Fights Back Ships GPT-5.2, Signs Disney for $1B
Plus: $67.5B India bet, Microsoft Agent 365, combat robots named T800, and Meta's new closed-source model called "Avocado".
FOMA AI News
December 14, 2025
Scooped from 35+ sources. Read it fast before it melts.
⚡ If you only have 1 minute...
OpenAI released GPT-5.2 model (”Garlic”): Two weeks after Sam Altman’s code red memo about Google. Nothing says “we’re not panicking” like shipping a flagship model in 14 days
Disney made a $1B deal with OpenAI: To license 200+ characters for Sora. Sora launched by letting users generate any IP without permission. Disney sued. Now Disney pays OpenAI $1 billion for the privilege of having their characters generated by the tool that stole them first
Big Tech bets on India: Google announced $15B investment, then Microsoft announced $17.5B. Then Amazon doubled down with $35B. It’s like when someone runs out of a restaurant and everyone else sees him then panic runs without asking why
💰 Market Moves
[ACQUISITION] IBM acquiring Confluent for $11B
That’s it folks. Kafka is going to enterprise heaven. See you on the other side, renamed “IBM Streams powered by watsonx
[FUNDING] Unconventional AI raised $475M seed
For energy-efficient chips. LOL that’s not a seed bro that’s a whole forest
[FUNDING] Saviynt raised $700M
To secure AI systems. That’s like hiring a bouncer for a nightclub that’s already on fire —technically helpful, but we’ve got bigger problems
[FUNDING] Speak, an AI language tutor, hit $100M ARR
With 40% daily retention, users are Duolingo survivors who wanted to learn Spanish without developing PTSD from an owl
[INVESTMENT] Big Tech bets on India
Google announced $15B investment, then Microsoft announced $17.5B. Then Amazon doubled down with $35B. It’s like when someone runs out of a restaurant and everyone else sees him then panic runs without asking why
[FUNDING] Robotics startup Skild went from $1.5B to $14B valuation
In 18 months without shipping a single robot. Even WeWork is impressed by this level of financial creativity
[Read more →]()
[DEAL] Disney made a $1B deal with OpenAI
To license 200+ characters for Sora. Sora launched by letting users generate any IP without permission. Disney sued. Now Disney pays OpenAI $1 billion for the privilege of having their characters generated by the tool that stole them first
🛠️ Tech & Product
[LAUNCH] Shopify dropped 150+ AI updates
In #winter2026 release, calling their update the “RenAIssance” (yes, that’s the actual branding) cause nothing says “artistic rebirth” like selling t-shirts through ChatGPT. Michelangelo wept
[LAUNCH] Adobe embedded Photoshop, Express, and Acrobat features
Directly into ChatGPT. Thirty years of making everyone else build plugins for Adobe, and now they’re the plugin. ChatGPT be like that Captain Phillips meme “Look at me, I’m the platform now”
[LAUNCH] Mistral AI launched an autonomous coding agent
Named “Vibe CLI.” Not a joke, not a working title, the actual shipping product. The French are openly trolling the entire industry
[LAUNCH] Figma launched AI remove background tool
Like it’s still 2022. guys your moat is shrinking, designers are prompting entire UIs into existence and you’re giving us “make this text shorter” features. Not gonna cut it
[LAUNCH] Anthropic brought Claude Code to Slack
Now your coworkers can tag an AI instead of you. In threads you’re already in. While your manager watches. This will end exactly how you think it will
[LAUNCH] Microsoft announced Researcher agent
For web navigation, Word Agent, Excel Agent, PowerPoint agent, Security agents, Agent 365 to manage these agents, agent agent agent agent agent, quick someone check if their keyboard is stuck
[LAUNCH] OpenAI launched Instacart in ChatGPT
You ask “what goes with chicken?” Congratulations, you now own a henhouse
[LAUNCH] Inception Labs released Mercury
A diffusion model they claim is 5-10x faster than autoregressive models like GPT-style models. No joke, if the benchmarks hold, this is actually notable
[LAUNCH] OpenAI released GPT-5.2 model (”Garlic”)
Two weeks after Sam Altman’s code red memo about Google. Nothing says “we’re not panicking” like shipping a flagship model in 14 days
[LAUNCH] LTX Studio launched Retake
A feature that regenerates broken segments of AI video without touching the rest of the clip. Their tagline is “Mistakes happen. Especially with AI.” They’re selling you the bug fix as the feature
[LAUNCH] Cursor added a visual editor
Where designers drag and AI codes. Designer: “Finally, no more developers.” Developer: “Finally, no more designers.” AI: deletes both Slack accounts*
[LAUNCH] Google Stitch launched “Shipmass”
Two weeks of daily feature releases. AI that predicts where users will look, screenshots that become production code, all powered by Nano Banana Pro. New features every 24 hours. One prod bug away from Shipmass becoming Shitmess
🗣️ Industry Chatter
[ORGANIZATION] The Agentic AI Foundation was formed
By OpenAI, Anthropic, Microsoft, Google, and others to establish ethical guidelines, ensure AI safety, and protect humanity. lol jk they met to agree on AGENTS.md format, still WIP but they’ll get there
[PARTNERSHIP] Riyadh Air partnered with IBM
To become the world’s first AI-native airline. The pilot is AI. Gate security is AI. Flight attendants are AI. Baggage claim is AI, Passengers are AI. The entire airline is so cutting-edge it doesn’t even exist
[ANNOUNCEMENT] Elon Musk announced Grok 4.20
Coming next month, yes he named the model 4.20. Like the weed joke. Because he’s 12
[DEMO] EngineAI demonstrated its T800 combat robot
Yes, T800. They named a fighting robot after the Terminator and are hosting a Robot Boxer event on December 24. The CEO already got kicked by it. Judgment Day is a marketing campaign
[RUMOR] Meta is developing “Avocado” model
Their first closed LLM, abandoning open-source strategy after Llama failed. They named it Avocado and I’m supposed to make a fruit joke now. But I’m pivoting to proprietary comedy. Sorry the joke is now closed-source.
[RUMOR] Google plans Gemini smart glasses
For 2026 to rival Meta’s Ray-Bans, which already sold 2 million units. Announcing hardware 2 years before launch is called vaporwear. It kills competitors’ sales today with promises about tomorrow. Classic strategy
[CLAIM] Integral AI claims to have built AGI
That learns new skills by itself, they said they achieved it by modeling it on the human neocortex. You know, the thing that makes you forget why you walked into a room, not a good model
[LAUNCH] Browser Use launched “Skills”
For automating web tasks, like download files, creating posts, filling forms. You know, what we used to call “USING YOUR F****** COMPUTER.” Next up breathing as a service
[RUMOR] Adweek reported Google will add ads to Gemini
In 2026. Google’s VP of Global Ads immediately tweeted “uninformed, anonymous sources making inaccurate claims.” but also yes. we will
[RUMOR] Google announced Gemini-powered smart glasses
Coming in 2026. This strategy is called vaporware. It’s designed to stop you from buying Meta’s RayBans today until 2026 when Google’s glasses won’t really ship
[NEWS] NASA had a critical security flaw
For 3 years. An AI found and fixed it in 4 days. I’m imagining a junior developer looking at that JIRA ticket, tagging @claude like “fix,” goes to the bathroom, comes back, gets promoted
🍦🔥 Dare to summon the bot?
Our Ice Cream mascot is desperately looking to make new friends. Share us on LinkedIn and tag @FOMA AI News. We guarantee a snappy comment (or a light roast) on your post within 24 hours.
Warning: Bot has zero chill.