Welcome, humans.

🔴 We’re LIVE right now for today’s marketing tools showcase. Come join!

Today, we're bringing in four founders to demo how they're using AI to transform marketing: from smarter ad buying to creator automation to creating the perfect messaging and marketing your products directly to ChatGPT (yes, really).

These four founders will show you what's actually possible w/ AI in marketing right now:

• 11am PST: Jonathan Gudai from AdOmni on how his Jeen AI tool is making ad buying smarter and more efficient.

• 11:30am PST: James Cadwallader from Profound on the wild new frontier: marketing your products directly to AI systems like ChatGPT and Claude when they're making recommendations to users.

• 12:00PM PST: Jonathan Meyers from Agentio sharing how to automate the entire creator campaign lifecycle (from matching to payment) using their platform.

• 12:30PM PST: Crystal Foote from DCG showing their Resonance Pathway tool that maps the fastest route between your brand and audience emotion.

Each gets 15-20 minutes to show you what their platform actually does, followed by Q&A. These are teams backed by Sequoia, Kleiner Perkins, and Khosla Ventures, already working with Uber, DoorDash, and Reddit.

P.S: OpenAI’s “Garlic” model (GPT 5.2) just dropped (more on that below)… stay to the end (1pm PST / 3pm CT) gand we’ll test it out live!

Here’s what happened in AI today:

1. OpenAI released GPT-5.2 with a 400,000-token context window

2. Disney invested $1 billion in OpenAI and became Sora's first partner

3. Stanford researchers built Artemis, an AI hacking bot that outperformed pros.

4. State attorneys general demanded AI companies fix chatbot outputs linked to suicides.

GPT-5.2 "Garlic" Just Launched—And Disney Bet $1 Billion on It

OpenAI just released GPT-5.2, internally codenamed “garlic,” and it's their most powerful model yet. The timing? Not coincidental: Disney simultaneously announced a $1 billion investment in OpenAI and became the first major content partner for Sora.

This is the biggest enterprise AI deal we've seen in 2025, and it signals where the industry is heading: from experimentation to full-scale deployment. Technically I think it’s Disney’s first big move into AI, too? But we’ll have to confirm that another time.

Here's what GPT-5.2 brings:

• GPT-5.2 is OpenAI's new flagship model, built specifically for coding and agentic tasks across industries.

• It has a massive 400,000-token context window (that's roughly 300,000 words), letting it process entire codebases or lengthy documents in one go.

• The model outputs up to 128,000 tokens per response and includes reasoning token support for complex problem-solving.

Pricing: $1.75 per million input tokens, $14 per million output tokens. For context, that's 40% more expensive than GPT-5, but OpenAI says the performance gains justify the premium for enterprise use cases.

The Disney partnership explained:

Disney isn't just investing—they're becoming one of OpenAI's largest customers. The deal includes:

• Sora access: Disney gets to generate short videos featuring 200+ characters from Disney, Pixar, Marvel, and Star Wars (think Mickey Mouse, Iron Man, Darth Vader).

  • Fans will create these videos, and curated selections will appear on Disney+.

• Internal deployment: Disney will use OpenAI's APIs to build new products for Disney+ and deploy ChatGPT company-wide for employees.

• Equity stake: Disney receives warrants to buy additional OpenAI shares alongside their $1B equity investment.

The companies emphasized their “shared commitment to responsible AI: and protecting creator rights (for example, the licensing agreement specifically excludes talent likenesses and voices).

Why this matters: Disney's $1B bet signals that Fortune 500 companies are ready to integrate AI at scale, not just pilot it in side projects. Disney also choosing today to send a cease and desist to Google for “massive” copyright infringement is also significant; pay up, or pay out.

As for more thoughts from us? We gotta test it first. Join us on the stream at 1pm PST to test GPT 5.2 live 😈

Read more: OpenAI's full announcement | GPT-5.2 model card

FROM OUR PARTNERS

From finding answers to getting work done

Knowledge management tools help you find information. Sana Agents takes it further.

With Sana’s AI platform, your powerful agents can search your company's knowledge, then complete the work. Pull that pricing doc from Drive, analyze the data inside it, generate a presentation with the insights, and update your CRM—all from one prompt across 100+ tools.

With Sana Agents, teams conduct hours of deep research in seconds, analyze complex datasets 10x faster, and cut sales prep time by 90%. Enterprise-grade connectors, fully maintained. Robust security with zero data leakage.

This isn't about finding documents faster. It's about getting work done.

See how it works

Prompt Tip of the Day

This is just so cool

Treats to Try

*Asterisk = from our partners (only the first one!). Advertise to 600K readers here!

1. *AI You Can Audit. Every Guru answer is cited, permission-aware, and logged—so your AI is finally accountable. Watch a demo.

2. Z.ai launched GLM-4.6V, an open-source vision-language model family, with a 106B flagship and a 9B Flash variant that supports native multimodal tool calling (can use connectors and call functions), has a 128K context window (so ~100K+ words), and competitive scores on 20+ benchmarks under an MIT license with weights on Hugging Face (meaning you can spin it up and run it).

3. Cohere's Rerank 3.5 reorders your search results to show the most relevant documents first, working across 100+ languages and complex data like emails, tables, JSON, and code—paid only ($2/1,000 searches).\

4. ClickUp 4.0 pulls your projects, chats, docs, automations, and its Brain AI assistant into one “converged” workspace so you can ask @Brain to answer questions, schedule work, and spin up agents instead of bouncing between tools—free to get started.

5. Helploom gives you unlimited live chat, a full help center, and an AI support bot that learns your knowledge base so you can support customers 24/7 without per-seat pricing—free tier available, then $29/month Base or $59/month with automated AI replies.

6. Dex turns your Chrome browser into a single AI workspace that remembers your tabs, tasks, and context so you can jump between research, planning, and execution while an in-browser copilot helps move work forward—no clear public pricing yet.

7. PrivacyFirewall is a local-first “AI privacy shield” that runs on your machine to block PII and secrets before you paste them into ChatGPT, Claude, or Gemini, using on-device rules and a lightweight transformer model to flag risky text as you type.

Around the Horn

1. Rivian will ship a built-in AI assistant to every existing EV in early 2026, using its new Rivian Unified Intelligence platform plus Gemini and Vertex AI to let drivers control car functions and connect apps like Google Calendar via an agentic framework.

2. Google launched “Disco,” a Gemini-powered Chrome experiment that turns your open tabs into lightweight web apps called GenTabs, suggesting interactive tools based on whatever you’re browsing and letting you build custom apps with plain-language prompts.

3. Runway released GWM-1, its first physics-aware world model for simulating reality, and updated its Gen-4.5 video generator with native audio so creators can generate visuals and sound together instead of stitching them in post.

4. TIME named the “Architects of AI” — a roster including Jensen Huang, Elon Musk, Sam Altman, Mark Zuckerberg, Lisa Su, Dario Amodei, Demis Hassabis, and Fei-Fei Li — as its Person of the Year for reshaping policy, geopolitics, and infrastructure with their AI bets.

5. Stanford researchers built an AI hacking bot called Artemis that beat nine out of ten professional penetration testers on their own network at a fraction of the cost, according to the WSJ, highlighting how offensive cyber tools are getting dangerously close to outperforming human hackers.

6. Medra raised $52M to build one of the largest autonomous labs in the U.S., where AI-programmed robots run and refine biological experiments end-to-end to speed up drug discovery.

7. Waymo confirmed that a passenger in labor gave birth inside one of its San Francisco robotaxis en route to UCSF Medical Center, with remote support flagging “unusual activity” and calling 911 as the car finished the trip on its own.

8. States attorneys general from across the U.S. sent a warning letter to Microsoft, OpenAI, Google and other AI giants to fix “delusional” chatbot outputs linked to suicides and other harms, demanding stronger safeguards, independent audits, and responses by mid-January.

FROM OUR PARTNERS

Can you scale without chaos?

It's peak season, so volume's about to spike. Most teams either hire temps (expensive) or burn out their people (worse). See what smarter teams do: let AI handle predictable volume so your humans stay great.

Check your options

Midweek Wisdom (Thursday Trivia will come tomorrow! Sending for urgency lol)

1. TheAhura wrote about using Claude as a coding partner to painstakingly recreate the 1996 Space Jam website, showing how an LLM can handle everything from scraping old assets to rebuilding retro HTML/CSS while still needing a human in the loop for polish and deployment.

2. SteerLabs warned about the “confident idiot” problem in agents, arguing that LLM-based systems will happily take wrong actions with total conviction unless we treat them like regular software—with hard rules, explicit state machines, and verifiable checks—instead of relying on fuzzy “vibe checks” and clever prompting.

3. Steven Yue argued that “AI should only run as fast as we can catch up,” framing AI progress as a verification bottleneck where teams that invest in fast, rigorous checking of AI’s work will win while everyone else drowns in “verification debt” from unchecked vibe-coding.

4. Google said Gemini 3 Pro delivers state-of-the-art document, spatial, screen, and video understanding, and highlighted how its improved screen reading powers more reliable computer-use agents that can click around real apps and analyze high-frame-rate video like golf swings.

5. Nicholas Liu argued in Jacobin that most AI art is “weird, sad, and ugly” because it’s born from a profit-driven system that severs people from real creative labor, framing gen-AI images as slop frothing up from capitalism’s gap between human desire and commodified output.

6. Pathway got a WSJ profile for its “Dragon Hatchling” architecture, pitched as a post-transformer system with built-in long-term memory and lifelong learning that runs on NVIDIA and AWS and could eventually enable more continuously learning AGI-style models.

A Cat’s Commentary