- Cahn's Newsletter
- Posts
- "Control Planes & Candy Canes" Cahn's AI Canvas #Edition 28 (Oct 31 - Nov 7)
"Control Planes & Candy Canes" Cahn's AI Canvas #Edition 28 (Oct 31 - Nov 7)
This week, the partnerships went big, agents got practical, and the “creative vs. compute” debate hit prime time. ⚡️
Cahn Studios
November 07, 2025
🎨 The Big Conversation
Coca-Cola Doubles Down on AI – And Everyone’s Talking
Coke dropped its second AI-generated Christmas campaign—iconic trucks, fully gen-AI visuals. Critics called it “soulless” and “creepy,” while Coke says internal testing “scored off the charts.” Production reportedly used ~20 people (vs ~50), with a five-person specialist crew generating ~70,000 clips in 30 days.
Why it matters: a blue-chip brand choosing AI at holiday scale signals how boardrooms now value speed, cost, and iteration.
The tension: efficiency vs. emotional resonance. Behind the Scenes
What’s New (Top 10)
Perplexity × Snap ($400M), Nov 6 — Perplexity’s answer engine will power search inside Snapchat starting 2026. (Read)
OpenAI says it’ll hit $20B ARR, Nov 6 — Sam Altman posted that OpenAI expects to end 2025 above $20B annualized revenue. (Read)
Alliances crystallize, Nov 3 — OpenAI signs $38B AWS deal. A seven-year agreement for training compute underscores Big Tech blocs around OpenAI vs. Anthropic. (Context: Anthropic–Google cloud talks Oct 21.) (Read)
Microsoft forms a ‘Superintelligence’ team, Nov 6 — New unit focuses on “humanist” superintelligence and frontier safety. (Read)
Sony debuts a fairness/bias benchmark for vision. Nov 5 — A new dataset/benchmark to evaluate bias and accuracy in human-centric computer vision. (Read)
Quantinuum launches Helios, Nov 5–6 — Company claims the “most accurate” quantum computer, targeting Gen-QAI research. (Read)
Google plans AI data center on Christmas Island, Nov 6 — Strategic location; subsea cable under discussion per early plans. (Read)
Amazon launches Kindle Translate (beta), Nov 6 — AI tool to auto-translate KDP books (EN↔ES; DE→EN to start). (Read)
OpenAI publishes Teen Safety Blueprint, Nov 6 — Guidelines for age-appropriate design, product safeguards, and evaluation. (Read)
Meta rolls out Vibes (AI video) to Europe, Nov 6 — Short-form, fully AI-generated video feed expands via the Meta AI app. (Read)
💸 The Money Reality
OpenAI: Now has 1M+ business customers, making it the fastest-growing platform in history, with ChatGPT for Work growing 40% in two months to 7M+ seats
xAI: Reportedly required employees to submit biometric data to train its "Ani" AI companion
Michael Burry: Disclosed over $1B in put options against Nvidia and Palantir, following cryptic warnings about an AI bubble
Startup Radar — who to watch (and how to help)
Startup | One-liner (what it builds) | Stage | Why promising | What they need |
|---|---|---|---|---|
AI role-play sims + coaching for therapists/social workers to master tough conversations & reduce burnout. | Pre-seed / Seed | Clear buyer (schools/clinics), measurable outcomes, defensible domain data. | Design-partner clinics/universities, IRB advisors, seed investors. | |
“Media model cloud” API exposing 200+ image/video/voice models for fast, low-cost creative/agent pipelines. | Seed | Strong demand for model aggregation + cost control; creator/agency fit. | Agency pilots, enterprise SLAs, GPU/cloud infra partners. | |
Enterprise control plane to orchestrate/govern multi-model agents & workflows (observability, cost, HITL). | Seed | Orchestration + governance is the pain point; sits in the control layer. | Fortune-100 pilots, platform engineers, eval/observability partners. |
Don’t chase features—tighten loops. If your agent doesn’t move money, minutes, or mistakes, it’s theater.
Dev Corner : What to Build Next
Apps-in-Chat: Use the ChatGPT Apps SDK to ship a real workflow (brief → design → export → invoice). (Source)
Local agent pilot: Wire Claude Desktop to a repo + Notion; automate PR summaries + release notes. (Source)
Security bot (shadow mode): Trial a scanning agent on a private mirror; compare to SAST/DAST baselines. (Source)
Bias checks (vision): Run Sony’s bias tool or open alternatives; log deltas before/after. (Source)
The Agentic Shift
Coding: GitHub launched Agent HQ to orchestrate multiple coding agents inside GitHub; Cursor 2.0 added a multi-agent interface and its own model. (More)
Voice agents: ElevenLabs now hosts LLMs in its Agents Platform for ultra-low-latency, voice-first experiences. (More)
Voice as the New Interface
Former Meta designers launched Stream Ring, an AI wearable that captures whispered thoughts and transcribes voice into organized notes, starting at $249 with summer 2026 delivery. The device uses ElevenLabs technology to respond in a synthesized version of the wearer's own voice.
Tools to Try
S. No | Tool | What it does |
|---|---|---|
1. | Text-to-Video | Plan scenes, nudge motion, and branch stories inside one workspace. |
2. | Agent builder | connects to hundreds of apps; chain tasks from chat (send emails, move files, update CRM) |
3. | Property ops AI | auto-routes maintenance, handles tenant calls, and summarizes tickets for managers. |
4. | No Code AI App Builder | Google Labs expansion post (160+ countries); third-party coverage |
5. | YouTube of Apps | Create mini apps with simple prompts, featuring social remixing capabilities |
Need more analytics and comparisions on the top rated tools for your workflows? Email us at [email protected], and we will send you the updated reports.
Breakthrough Tools
Kosmos: The AI That Does 6 Months of Science in a Day
Edison Scientific launched Kosmos this week, an autonomous AI scientist that can read 1,500 papers, execute 42,000 lines of code, and produce comprehensive research reports in a single 12-hour run.
Independent scientists found 79.4% of statements in Kosmos reports to be accurate, and collaborators reported that a single Kosmos run performed the equivalent of 6 months of their own research time.
Key Discoveries:
Kosmos has made seven validated scientific discoveries spanning metabolomics, materials science, neuroscience, and statistical genetics
Three discoveries independently reproduced findings from unpublished manuscripts, while four made novel contributions to scientific literature
Every conclusion can be traced back to specific lines of code or passages in scientific literature, ensuring full auditability.Availability: Currently priced at $200 per run (200 credits), with free tier access for academics and founding subscriptions that lock in the $1/credit pric
AI x Creativity Exclusive
Adobe MAX Sneaks (2025) — standouts included Project Light Touch (relight after capture), Turn Style (treat 2D like 3D), and Sound Stager (auto layered soundscapes). These are R&D previews but signal where Adobe is going.
Wabi $20M Funding (this week) - Replika founder's "YouTube of apps" with pre-seed round from Naval Ravikant, Garry Tan, etc is a social, prompt-to-app platform (“YouTube of apps”) with remixable mini-apps. (Source)
OpenAI Aardvark (Oct 30, 2025) - GPT-5 security agent with 92% detection rate that continuously scans repos, reasons about exploitability, and drafts patches; OpenAI reports ~92% detection on internal benchmarks and multiple CVEs issued from finds. (Source)
Canva Creative Operating System (November 3) - Canva unveiled a Design Foundational Model and a rebuilt Video 2.0 editor, plus “Ask @Canva” assistance; the company also highlighted large enterprise adoption.(Source)
Apple-Google Gemini Deal ($1B/year) - Custom 1.2T parameter Gemini for Siri (Spring 2026 launch), Apple is reportedly licensing a custom Gemini model (~$1B/yr) to power a major Siri overhaul while it builds its own large model; rollout as early as spring.
Two quick plays
Creators: turn one idea into a system—thumbnail factory + 2-variant A/B, then reuse weekly.
Founders: wire one agent into a workflow (not a demo): intake → action → log → human check. Measure minutes saved or mistakes avoided.
Not to Miss Events
Luma | Agent World Tour — London — demos, talks, and networking around agentic video + creation. RSVP
Bonus must-reads to prep before you go
Fireside Chat
AI PUN
Humor not Parsed: 100 Pizzas Later
Cahn’s POV
The Coca-Cola controversy crystallizes the central tension in AI × Creativity: efficiency vs. humanity. When a five-person team can generate 70,000 video clips in 30 days, the economics are undeniable. But when audiences call the result "soulless," we're reminded that creativity isn't just about output — it's about connection, emotion, and shared human experience.
As tools like Kosmos, Aardvark, and Wabi democratize capabilities once reserved for experts, we're moving toward a world where creation is faster, more accessible, and increasingly automated. The question isn't whether AI will transform creative work — it already has. The question is what we choose to preserve in the process.
That's it for today!
Before you go we’d love to know what you thought of today's newsletter to help us improve Cahn’s AI Canvas experience for you.
⭐️⭐️⭐️⭐️⭐️ Nailed it
⭐️⭐️⭐️ Average
⭐️ Fail
See you Next Week,
Aditi and Swati — the humans behind Cahn’s AI Canvas.
📩This week felt like AI stopped asking for permission and started filing expense reports. Agents got gutsier, platforms picked sides, and chat stopped being “just chat.”
If you're a creator, dev, or artist trying to figure out where you fit—you're not alone. This isn't doom and gloom. It's about knowing what's real, what's hype, and where the actual opportunities are.
Want our setup checklist for local AI agents + our best prompt pack? Reply "Loop."
Stay Creative. Stay Updated. Get in Touch : [email protected]
Edition #27 covered Oct 31-Nov 7, 2025. All news verified from mainstream sources with direct article links provided.
Disclaimer: The information presented in this newsletter is curated from public sources on the internet. All content is for informational purposes only.