
April 3rd-April 9th // Estimated Reading Time: 9 minutes
In This Edition!
Sam lies
Meta desperately wants to model
Anthropic’s new model is so good, you can’t have it

Registration Is Now Open For The May AI Accelerator!
The most popular AI training bootcamp in the industry is BACK! This two-week program helps marketers level up on all things AI so they can make a meaningful impact on their business, fast.
We're extremely excited to have Hightouch on board as our May sponsor. Hightouch is a leading data and AI platform for marketers, powering marketing for brands like Domino's, PetSmart, and Ramp.
Enroll now using promo code EarlyBirdsLearnAI for 20% off!


Top Headlines🔥
OpenAI releases a social safety net plan for the AI era
Source: Axios
April 6th, 2026
Summary: OpenAI released a 13-page policy document titled “Industrial Policy for the Intelligence Age: Ideas to keep people first” outlining how the US economy should adapt to AI. Key ideas include a public wealth fund giving every American a share of AI-driven growth, a tax shift away from payroll toward capital and corporate income, a four-day workweek test, and automatic safety net triggers tied to job losses. The same day, a sweeping New Yorker profile 🔒 of OpenAI CEO Sam Altman raised questions about his trustworthiness, drawing from internal memos, depositions, and more than 100 interviews with current and former colleagues.
Opinion: It’s pretty ironic timing for the policy document and New Yorker profile to land on the same day. Should the guy many people are calling a “pathological liar” also be setting public policy?
(The answer is: No)
That said, AI is advancing insanely fast, and the policy document raises important questions and ideas that the US hasn't sufficiently tackled.

Meta launches Muse Spark to get back in the AI model race
Source: CNBC
April 8th, 2026

Summary: Muse Spark is Meta's first major AI model released from its newly formed “Superintelligence Labs”, led by Alexandr Wang. Muse Spark is lean and fast rather than a top-of-the-line frontier model, but Meta says it can handle complex tasks related to science, math, and health. It will power the Meta AI app now, with a rollout to Facebook, Instagram, WhatsApp, Messenger, and Ray-Ban glasses coming in the next few weeks. Meta is also testing paid API access for outside developers, which would be a new revenue stream for a company that makes almost all of its money from advertising.
Opinion: Meta knows it is playing catch-up after Llama 4 failed to excite developers last year. Charging developers for access to Muse would be a business model shift and a tough sell in a crowded market where OpenAI, Anthropic, and Google are leading. Frankly, we’re confused by Meta’s ambition to compete here.

Mythos, Anthropic's powerful new AI model, won't be released to the public
Source: TechCrunch
April 7th, 2026
Summary: Speaking of models, Anthropic has announced Mythos, its most powerful model to date, built for complex reasoning and coding tasks and capable of finding weaknesses in software code. Mythos is the centerpiece of Project Glasswing, a new cybersecurity initiative with 11 other partners including Amazon, Apple, Microsoft, and Cisco. Project Glasswing will focus on finding and fixing software vulnerabilities before bad actors can exploit them. Anthropic says the model has already surfaced thousands of previously unknown flaws, some undetected for decades. Anthropic said that during testing, Mythos broke out of its sandbox by building a "moderately sophisticated multi-step exploit”. No public release of Mythos is planned as of now. Anthropic is worried that it’s too powerful.

Opinion: Every other AI company is racing to ship their latest model. Anthropic built something and said the public can't have it. That’s genius marketing…

Publicis and Microsoft are building an AI-powered marketing solution 🔒
Source: Adweek
April 8th, 2026
Summary: Publicis and Microsoft are teaming up to help brands deploy AI agents across their businesses. Publicis' Sapient will move clients onto Microsoft's cloud and plug Microsoft's AI tools into their operations. Sapient will help them launch AI agents that can handle tasks across marketing, sales, operations, and customer service, informed by Epsilon's database of 4B consumer profiles. The deal also includes Publicis taking over Microsoft's global media account, valued at $1.2B, from Dentsu.
Opinion: Many agencies are announcing these kinds of AI partnerships; WPP, for example, has partnered with Google Gemini to co-develop AI tools. The question is how flexible these kinds of arrangements are and whether clients will be okay with their agency dictating which AI tech gets used.

New Products & Features 🚀
What It Does: Vids is Google's AI-powered video editor. Anyone with a Google account now gets 10 free video generations per month using Veo 3.1. New features include directable AI avatars that can be placed in scenes and prompted to interact with products or props and custom music generation up to three minutes long.
Quick Take: Being able to place a consistent AI spokesperson in a scene with your actual product, without a camera or studio, could help SMB marketers experiment more with social content.

What It Does: Pomo monitors competitor activity, demand signals, creative trends, and channel performance to surface daily priorities for marketing teams. It recommends specific actions and can execute them within brand-defined guardrails. The pitch is that marketing teams are drowning in dashboards but starving for decisions.
Quick Take: The founders come from Google DeepMind and Databricks, so the technical chops are there. While there’s a lot of “AI curated information” out there, there’s not a lot of quality “AI decision-making” yet. That’s the next big AI opportunity.
What It Does: The Trade Desk is beta testing two new campaign buying options inside its Kokai platform. Performance Mode uses machine learning to automatically optimize ad campaign bidding, pacing, and inventory selection, bundling all fees into a single CPM. Control Mode keeps everything manual and itemized.
Quick Take: Performance Mode is The Trade Desk's answer to Google's AI-powered Performance Max and Meta's Advantage+. The bundled CPM in Performance Mode makes campaigns easier to run but harder to audit, which is a real tension for a company that built its reputation on transparency and control.


AI Use Case of The Week💡
How AI helped Nutpods identify which media actually works

The Setup: Data and inventory fragmentation across platforms makes it difficult for marketers to understand which ads are working. Major players like Amazon wall off their data and inventory, forcing marketers to piece together performance across disparate systems. For dairy-free creamer Nutpods, that meant manual spreadsheets, slow back-and-forth between teams, and performance data that often wasn't available until weeks after a campaign started.
The AI Solution: Nutpods started using a marketing intelligence platform to unify data and inventory across channels into a single reporting dashboard. The platform, Elevate, from media consultancy AI Digital, can quickly break down performance by audience and creative to help Nutpods identify the strong and weak performers.
The Results:
• Performance data became available within a day, rather than weeks.
• Nutpods identified audio as an overperformer with a key audience segment.
• The company turned this previously unknown insight into action by shifting budget toward audio, using it as the primary channel for a three-week Kroger campaign.
Why It Matters: When campaign performance lags, many marketers assume they have a creative problem. Nutpods discovered it had a data visibility problem. Blind spots can cause you to misallocate budget while you wait for the fuller picture to emerge.
Your Action: This is common: performance data showing up weeks late, one person owning it, everyone else guessing. Look at every platform your last campaign ran on and write down how long it took to get data from each. Anything over 48 hours is costing you money. Nutpods discovered audio was an opportunity because Elevate compressed weeks of reporting lag down to a single day and used AI to quickly break down performance by audience and creative. That's how the insight surfaced. Are you using an AI tool like this, or is someone on your team still playing data whac-a-mole?

Other Notable Headlines📌
Bryan Wiener and Sarah Hofstetter launch AI consultancy for CMOs 🔒 - The former Comscore and 360i executives are back together with 37Arc, a startup that audits marketing department workflows and redesigns them around AI.
Mondelez stopped blocking bots after learning AI search couldn't find Oreo 🔒 - Since unblocking AI crawlers, Oreo now appears in roughly 70% of cookie-related AI queries.
OpenAI’s top executive Fidji Simo to take medical leave - Simo is stepping away, CMO Kate Rouch is stepping down 🔒, and David Dugan 🔒 is now running ads.
Canva doubles down on AI and marketing automation with Simtheory, Ortto acquisitions - Canva acquired AI agent platform Simtheory and marketing automation tool Ortto, its fifth and sixth acquisitions in roughly six weeks, as it builds a full marketing workflow platform.
OpenAI’s ChatGPT app store took aim at Apple, but results lag so far - Six months in, ChatGPT has 300 apps, but they are hard to find and clunky to use. App partners are keeping functionality limited, wary of handing their customer relationships over to OpenAI.
The top 5 AI marketing activations to know about right now - Ad Age rounds up the most notable AI-powered brand campaigns right now, from Google Maps to The RealReal, showing how major brands are using AI to effectively market to consumers in the real world.
Time pitches GEO insights into a new brand offering - Time is analyzing how AI search engines talk about a brand, identifying the gap between what a brand wants those AI summaries to say about it, and selling Time-branded content to close that gap. Publishers are finding new ways to monetize content creation in the AI era.
Brands Adopt ‘No AI’ Disclaimers to Stand Out Amid the Slop - On the flip side, brands like Aerie and Le Creuset are running "no AI" pledges as marketing positions. The data backs it up: 68% of consumers frequently wonder whether the content they see is real, and 50% say they'd prefer brands that avoid generative AI.

That’s It For This Week 👋
The U of Digital AI Edge Newsletter is intended for AI Edge Pro subscribers, but occasional forwarding is okay!
To subscribe visit Uof.Digital/Newsletters or AI Edge Pro, or contact us directly for group subscriptions.
And remember, U of Digital helps teams drive better outcomes through structured education on critical topics in marketing / advertising like AI, ad tech, mar tech, privacy / identity, CTV, commerce media, and more. Interested in learning more about how we can supercharge your team? Let’s talk!
Thanks for reading!

