why taste matters more than ever
Plus: ChatGPT checkout converted 3x worse for Walmart, 81,000 AI work interviews analyzed, and Merriam-Webster and Encyclopedia Britannica sue OpenAI.
what you need to know this week in ai + marketing in 15 seconds:
Is AI ruining our taste? - as AI makes polished work easy, more things look the same. The shift is from execution to taste. Tools are widely available, choices aren’t (NYT)
The future of content belongs to the tastemakers - when content is cheap and endless, judgment becomes the scarce resource. Brands that know what to say, not just how to say it (Contently)
AI can optimize marketing, but can it build belonging? - efficiency doesn’t automatically create cultural relevance. Optimization is great until it shrinks your world (Newsweek)
Walmart says ChatGPT checkout converted 3x worse than its own website - people may browse with AI, but prefer familiar environments to complete purchases. Discovery is moving, trust is slower (MarTech)
81,000 interviews show how people are using AI at work - Anthropic analyzed real conversations. Most usage clusters around writing, analysis, and support tasks rather than full automation. Reality is still “assist” not “replace” (Anthropic)
AI commoditizes marketing execution and elevates judgment - execution getting cheaper shifts value toward decision-making, prioritization, interpretation. Easy parts automated, hard parts becoming the job (MarTech)
Entry-level marketing work is reshaping - roles moving away from drafting toward editing, QA, managing AI-assisted workflows. Junior roles not gone, just less forgiving (MarTech)
Merriam-Webster and Encyclopedia Britannica sue OpenAI - publishers suing over how content is used to train models. Legal battle over data ownership nowhere near settled (TechCrunch)
🗞️ top ai + marketing news
taste, creativity, and what actually matters now
Is AI ruining our taste?
The New York Times | ~6-min read
As AI makes it easier to generate polished work quickly, more things start to look the same. The piece argues the real shift is from execution to taste. The tools are widely available. The choices aren’t.
My take: The work is easier but deciding what’s worth making is harder.
The Future of Content Belongs to the Tastemakers
Contently | ~5-min read
When content is cheap and endless, judgment becomes the scarce resource. Brands that stand out will be the ones that know what to say, not just how to say it.
My take: This is quietly becoming the whole job.
The human skill AI still struggles with in writing
The Atlantic | ~7-min read
AI can replicate structure and tone, but still struggles with originality, restraint, and knowing when something actually lands. The gap shows up most clearly in creative work.
My take: You can generate ten versions. You still need someone to pick one.
AI can optimize marketing. But can it build belonging?
Newsweek | ~4-min read
AI makes it easier to target and optimize for known audiences, but that can narrow reach and limit long-term brand growth. Efficiency doesn’t automatically create cultural relevance.
My take: Optimization is great until it shrinks your world.
commerce, agents, and where the money moves
Shopify is preparing for AI shopping agents to change everything, exec says
TechCrunch | ~4-min read
Shopify is leaning into a future where AI agents handle discovery and product recommendations. That shifts how consumers find products and how brands compete for attention.
My take: If the agent is picking, your brand has to give it a reason to.
Walmart says ChatGPT checkout converted 3x worse than its own website
MarTech | ~4-min read
Walmart found that checkout inside ChatGPT significantly underperformed its own site. People may browse with AI, but they still prefer familiar environments to complete purchases.
My take: Discovery is moving. Trust is slower.
the stack, org design, and what work actually looks like now
AI commoditizes marketing execution and elevates judgment
MarTech | ~5-min read
Execution is getting cheaper and faster, which shifts value toward decision-making, prioritization, and interpretation. The work doesn’t disappear, but the center of gravity moves.
My take: The easy parts are getting automated. The hard parts are becoming the job.
AI is reshaping what entry-level marketing work looks like
MarTech | ~5-min read
Entry-level roles are moving away from drafting and repetitive tasks toward editing, QA, and managing AI-assisted workflows. The ladder still exists, but the first rung looks different.
My take: Junior roles are not gone. They’re just less forgiving.
Scaling AI adoption requires more than tools
Google | ~6-min read
Google focuses on the operational side of AI adoption: leadership alignment, training, and integrating AI into actual workflows. Access to tools is not the bottleneck anymore.
My take: Most teams are not blocked by tech. They’re blocked by how they work.
agencies, economics, and industry pressure
AI is squeezing marketing agencies from both sides
Search Engine Land | ~5-min read
Agencies are getting pressure from clients expecting efficiency and platforms making work more self-serve. That pushes agencies toward higher-level strategy or forces them to justify their existence.
My take: Execution is harder to sell when everyone has the same tools.
platforms, products, and the model race
OpenAI’s side projects show how fast the ecosystem is expanding
The Wall Street Journal | ~5-min read
OpenAI’s internal experiments and side projects highlight how quickly new use cases are being explored, often outside the core product. The surface area keeps expanding.
My take: The roadmap is starting to look like a scatter plot…
OpenClaw moment sparks concern AI models are becoming commodities
CNBC | ~4-min read
As models become easier to replicate or compete with, differentiation shifts away from raw capability toward distribution, ecosystem, and product experience.
My take: When the model becomes table stakes, everything else matters more.
legal, data, and ownership
Merriam-Webster and Encyclopedia Britannica sue OpenAI
TechCrunch | ~4-min read
Publishers are suing OpenAI over how their content is used to train models, continuing the broader legal battle over data ownership and compensation.
My take: Oy…this fight is nowhere near settled.
research, insight, and how people actually use AI
81,000 interviews show how people are using AI at work
Anthropic | ~6-min read
Anthropic analyzed tens of thousands of real conversations to understand how AI is being used across roles. Most usage clusters around writing, analysis, and support tasks rather than full automation.
My take: The reality is still much more “assist” than “replace.”
👩💻 thought-leader highlights
Emily Kramer – “Build marketing strategy skill in Claude Code”
Emily warns that teams who jump straight to execution with Claude Code miss out on developing the marketing strategy underpinning their work. She urges marketers to build and document strategy inside Claude so it becomes a working guide in everyday tasks.
Andrew Yeung – “Taste is the most important skill in tech”
Andrew writes that as AI lowers production costs, taste—an eye for what to create—becomes the real competitive edge. He shares six ways to hone it: keep a “taste file,” cross‑pollinate ideas, learn from making bad work, invest in learning, edit ruthlessly and develop a coherent style.
Cecilia Stallsmith – “Why campaigns fail: it’s not the idea, it’s the ship date”
Cecilia invites us all to a virtual vibe marketing webinar on April 1. She also writes that most campaigns flounder because teams never ship them, not because they’re bad ideas.
Jonathan Martinez – “Your marketing stack should live in GitHub”
Jonathan believes marketing teams should organize their AI skills, docs and workflows in GitHub so Claude Code always has context. He sees the rise of a “growth engineer” mindset where marketing adopts engineering discipline.
Kristen van Laren – “How our CEO’s post hit 300 K impressions with Claude”
Kristen explains that by feeding Claude a dataset of top‑performing posts, analyzing patterns and mining comments for ideas, her team built a content formula that helped their CEO craft a 300‑K‑impression post.
Jason Widup – “Boost marketing efficiency with AI: Claude, Clay and ICP analysis”
Jason uses Claude to automate tasks like ad‑copy and creative‑brief generation and built a marketing library from product docs and transcripts.
🛠️ latest ai + marketing tools
Perplexity Personal Computer Always-on AI running on a dedicated Mac mini that merges your local files and apps with Perplexity Computer to work as your 24/7 digital proxy. $200/month for Max subscribers, with kill switch and full audit trail.
Perplexity Health Connects Apple Health, wearables, and health records from 1.7 million providers to give you answers grounded in your actual medical history — not SEO content. Rolling out to Pro and Max subscribers in the US.
OpenAI GPT-5.4 Mini and Nano Smaller, faster versions of GPT-5.4 for high-volume workloads. Mini approaches flagship performance at 2x speed; Nano costs $0.20 per million input tokens for classification and lightweight coding tasks.
Microsoft MAI-Image-2 Microsoft’s in-house image model lands at #3 on Arena.ai behind Google and OpenAI, with standout photorealism and readable text in images. Rolling out to Copilot and Bing Image Creator.
Google AI Studio Apps Full-stack vibe coding upgrade — the Antigravity agent understands entire project structures, provisions Firebase backends, and deploys production apps from prompts. Free tier: 60 requests/minute, 1M tokens/day.
Google Stitch AI-native design canvas that turns voice and text into high-fidelity UI prototypes in minutes. Exports to Figma or HTML/CSS — Figma shares dropped 8% on the announcement.
Google Personal Intelligence Expansion Now free for all US users across AI Mode, Gemini app, and Chrome. Connects Gmail and Photos to personalize search results — shopping recs from past purchases, travel plans from hotel confirmations.
Gemini Workspace Updates Gemini now drafts docs from your files and emails, builds spreadsheets from plain English, and answers questions across your entire Drive. “Fill with Gemini” predicts data from existing cells or live search.
Cursor Composer 2 In-house coding model fine-tuned on Kimi K2.5 with 200K context, beating Claude Opus 4.6 on Terminal-Bench while costing 86% less than Composer 1.5.
NVIDIA NemoClaw Open-source stack for running OpenClaw agents securely, announced at GTC with the new OpenShell runtime. Hardware-agnostic, enterprise-focused — early preview, not production-ready.
Manus My Computer Meta’s Manus AI goes from cloud-only to a desktop app for Mac and Windows that works directly with local files and terminals. Paid plans from $20/month.
Okara AI CMO Enter your URL and it deploys agents for SEO, GEO, content, Reddit, Hacker News, and X. Priced at $99/month — the launch video hit 8.6 million views.
Accenture Invests in DaVinci Commerce Strategic investment in the agentic commerce platform powering branded shopping experiences inside ChatGPT and other LLMs. Signals enterprise focus on AI-native storefronts.
Ahrefs MCP Server Remote MCP server that works directly in Claude, ChatGPT, and Cursor — no local setup needed. Pull SEO metrics and competitor data through natural language prompts.
Picsart AI Agent Marketplace Hire AI assistants for resizing, remixing, and Shopify product editing. Four launch agents accessible via WhatsApp and Telegram with adjustable autonomy levels.
Alibaba Wukong Enterprise platform coordinating multiple agents for document editing, spreadsheets, approvals, and transcription. Built on a full DingTalk rewrite, with Slack and Teams integrations planned.
Snowflake Project SnowWork Turn conversational prompts into finished workflows — board decks, churn spreadsheets, supply chain analysis. Executes end-to-end on governed Snowflake data, now in research preview.
Mistral Forge Enterprise platform for building custom AI models from scratch on proprietary data — not just fine-tuning. Early adopters include ASML, Ericsson, and the European Space Agency.
💼 cool ai + marketing jobs
Siro - Content Marketing
Ideogram - Director of Growth
Stripe - Integrated Campaign Manager, Startups & AI
Lovable - Senior Solutions Marketer
Listen Labs - Growth Lead, Acquisition
AirOps - Head of Product Marketing
Anthropic - Head of Product Communications
Reply anytime and let me know:
Topics you’d like covered in future weeks
Sections you love or could do without
Questions about AI, marketing, or anything else
Love y’all,
Carley