Telex brings vibe-coding to WordPress sites 

Hey AI Enthusiast,

WordPress's Telex is already powering live production sites.

Only months after its September debut.

Price comparisons. Cost calculators. Real-time store information. All built with AI. All shipping.

"Things that used to cost tens of thousands of dollars to build, we're now able to do in a browser for pennies," said Matt Mullenweg at this week's State of the Word event.

But here's what makes this different: the person building these live blocks can't write code.

"I can't write a single line of code, but I can describe what I want to Telex, and it can make it for me," said Nick Hamze.

This is the shift. Non-developers building production software.

But first, today's prompt and the AI feature most people aren't using (then more on Telex...)

🔥 Prompt of the Day 🔥

Smart Social Listening AI Assistant

Act as a social listening specialist. Create one monitoring system that tracks brand conversations and identifies engagement opportunities.

Essential Details:

• Brand Keywords: [TERMS TO MONITOR]

• Platform Scope: [SOCIAL NETWORKS]

• Mention Volume: [EXPECTED DAILY]

• Response Priority: [WHAT NEEDS REPLIES]

• Sentiment Focus: [POSITIVE/NEGATIVE/QUESTIONS]

• Team Capacity: [RESPONSE BANDWIDTH]

Create one listening system including:

1. Keyword and hashtag tracking setup

2. ChatGPT sentiment classification prompts

3. Opportunity scoring algorithm

4. Response suggestion generator

5. Escalation trigger identification

6. Performance analytics dashboard

Never miss a conversation with AI.

✅ Tips & Tricks Thursday ✅ 

Stop Wasting Time Re-Explaining Everything to AI

Most people treat every AI conversation like starting from scratch.

They re-explain their role. Their industry. Their preferences. Their brand voice. Every single time.

This is killing your productivity.

ChatGPT and Claude now have memory features. They remember context across conversations.

But almost nobody uses them correctly.

What AI Memory Actually Does

ChatGPT's memory feature saves key information from your conversations. Your role, preferences, common project details, how you like things formatted.

Claude has a similar system called "custom instructions" and conversation history awareness.

Instead of typing "I'm a marketing consultant who works with SaaS companies" every time, you tell it once. It remembers.

Instead of explaining your brand voice in every prompt, you set it once. It applies it automatically.

Instead of re-describing your current projects, you reference past conversations. "Use the framework we discussed last week."

How to Set Up AI Memory Correctly

Step 1: Define Your Core Context Once

Open ChatGPT or Claude. Start a conversation specifically to set your context:

"Remember these details about me:

• I'm a [YOUR ROLE] working in [YOUR INDUSTRY]

• My target audience is [SPECIFIC AUDIENCE]

• My brand voice is [DESCRIBE TONE]

• I prefer responses that are [LENGTH/STYLE]

• Common projects include [LIST TYPES]"

ChatGPT will save this to memory. Claude will reference it in future conversations.

Step 2: Add Project-Specific Context as You Go

As you work on projects, let AI build context:

"I'm launching a new product called [NAME]. It's [DESCRIPTION]. Target market is [AUDIENCE]. Key differentiators are [POINTS]."

Now in future conversations, you can just say "Update the landing page copy for [PRODUCT NAME]" and it knows what you're talking about.

Step 3: Reference Past Conversations Instead of Re-Explaining

Stop retyping the same information.

Instead of: "I need help with my email sequence for my SaaS product that helps marketing teams automate reporting..."

Say: "Improve the email sequence we worked on last Tuesday."

AI pulls the context. Saves you 30 seconds every time.

Step 4: Update Memory When Things Change

Your needs evolve. Your projects shift. Your preferences update.

Tell AI explicitly:

"Update my memory: I no longer work with B2C clients. Focus only on B2B examples going forward."

"Remember: I prefer all code examples in Python now, not JavaScript."

Don't let stale context slow you down.

Step 5: Build Multiple Specialized Assistants

Don't use one conversation for everything.

Create separate threads for different contexts:

• One for content creation

• One for technical documentation

• One for client strategy

• One for internal planning

Each maintains its own specialized memory and context.

The Compound Effect

This isn't about saving 30 seconds per prompt.

It's about AI getting smarter about you over time.

Week 1: AI knows your basic role and preferences.

Week 4: AI knows your current projects, your writing style, your common workflows.

Week 12: AI anticipates what you need. Suggests improvements in your voice. References past work without prompting.

Context continuity multiplies your productivity exponentially.

What Most People Get Wrong

They use AI memory passively. They hope it picks up context automatically.

It does pick up some things. But not strategically.

Active memory management means:

• Explicitly telling AI what to remember

• Updating it when things change

• Referencing past context deliberately

• Building specialized conversation threads

Treat your AI like a new team member. You wouldn't expect them to figure everything out by osmosis. You onboard them properly.

The Real Shift

AI memory turns one-off interactions into continuous collaboration.

You're not using a tool anymore. You're building a personalized assistant that gets better the longer you work together.

Most people reset this relationship every time they open a new chat.

That's the waste.

Did You Know?

Hotels use AI that adjusts room pricing every few minutes based on factors like local weather changes, your browsing history, and even the battery level of your phone.

🗞️ Breaking AI News 🗞️

 WordPress Telex: The Full Story

WordPress announced Telex in September as an experiment.

By December, it's powering production sites.

At the State of the Word event in San Francisco, Matt Mullenweg showed examples built by Nick Hamze:

Pricing comparison tool - Rich, interactive element that used to require custom development.

Real-time store information block - Pulls current hours, phone number, directions link into the site header.

Partner logo carousel - Rotating display of business partners.

Custom pricing calculator - Interactive tool for quotes or estimates.

Google Calendar integration - Embeds live calendar data.

Equal-height post grid - Homepage layout where all post cards match height automatically.

These all shipped. Live. In production.

The Person Building Them Isn't a Developer

Nick Hamze: "The thing that blows my mind and should blow yours is I'm not a developer. I can't write a single line of code, but I can describe what I want to Telex, and it can make it for me."

He continued: "I think as long as people think of these tools as 'developer' tools, they are missing the point on what they can really accomplish, which is letting regular folks do things they never could have done before."

Another creator, Tammie Lister, built a new Gutenberg block every day in October. Playable ASCII Tetris. Halloween trick-or-treat block. Seasonal UI elements.

Not for a demo. For practice. For fun. Because she could.

What Telex Actually Does

Telex translates natural language into functional Gutenberg blocks.

Gutenberg blocks are the modular components—text, images, columns, interactive elements—that make up WordPress sites.

Instead of writing React and PHP code, you describe what you want. Telex generates the block code, configuration, and styles. You iterate in the browser.

This is vibe-coding: define software by describing the result, not every step to get there.

The Foundation WordPress Built

Telex works because of architectural changes WordPress made to support AI:

Abilities API - Defines what WordPress can do in machine-readable language so AI systems can understand it.

MCP Adapter - Exposes those abilities to any MCP-compatible tool. WordPress installations can now participate in AI workflows through Claude, Copilot, and other platforms without custom integrations every time.

"This adapter pattern means WordPress can participate in AI workflows without duplicating logic or creating separate integrations for every AI platform," Mullenweg explained.

Developers are already using AI in daily workflows through Cursor, Claude Code, and next-gen CLIs. They refactor projects, search codebases, automate tasks, run WP-CLI scripts alongside AI agents.

Then they use Telex when it's time to add a new block.

What's Coming in 2026

WordPress will introduce benchmarks and evaluations for AI models to test on WordPress-specific tasks.

Changing plugins. Editing content. Manipulating the WordPress interface using browser agents.

This creates scoring for models on real CMS work, not just generic coding tests.

Why This Matters Beyond WordPress

WordPress powers 43% of all websites.

Custom development for sites has always been expensive and time-consuming. Agencies charge thousands for interactive elements. Development cycles stretch weeks or months.

Telex collapses that timeline.

Small businesses can launch pricing tools, appointment widgets, location modules—elements that directly affect conversion—in hours instead of weeks.

Enterprise teams can ship dozens of brand-aligned blocks across properties without bottlenecking designers and engineers.

But the bigger shift is accessibility.

Nick Hamze can't code. Now he's shipping production blocks.

That's not "making developers more productive." That's giving non-developers developer capabilities.

The barrier to building custom web functionality just dropped to near-zero for anyone who can describe what they want.

That's the real story.

Over to You...

Have you tried any AI coding tools yet, or still building everything manually?

Let me know what's working for you.

To smarter development workflows,

Jeff J. Hunter 
Founder, AI Persona Method | TheTip.ai

NEW: “AI Money Group” to Learn to Make Money with AI