Welcome, humans.

“My boss tells me he needs agents, and I just keep building Make workflow automations and I tell him these are agents, and he's very happy.”

That's an actual quote from a customer that Darin Patterson, VP of Market Strategy at Make, shared with us. And honestly? It might be the most relatable thing we've heard all year.

Everyone's rushing to deploy AI agents right now. But here's the dirty secret: most companies are setting themselves up for failure because they're skipping the automation foundation entirely.

This tracks with a quote from Eric Schmidt that Maithra Raghu just shared:

She says 2026 will be different. True agents, where the AI decides how to plan, leverage tools, and act on intermediate output, are becoming real (hence the Claude Code hype you keep seeing).

But that doesn't mean workflows are dead. It means knowing when to use which one is the new skill.

In our latest podcast episode, Darren breaks down when you actually need an AI agent vs. when a good old-fashioned workflow will do the job better, faster, and cheaper:

Here’s what stood out to us most:

• (2:58) What actually counts as an “AI agent”, and why most tools calling themselves agents... aren't.

• (5:39) When AI agents genuinely outperform traditional workflows; hint: it's about rules that change, not complexity.

• (10:10) “I met a customer whose boss demanded agents...”: start with the problem, not the buzzword, and solve accordingly.

• (20:20) Demo: Mixing deterministic logic with AI , a.k.a the “scalpels not sledgehammers” approach.

• (23:14) Why Make might be the best MCP server you're not using, and how it's more token-efficient than connecting raw APIs (more on that below).

• (25:20) The great irony of AI context: “the more you throw at it, the less successful it is.”

• (39:10) Make Grid: See your entire automation landscape in real-time , a visual map of every workflow, every dependency, every data flow.

• (55:56) A heartwarming use case: how a pet adoption center uses AI to fuzzy-match names on a "do not adopt" list.

In addition to being a fun interview, you’ll also get an overview of how to use Make.com with some live demos that simplify the more complex, node-based workflows, fields you need to fill in, and how to think about using connectors (MCP servers that connect to other tools), and even get a birds-eye-view of all your automations.

More tools need that last capability. Imagine if you could see all your Claude Code automations like that? You can certainly try to code yourself a similar UI! We like Cursor’s example of their own coding agents building an entire browser from scratch. Now imagine that, but real time.

Bottom line: The companies winning with AI aren't the ones deploying the most agents, they're the ones who can see their entire automation stack and know exactly when AI adds value vs. when it just adds risk.

Listen now on YouTube | Spotify | Apple Podcasts

Keep scrolling for details about our upcoming live stream this Thursday, a few videos from the Archive we think you’ll love, and a video to show you how to take your newfound Make.com knowledge and combine it with Claude Code to automate your entire life (whoa dude).

🔴 THIS THURSDAY: Is AI Actually Working?

🔴 Join us LIVE on Thursday, Jan 22 2026 @ 12pm PST | 3pm ET | 8pm GMT:

In this live episode, we’re hanging out down with Dennis Salguero, a.k.a Data Science With Dennis, to cut through the AI hype and talk about what actually matters.

We’ll get Dennis’ hot takes on what the best model is for data science tasks, how to tell if you're ready for AI adoption, and what realistic AI implementation looks like in 2026.

Plus, we’ll do a LIVE comparison of three AI coding tools for data science work: Claude Code, Claude Cowork, and OpenAI’s Codex (Dennis's personal favorite). Watch us test them side-by-side to see which one actually delivers when you're working with real data. If you work in datascience and need help getting started with AI, this ep’s for you.

Click here to get notified when we go live!

🔧 Bonus: Connect Make to Claude Code

After you get an intro to Make, check out Nate Herk's viral tutorial on building apps with Claude Code + n8n. N8N?! You might ask. But you just told me all about Make!  

Here's the good news: you can do the exact same thing with Make.

Make now has an official MCP server, which means Claude Code can:

• See and audit your Make scenarios.

• Suggest optimizations (like swapping Watch triggers for Webhooks).

• Build frontend apps that hit your Make webhooks.

The setup is nearly identical:

1. Generate an MCP Token in your Make.com settings.

2. Connect it to Claude Code following these docs (throw them into Claude and ask it to help you).

3. Start building!

One difference: Make doesn't have a “Skills” repository yet (like n8n does), so Claude might be slightly less fluent at building complex Make modules from scratch. The workaround? Create a simple SKILL.md file with your own Make best practices.

This is probably safer than trying to spin up your own code to make automation workflows right out of the box with Claude Code. For that, you’ll want to watch this (for an overview of Claude Code automations), then this (showing how to replace an agentic automation with Claude Code), and if you want to go deeper in coding with Claude, then this (its 80+ minutes, but its very in depth!).

Dive deeper with these resources:

• Make Academy — hands-on training to get started.

• Make Playbook — for executives thinking about AI transformation strategy.

• OpenAI's guide on when to use agents — the paper Darin references in the episode.

Stay curious,

The Neuron Team

P.S. Darin revealed that Make is working on “Maia by Make”, an in-product AI that helps you build automations through natural language conversation, not one-shot prompts. It’ll ask clarifying questions and build alongside you. Y’know, so you can drink pina coladas on the beach like the future Darin wants. We'll cover it when it launches. 🚀

IN CASE YOU MISSED IT…

Check out our other recent favorite episodes below!

David Cox: The Case Against Friendly Robots: IBM's VP of AI Foundations argues the industry has it backwards — we should treat AI like processors, not personalities. Granite 4.0 is their answer: hybrid architectures that cut memory 10x, models that run on phones at 90 tokens/second, and the first open model with external security certification. (YouTube | Spotify | Apple)

Carina Hong: The 24-Year-Old Building an AI Mathematician: She dropped out of Stanford's PhD program to build “mathematical superintelligence” and just raised $64M. Axiom has already solved a 130-year-old problem and disproved a 30-year-old conjecture. (YouTube | Spotify | Apple)

Zuzanna Stamirowska: The Post-Transformer Architecture: The CEO of Pathway breaks down BDH (Baby Dragon Hatchling), a brain-inspired architecture that introduces true temporal reasoning and continual learning. Backed by Lukasz Kaiser, co-inventor of the Transformer itself. (YouTube | Spotify | Apple)

And if you haven’t subscribed yet, please do! Click the image below to go to our channel and hit “subscribe” to get notified right when new videos go live.

We have a goal to hit 50K subscribers by the end of the year (if not 100K), and we’re only ~37K away! If you like learning about AI, and already watch some of our videos, do us a favor and click here to subscribe today.

Stay curious,

The Neuron Team

P.P.S: Love the newsletter, but don’t want to receive these podcast announcement emails? Don’t unsubscribe, adjust your preferences to opt out of them here instead.