THIS WEEK: THIS IS WHAT PEAK MARKETING LOOKS LIKE IN 2025

🌟 Editor's Note

Thanks for reading H4CKER, where I share my nerdiest experiments, in-the-trenches perspectives, and hot takes about the future of AI, marketing, and automation. We’re all learning when it comes to AI, so let’s trade follows on social media so we can learn from each other. Find me on X at @nathanabinford and / or let’s connect on LinkedIn.

😖 All The Best Marketers Are Now Coders

I’ve been a marketer for 20+ years and I’ve never been so bearish on the industry.

Marketing as a practice, and a skillset in-demand, isn’t going anywhere. But I’m souring on marketing as a career (and marketing agencies as a business model). The work has become meaningless and turned marketers into mindless button-pushers and unpaid sales reps for Google and Meta.

Small companies are too broke for the fun stuff, like brand-building, and they’re too content with the drip feed of Google Ads and Google Business Profile traffic to see the value of brand equity.

Agencies are getting squeezed on both ends while chasing diminishing profits and doing cheap, commodity work. It’s just not a fun, innovative space anymore.

If you crave newness, fresh ideas, and liberty to experiment (like I do) then building software is where all the good energy is flowing right now.

Ironically, some of the biggest problems to solve right now with software are marketing problems: how to reach more people, faster, and with more personalization, is a pressing problem for almost any business these days.

And that kind of problem is ripe for new software solutions. So it’s a great time to be a software engineer with a background in marketing.

Peak marketing in 2025 is less about content and campaigns and more about agents and automation.

We’re all going to end up in the same place sooner or later, but it will happen more gradually for some and may look quite different in different scenarios.

The common element is though, will be building with AI.

🤓 AI’s For Nerds

Nerds, meaning those of us with the technical proficiency, have an advantage in an agentic, AI-first world.

Pressing that advantage means doubling down on up-skilling in AI and related technology.

And it’s obvious to anyone paying attention that this generation of AI models is skewed towards coders / builders.

To squeeze the most value out of this point in the timeline means leveraging AI where it has the most monetizable juice. And these days, that’s definitely building software.

As marketers, we win when our brands (or our clients) win. Software fits into this dynamic in 3 ways:

• SaaS: Building marketing tools and tools for marketers

• Marketing as a Service: Productized services wrapping software automations

• Customer Interfaces: Sites, apps, tools and magnets, bots / chat bots

Simple automations: Automations for scheduling posts, gathering research, generating replies for reviews, and so on, are best built with n8n automations (like Zapier, but much better).

Complex workflows with custom logic: Proprietary business processes, on-demand integrations (e.g. CRM + legacy systems), or replacing human roles; these undertakings require automations with lots of customization and might as well be bespoke software.

I don’t think a total amateur with no software engineering experience can realistically vibe code a real platform for thousands of users.

But serving a few dozen (or a few hundred) simultaneous users is actually pretty feasible.

Plus, as I show in this week’s video (below), you can use agentic AI to help manage all the most technical aspects that you don’t already understand yourself.

To be fair, the tech is still a bit rough…you’ve got to be a pretty experimental soul to let agents run your computer and perform tasks you don’t completely understand.

But let’s be clear: Eventually, this is what we’re ALL going to to be doing: using AI to automate and scale our marketing activities as much as possible.

🍖 Where’s The Meat On The Bone?

The biggest value in AI right now is speed. Accuracy-wise…we’re far from perfect, but AI is really fast.

So focus your efforts on gaining speed through efficiency. That means building solutions that reduce inefficiencies, eliminate redundancies, scale output, and optimize time and effort per win.

How does custom software fit into this?

Here are just a few examples of the kind of projects currently on my own vibe coding Christmas list:

1. Formatting, images, and links automatically added to WP posts

2. Content repurposing pipeline (1-to-many)

3. Syncing and enriching contacts across systems

4. Analyzing customer data for marketing opportunities

5. Automating customer interactions

These activities are no longer worth paying people to perform; but they’re still needed (and not going anywhere).

So they’re ripe for automation, and building custom software / workflows give you the power to sculpt technology to support your business rather than forcing your business to fit the available tools.

Getting started on this journey means getting comfortable with platforms like Zapier, Make, and n8n, or learning how to vibe code (and consistently generate results).

Or, at least, it means partnering with a nerd who does… 😉

⚙️ The Laboratory: Prompts & Automations

Watch Cursor’s coding assistant agent deploy an app to a new Digital Ocean droplet for a lightning-fast, and hands-free, shipping of one of my new vibe coded tools to production.

Here’s what I show in the video:

✅  Using Cursor To Manage Git : Implement GitHub versioning control over projects with zero experience and nothing more than the English language.

✅  How I Manage Deployment With Cursor : Cursor Agent makes deploying complex projects almost effortless (and very fast)!

✅ How The Agent Solves Problems In Real-Time : Watch the agent encounter several problems during the deployment and work around them in real-time.

🚀 AI / Marketing News

AI Skills Boost Salaries by $18K — And It's Not Just Tech Jobs Benefiting

The research is in. Up-skilling in AI, even in non-technical roles, is a quick way to boost your salary by 5 figures. And this is just the beginning…

Source: CNBC

Takeaways:

• AI skills boost average salaries by $18,000 per year across industries

• This applies to roles beyond tech, like finance, marketing, and ops

• Hiring managers prize AI fluency, even in unrelated jobs

• LinkedIn data shows 56% of executives believe AI will reshape hiring

• Cited skills: prompt engineering, data analysis, and ML tools

Knowing your way around ChatGPT or Midjourney isn’t just “for geeks”. It’s how to put your career in growth mode.

No, you don’t have to be a data scientist wearing a Patagonia vest at a VC-backed startup. If you can string together a few AI tools to get results or speed up workflows, you’ve entered rarified income territory.

AI literacy has become the new Excel: you don’t need to build the engine, just drive it better than the next guy.

The future belongs to the folks who actually use the tech, not just tweet about it.

Coinbase CEO Wants AI Writing 50% Of Their Code By October

Coinbase is leading the way, pushing aggressively for 50% of the company’s code to be written by AI by next month.

Source: CoinTelegraph

Takeaways:

• Coinbase CEO wants AI to generate 50% of their code by next month

• He sees AI as a force multiplier for small teams under heavy regulation

• Coinbase built its own AI coding assistant, inspired by GitHub Copilot

• Their goal is to dramatically improve dev productivity

• AI code generation is managed via human-in-the-loop oversight

My Take:

CEO Brian Armstrong’s bold goal for 50% of its code to be AI-generated by October isn’t just ambitious, it’s motivated by battling tech debt, red tape, and the constant rug pulls of crypto regulation.

Are they playing fast as loose? Maybe…maybe not. Building their own AI coding assistant is next-level hacker energy and signals they’re committed.

This signals to every CTO and marketing ops boss out there: if your devs are still hand-typing boilerplate like it’s 2012, you’re wasting burn.

AI isn't replacing developers (not the good ones). Instead it's turning them into dev-mages who ship faster, fix bugs quicker, stay compliant without crying, and maybe even enjoy their jobs.

Armstrong is moving fast here, but he can’t be that far ahead of everyone else.

Venezuela Says Trump Shared AI-Faked Military Strike Video

Trump ordered a massively overpowered missile strike on Venezuelan narcos?

“@Grok, is this real?”

The really interesting thing here isn’t potential deep fakes or government propaganda but the government potentially crossing over into using deep fakes to make themselves look more aggressive than they really are…

Source: Gizmodo

Takeaways:

• Donald Trump posted a video of a strike on drug smugglers near Venezuela

• The Venezuelan government responded, saying the video is fake and appears to use AI-generated content.

• Experts noted telltale signs of AI manipulation, including oddly looping explosions and inconsistent visuals.

• There’s no official record of a recent U.S. Navy operation matching the video’s events.

My Take:

Well, this one’s a twisted cocktail of political theater, deepfakes, and geopolitical finger-pointing.

Trump tosses out an 'explosive' video (pun intended) of alleged Navy heroics, and Venezuela cries foul, claiming it’s AI smoke and mirrors.

Whether it’s real or propaganda, the fact that we’re now questioning reality in a campaign ad means doubling down on our bullshit detectors and realizing that reality’s a game and everyone’s playing…

Google Declared a Monopoly... But Escapes With a Slap on the Wrist

Not sure why anyone expected this to be different. America LOVES monopolies. It was built by monopolists and has always positioned itself in support of the behemoths.

Nothing new to see here…

Source: Wired

Takeaways:

• U.S. DOJ ruled Google will not be broken up despite being a monopoly

• Google retains control over Chrome, Android, Ads

• The court barred Google from making exclusive search distribution deals

• But ruled payments to partners for placement are still allowed

• Google must share index and user interaction data with 'Qualified Competitors'

• Syndication of search and ad services to competitors must be allowed

• Judge Mehta (a stooge) stated harsher remedies would “jolt the system”

My Take:

So Google gets officially tagged as a monopoly and the penalty is... sharing search index crumbs with 'qualified competitors' and being told to play nicer with others.

No divesting Chrome or Android, no ban on stuffing their apps into every device they can pay to hijack; just a gentle nudge not to play exclusively with their distribution pals.

The ruling reads like a gentle slap followed by a back rub, citing 'market forces' like they're some magical invisible hand as opposed to the actual reason Google is a digital feudal lord.

Some things never change (and you shouldn’t expect them to)…

Nathan Binford
AI & Marketing Strategist

I hope you enjoyed this newsletter. Please, tell me what you like and what you don’t, and how to make this newsletter more valuable to you.

And if you need help with AI, marketing, or automation, grab time on my calendar for a quick chat and I’ll do my best to help!

Thanks For Reading H4CKER!