> ## Content Index
> Fetch the complete content index at: https://kelfordinc.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# The discernment to do less — Kelford Labs
- URL: https://kelfordinc.com/the-discernment-to-do-less-kelford-labs/
- Published: 2026-10-29T11:06:00.000Z
- Updated: 2026-10-29T11:05:59.000Z
- Description: So I have the time to do better.
- Author: Joel Kelly
- Tags: ai, advantage, change, constraint, content, control, decision, experiment, mistake, optimization, pace, priority, strategy, thinking, tools, value, Weekly Labs, workflow, writing

I’m retiring the [Daily](https://kelfordinc.com/tag/daily/) version of [Kelford Labs](https://kelfordinc.com/).

For a few years now, I’ve published a newsletter almost every single weekday. But for the past few weeks I haven’t published anything, and, in the coming weeks and months and years, I’ll be publishing a lot *less*.

These are related phenomena.

> “That’s why when Colin Chapman the founder of Lotus cars, was asked why he concentrated on lightness instead of power, he said:

> ‘Adding power makes you faster on the straight.

> Subtracting weight makes you faster everywhere.’”  
>  
> — Dave Trott

**Time was, speed and volume were a demonstration of my value.**

The fact that I both could and did publish that much content, that regularly, said something. It said that I was good, perhaps uniquely so, at coming up with marketing content ideas and getting them out and into the world.

But that’s just not a thing anymore, is it? AI solved the volume problem, anybody can publish at any speed they wish. Daily, sure. Hourly, if you want. *Continually*, even.

**When it’s all just tokens paid for on a meter, the only barrier to publishing *something* is access to an AI plan.**

> “Experts don’t build brands, they build track records.”  
>  
> — Leah Sanford

Now, when I publish something quickly, I always have this thought: Am I publishing it quickly because I *can*, or because it’s the best version of itself?

I don’t want to contribute to the *volume* of content available on the internet or in your inbox, I want to contribute to the *value*, you know?

The instinct to press ahead, at full speed, comes naturally to me and most others, especially entrepreneurs.

But I don’t think that means it’s best to give in to that temptation. Sometimes, and this is one of those times, it might be best to avoid indulging it.

**Because** [***everyone else***](https://kelfordinc.com/kelford-labs-weekly-everyone-has-the-same-bright-idea/) **is giving in to it.**

> “Speed is the second oldest animal craving in our nature, and our generation is fortunate in being able to indulge it more cheaply and generally than our ancestors.”  
>  
> — T. E. Lawrence

Now, I believe that what I was producing with those Daily newsletters was *better* than what others churn out with AI, but better’s not enough. At least, it’s not enough for me.

I think plenty of people would be happy to take AI output, slap a brand on it, or put it in a package, and sell it as their own thinking at a markup.

**The bottled water of insight.**

But I just... can’t. At a certain point I’ve got to actually care and I just don’t care to be that uninvolved in my work.

So that requires *something else*. Something not “better” than AI, but different. Further, higher up the stack, exerting more pressure and influence where it matters more.

Because here’s what I’m worried about the most: Not that AI can do what I do better or faster than I can.

**But that, if I try to fight that fight, I’ll get stuck *doing what AI can do at all.***

> “The very speed of operation of modern digital machines stands in the way of our ability to perceive and think through the indications of danger.”  
>  
> — Norbert Wiener, *Cybernetics*

See, I’ve noticed something, and maybe you’ve noticed it too:

The people who use AI the most for the output of their work aren’t even that good at using AI.

Like, when I talk to someone who’s using Claude to produce all their marketing content, if I ask them what model they’re using, they just say... *Claude*.

Not “Opus 5.5,” or “Fable 5.1”. If I ask why they prefer Claude over ChatGPT, they don’t talk about the relative differences in the Codex harness versus Claude Code. They tend to just say, “Everyone seems to be using Claude now.”

If I ask about their preferred infra provider for open models, I get blank stares.

**Because the drive to use AI for everything isn’t driven, in these cases, by a desire to increase their value or leverage in the priority stack.**

It’s just that desire for speed.

“If I can do more, I can sell more. If I can sell more, I can make more.”

But what happens when everyone has the same bright idea? Isn’t that just standing on your tip toes at a parade, as Warren Buffett once put it? “Nobody can see any better, but they’re all worse off.”

Speed for its own sake becomes a race to the bottom. And not just of cost.

**But of something deeper and more important: *Discernment*.**

> “Their movement becomes headlong—faster and faster and faster. They put aside all thought of obstacles and forget that a precipice does not show itself to the man in a blind rush until it’s too late.”  
>  
> — Frank Herbert, *Dune*

You’ve heard about [slop grenades](https://fortune.com/2026/09/17/shopify-tobias-lutke-ai-slop-grenades/), right?

It’s the name for sending someone AI output that wasn’t reviewed or considered. It’s now just the receiver’s problem to deal with.

And the people talking the most about it are the executives at companies that mandated their employees all use more and more AI. Well, be careful what you wish for.

**Or, be careful what you incentivize.**

Because people are just absolutely *cooking* their brains with this stuff. I’ve witnessed firsthand colleagues and acquaintances seemingly losing their ability to read and write without AI assistance, and it’s tragic and extremely frustrating.

I’ve had the unfortunate experience of working with someone who requires everything to be sent to them in increasingly simplified bullet points, because if you send a paragraph or a spreadsheet of data they just can’t comprehend it anymore.

**They *could*, but now they *can’t.***

> “Some men are born mediocre, some men achieve mediocrity, and some men have mediocrity thrust upon them.”  
>  
> — Joseph Heller, *Catch-22*

Did you see that Marissa Mayer quote going around?

Brace yourself, it’s brutal:

![](https://storage.ghost.io/c/5a/75/5a75dc74-aafc-42ff-a23a-da6b80d4c2f9/content/images/2026/10/marissa-quote.webp)

From this [NYT profile of Marissa Mayer](https://www.nytimes.com/2026/10/25/style/marissa-mayer-ai-assistants-women.html)

As others have already said on social media, that’s the kind of admission you couldn’t torture out of me. But she just... said it.

As Greg McKeown wrote in *Effortless*, “No matter how simple the step, it’s still easier to take no step.”

And AI allows people to take no step at all.

> “Not to take the first step without thinking what may be the last.”  
>  
> — Clausewitz, *On War*

Personally, my goal has always been to move up the stack of responsibility, not to find ways to shirk it.

I started my career in advertising as a production coordinator, making sure the clients who booked ads in publications actually got their ads published.

It sounds ridiculous now, but that job required entering data into two separate databases, manually, printing off Insertion Orders, and physically taking those from the salespeople, to the graphic designers, and then *faxing the designs* to the client, and repeating that cycle until sign-off.

**Luckily, I didn’t insist on doing that same job forever because much of that one doesn’t exist anymore.**

Instead, I decided to get a job at an advertising agency, where the real creative decisions seemed to be happening.

That job had me planning the digital media buys for our clients, choosing which websites and, eventually, social platforms to place their ads on, and then doing the actual negotiations and purchases.

**But I could see that the real leverage was in the creative strategy that ultimately determined the direction of those campaigns.**

So I got a job at another agency as a “Digital Strategist,” planning social media campaigns and activations, alongside the creative and media teams.

There, I noticed that social media marketing was not really “digital,” it was just the means by which word-of-mouth spread and distributed itself.

So I took a job as a Marketing Specialist at the premier social media measurement platform at the time, creating the product marketing pages, presentations, and videos that promoted the company to its Fortune 500 clients and prospects.

Leveraging my knowledge of social media measurement and analysis, as well as digital strategy and online reputation management, I ended up at a PR agency, moving up the stack again, creating the strategies for major PR and marketing campaigns (winning that agency its first marketing award).

Later, I got the opportunity to build my own department at a global tourism marketing agency, leading a team of social media marketers. But I saw that social media was just a subset of content, so I expanded the team’s purview into all of content marketing.

Then, to marketing generally, developing the strategies, content, and copy for TV spots, virtual reality campaigns, and mobile apps.

**Up and up the stack, closer and closer to where the most important decisions get made that really *make* marketing successful or not.**

I made the decision, from there, to go out on my own, to the very top of the organizational decision stack by starting my own company.

[Here](https://kelfordinc.com/), we work with the executives and leaders of advisory firms and non-profits to make the decisions about where, how, and when to do their marketing.

At the very top of the stack, where all the leverage is exerted, and where most of the value is created.

**I say all this because I think AI is an inherently stack-destructive technology if used exclusively or too abundantly.**

It puts the tool above the user, the technology above the tactic. The execution above the intent.

Frankly, as a millennial, I think a lot about quicksand (if you’re a millennial you know what I mean). And I think AI can be a lot like quicksand in movies:

***The faster you move, the more you sink*.**

> “‘Doing things badly,’ Jiménez would say to critics or editors or even impatient readers, ‘does not give you the right to demand haste from the person who does them well.’”  
>  
> — Ryan Holiday, *Discipline is Destiny*

Because to stay *ahead* of AI, I need to stay *above* it, not behind it or shackled to it.

**I need to be the person deciding *which AI tool to use*, or *if to use it at all*.**

I need to know which model, harness, or infra provider is best for which use case, and which to avoid.

And I need to be able to help others make these same decisions, at the top of the stack, not from under it.

I guess what I’m saying is I don’t want to be the person who *reads* LinkedIn posts about which AI tool is most popular right now, I want to be the person *writing* the post about how to *think* about the tools and how to *decide* for oneself what to use or whether to.

> “What is the lesson? It is that none of us can afford to rely on judgement or experience. We must feel our way. New problems require new experience. We must test our undertakings in the most exact way possible. Learn our mistakes and correct them. Watch every appealing lead.”   
>  
> — Claude C. Hopkins, *My Life in Advertising*

Which means taking time, or making it.

**It means *slowing down*.**

Trying to “keep up” with AI is a mistake. Better to be *above* it. Better to be the person the AI labs are trying to beat, than to be the person following the AI labs’ direction.

I want to do things AI can’t currently do, even if they one day will. That gap is my opportunity.

That gap is my *interest*.

That gap, and building it, and showing others how to create their own, is what I do.

**Giving up quality for speed is a false bargain, because I’d end up losing both at the same time.**

If I stop thinking and just use the tool, I’ll get worse at using the tool, too. So I’m not even the fastest anymore. Because I’m not in the driver’s seat, I’m hitched to the back and I go where the tool takes me.

> “If you were driving a car with fogged windows or faulty breaks, the first thing you would do to avoid overshoot would be to *slow down*. You would certainly not insist on accelerating.

> ...

> Even a car and driver functioning perfectly are unsafe at high speeds. The faster the growth, the higher the overshoot, and the farther the fall.  
>  
> — Meadows, Randers, Meadows, *Limits to Growth*

I’ve come to realize that what I sell, when it comes to marketing, is not content.

It’s not ad copy.

It’s not even decision-making.

**It’s *discernment* itself.**

It’s the ability to judge a marketing scenario, challenge, or opportunity and to understand it intuitively. And to know, and be able to explain *how I know*, what decision to make. Why, and when.

That takes time, that takes broad experience, and that takes a sustainable pace.

Yes, I use AI for *lots of things* (my token burn is measured in the hundreds of millions), so it’s not like I have some moral high ground to argue from, as much as I might wish I did.

**But I also *don’t* use AI for lots of things. So that I maintain the ability to see it as a tool, not as an oracle or the exclusive font of knowledge from which to draw.**

Warren Buffet once said he wouldn’t invest in businesses where the technology went way over his head. I think that’s a good lesson for entrepreneurs and marketers, too.

If I didn’t know how AI works, it would be such a massive risk to invest in it heavily.

So I spend a lot of time learning about it, experimenting with it, and watching how other people use it.

> “What we call expertise is really just ‘vast amounts of knowledge, pattern-based retrieval, and planning mechanisms acquired over many years of experience in the associated domain.’ In other words, a great memory isn’t just a by-product of expertise; it is the *essence* of expertise.”  
>  
> — Joshua Foer, *Moonwalking with Einstein*

Because I want to have an intuition about it, instead of just following the latest trend or “[best practice](https://kelfordinc.com/daily-lab-best-practices-are-classic-mistakes/).”

Frederick II, in his *Instructions for his Generals*, talked about the “coup d’oeil,” or the ability to assess a situation at a glance: “The judgement that is exercised of the capacity of the enemy at the commencement of a battle is called *coup d’oeil*. This latter is the result only of experience.”

Now, I don’t like talking about marketing in military terms when I can avoid it, but, in a sense, using technology is a battle. To know what to use, when and how to use it, and how it interacts with how *other people are using it* has the dynamics of a struggle.

**And I want to have the best *coup d’oeil* I can have, when it comes to marketing and marketing technologies.**

That requires experience, more than anything else.

> “Understanding is something that is internal to a person. It has to do with his ability to conceptualize and to match those concepts to his experiences, on the one hand, and to the expressions of the new language on the other.”  
>  
> — George Lakoff, *Women, Fire, and Dangerous Things: What Categories Reveal about the Mind*

All of this is to say, trying to compete on volume, on speed, is a mistake, at least for me.

Yes, I’m a fast writer, but my tokens per second would be embarrassing for any AI model.

**What I offer is discernment, a level higher up in the stack that understands why to use AI, or if to.**

How to use AI, and how not to.

So I’m going to be writing [less, better](https://kelfordinc.com/do-less-better/). I’m going to be sharing more of what I’m learning about tools and technology, and what that learning means for both me and you.

**I’m going to be helping you, if you’d like that, to build more intuition about these technologies and tools, so you can have more discernment about them, too.**

Instead of relying on them solely, or taking the words of the AI providers at face value.

> “Intuition depends on the use of experience to recognize key patterns that indicate the dynamics of the situation.”   
>  
> — Gary Klein, *Sources of Power*

In *Limits to Growth*, Donella Meadows and her co-authors wrote:

“One can’t learn without making mistakes, telling the truth about them, and moving on. Learning means exploring a new path with vigor and courage, being open to other people’s explorations of other paths, and being willing to switch paths if one is found that leads more directly to the goal.”

**So, what *is* that goal?**

Is it speed? Is it volume? Is it the sheer appearance of marketing?

Or is it something else, something greater, something higher than that?

For me, marketing is about demonstrating value at a distance.

So that more people take the actions that are advantageous to them and the service provider they choose to work with.

**That’s why I’m retiring the Daily Labs: So I can spend more time, and more discernment, where it matters most.**

I’ve spent more than 20 years now building a career that elevates my position within that stack to the point where I don’t have to yield to the pressure of new technology.

But instead I can wield it as one of many possible tools.

**And that’s what I’d like to help you do, too.**

Let me know what you think.

*If this has you thinking about how, or whether, to use AI or other tools in your marketing, I’d love to chat. Just reply to this email.*

---

[***Kelford Inc.***](https://www.kelfordinc.com/) ***are content marketers for expert advisors. If you’re struggling with what to say and where to say it to attract your ideal clients,*** [***we’ll show you the way***](https://kelfordinc.com/contact/)***.***