FIELD NOTES #00: Vibe-Coded Founders, AI Slop & Other September Debris

Hey, you with the eyes and the healthy attention span!

I applaud your ability to make it to the end of one, if not all, of my mad ramblings. What a titanic feat. 

Not lost on me in 2026, when most people suffer from constant overdoses of dopamine and stimuli.

When I first started Gonzo Capital this summer, I had my doubts:

• Does the world need another newsletter?
• Are my thoughts enough to carry any influence or interest?
• Do the topics that I obsess over interest anyone else?
• Can people even read good anymore?
• Do I even have time to churn out a small novella each Wednesday?
• What happens if I run out of things to write about?

And now a new worry has entered the room:

“Can I ever take a week off?”

So I had the idea to do Field Notes!

A monthly newsletter that gives me a break from essay writing. That allows me to reflect on the previous month and gives me some time to develop the next slate of writing to unleash the insane thoughts and existential worries rattlin’ around my overworked coconut.

Well, Field Notes was supposed to be this sort of easy, quick, pain-free way to avoid writing a thesis every week. Now it’s somehow become more difficult and agonizing. Despite being more sincere and casual, too. Especially this first one, where I’m trying to find my footing.

My incredibly wise better half also encouraged me to do something like this in an effort to make this more collaborative and less “old man screams at the clouds.” To better understand the fine folks that have subscribed and taken an interest here, and further develop that into a living, breathing organism.

So I’ll keep it short and sweet this time around for the opening bit.

So without further ado, welcome again to Field Notes, welcome to Gonzo Capital, and thank you again for your support. It’s truly a pleasure and a privilege to have you in whatever Gonzo Capital is currently becoming.

Things Occupying My Brain

Not everything that survives the month becomes an essay. Or deserves one. Maybe some of these eventually do become their own essays. Let’s wait and see. Feedback welcome!

Some ideas are still missing pieces, some are probably wrong, and some become less interesting the moment you try to force a conclusion onto them. 

These are the ones that are still alive enough to keep around:

The Vibe-coded Founder

AI (and social media) has made it dramatically easier to manufacture the visible evidence of a startup.

You can build a respectable website in an afternoon, generate a deck, produce product screenshots, create market research, write the copy, make a demo, and dress the whole thing in the tasteful geometric gradients that apparently signal technological seriousness in 2026.

And that’s mostly a good thing. Mostly.

Small teams can test ideas without spending months or huge amounts of money. 

Someone without a traditional technical background can turn an idea into working software instead of waiting for permission from someone who can.

With this seismic sea change, we may be witnessing the emergence of a new entrepreneurial subspecies, Homo conditor vibecodicus, identifiable by a fully functioning product, immaculate deck, tasteful gradient website, and only a passing acquaintance with the problem being solved.

To be clear, I ain’t talking about founders who use AI to code. 

That is rapidly becoming the new normal.

I mean founders whose company has been assembled faster than the underlying problem has been understood. Usually to jump on the latest meta and raise funds faster than the other twelve million people with the same idea that may or may not have come from asking an LLM.

The deck is excellent. The product exists. The market is enormous. The branding looks like the company has already raised three rounds.

Then you ask why anybody actually needs it.

Something begins rattling underneath the hood.

For years, polish was at least a weak signal of effort or lack of it.

A good pitch suggested somebody had spent enough time with the company to explain it coherently. 

A working product implied a certain amount of technical ability. Research suggested somebody had actually researched something.

Those signals are getting a hell of a lot cheaper. Well, as cheap as a Claude or ChatGPT subscription.

The cost of looking credible is falling much faster than the cost of being right.

Taste is beginning to matter more. As is judgment and experience. I’m more inclined to trust people who have been working in that niche for ages. People who intimately understand the nuances and core issues plaguing it. 

Knowing which problem deserves solving matters more. Understanding the customer matters more. The ability to make something suddenly matters slightly less when everybody can make something.

The Pitch Deck Industrial Complex has officially acquired automation. So now we must adjust how we think of founders and how we invest in them.

The Different Types of Slop

Slop is slop.

I don’t care if a human created it with or without an LLM.

Accusing something of being AI slop is the modern witch trials.

If you don’t like someone or something, you can easily discredit it by calling it AI slop.

For now, I’m not ready to say that AI isn't capable of making something that isn’t sloppy. I saw a trailer this morning for an all-AI movie that looked decent enough to at least entertain the idea of sitting through.

So if AI doesn’t automatically mean slop (it may; the verdict is still out on that one), does slop automatically mean AI?

I think we may struggle to think back to the days of social media before LLM usage became ubiquitous. 

My timeline was riddled with low-effort, copycat, empty-calorie, downright misleading content. 

Human-generated slop.Anyways, if you hate AI slop, that’s fine. I do too. If you hate all AI output, I fully understand.

But let’s be clear where we stand with it. I don’t hate AI slop because it’s AI. I hate it because it’s slop. Lazy slop. 

You can argue that the creator spent even less time on it and therefore it’s even lazier slop. But slop is slop, and I wish I could scrub it all from the ever-cluttered internet.

Experimental 

Yes, I want to do experiments each month. 

To keep me honest, and growing, and learning. Also to push my comfort zone and perform a humiliation ritual for you wonderful lot taking the time to read about it.

Since I have only just now thought to build this section, I’m going to let you in on a little secret: I didn’t really do any experiments this month.

So I’ll cheat and use this essay, The ROI of Staring at a Wall, instead!

I literally challenged myself to stare at walls for 20 minutes a day for a meager 2 weeks, and struggled mightily, but I learned a lot about myself and how I constantly need to find tangible and immediate value in anything I do.

Oh, and also that meditation can be a powerful tool to fight cognitive backslide and brain swampiness in the age of… what are we supposed to be calling it now?

Super intelligence?

No, I’m sticking with AI, sorry.

In the age of AI.

Give it a read, and give me your best ideas for October experiments; I want to hear them all!

Marginalia

A few smaller observations and musings from the month that I totally 100% have been writing down as I go and didn’t wait until the last minute to jot down:

• AI has made me more interested in ugly work. A rough draft now contains information that polished output sometimes removes.
• The most dangerous phrase in a strategy meeting may be “we should also.” I’m guilty of this.
• AI could make eccentricity economically valuable. If competent, polished output becomes abundant, the weird human sentence starts looking like provenance.
• We can measure whether someone stopped scrolling for 1.7 seconds, but remain strangely incapable of measuring whether anything resonated or stuck with them.
• I think we should stop using the word robot. It comes from the Czech word robota, which means forced labor. I feel like this may someday be offensive to our machine friends.
• The amount of spending on weapons and “defense” right now is horrifying. And most of it is offensive.

One Useful Thing

I don't want Field Notes to consist entirely of me staring suspiciously at modern civilization, or screaming at the sky. There’s got to be some actionable wisdom here.

One useful idea each month. No elaborate productivity doctrine. No downloadable system involving seventeen databases. Just pure unadulterated recommendations from a mad scientist.

The Explain-It-to-a-Normal-Person Test

I recently wrote about how unrelatable and bad at explaining itself the tech industry is. Tech doesn’t understand why you don’t like it!

So how do you validate your ideas and business plan outside of your echo chamber?

One useful test for startups, investing, and technology is embarrassingly simple:

Can you explain the thing to an intelligent person who does not work in your industry?

Not your co-founder. Not another investor. Not someone who already knows what an agentic workflow is. A normal person.

If the explanation immediately requires six acronyms, three market categories, and a brief history of decentralized compute infrastructure, something may be wrong.

Now, this need not mean complicated things need to be dumbed down. Nuclear reactors are complicated. Semiconductor manufacturing is complicated. I won’t pretend that I know the inner workings of nuclear reactors or humanoid robots just because I’m an investor who has no life and obsessively reads about new tech.

But complexity and incomprehensibility are not the same thing.

A genuinely complicated company can usually still answer basic questions in plain language. In fact, the best have mastered this art.

What does it do?

What problem disappears if it works?

Why would somebody pay for it?

I have noticed that weak ideas often become stronger-sounding as the vocabulary gets more specialized. 

And the jargon? It never clarifies ideas; it gatekeeps them with a large moat to keep pesky non-techie outsiders on the outside.

“AI-powered orchestration layer for autonomous enterprise workflows” may describe something genuinely useful. If only you could understand what those words mean.

It may also mean software that helps employees move information between applications.

Those are very different situations. So strip away the vocabulary. Unpack what they mean, and simplify them.

Explain it like the person listening is smart, curious, and completely uninterested in pretending they understand you.

If it suddenly becomes ordinary, that can also be useful. Ordinary problems make excellent businesses.

If the idea completely falls apart once the specialist language disappears, investigate why.

All this from a guy who will continue to use tech jargon and forget this advice from time to time. Classic!

The Idea Graveyard

Most bad and some decent ideas disappear quietly into Notes folders. 

This seems unfair. Some deserve a small funeral, particularly the ones that looked excellent for several hours before reality arrived carrying paperwork. Some could be solid if I had a cloning device or a time machine.

The Giant Open Founder Community

Thousands of founders helping each other, swapping resources, making introductions, trading war stories, comparing battle scars. Basically collectively becoming better at building their own companies.

The probable reality contains seventeen thousand muted notifications and somebody posting:

Hey everyone! Excited to share what we’ve been building!

followed by a Calendly link.

Cause of death: signal-to-noise ratio.

The underlying idea may survive, but I now think selection has to be part of the product.

Publishing Only Fully Formed Content

This sounded responsible.

It is also an excellent strategy for publishing approximately once every solar eclipse. I recently gave this advice to a friend who has also been reluctant to get out there and start doing the damn thing.

Put your unfiltered thoughts out there and start getting on the record about what you believe in. Ok well, filter them a little bit. But you get the gist.

One reason Field Notes exists is to give ideas somewhere to live before they deserve an essay.

Cause of death: perfectionism wearing business casual.

Things Worth Your Attention

There is more worthwhile material available than anyone could consume in several lifetimes, buried underneath an even larger quantity of things that absolutely do not deserve your Tuesday afternoon. 

This is a small part of what made it through my filter.

Left Hand of Darkness

I’ve got an unhealthy obsession with science fiction. It may or may not be part of why I like investing in tech companies.

I’ve just finished Left Hand of Darkness by Ursula K. Le Guin, and it was remarkably breathtaking and surprisingly light on the actual science or tech. 

What if a species had no fixed sex? What would society look like?

Gethenians are generally androgynous and become sexually male or female during kemmer.

Well, I suppose it’s more focused on social science and anthropology than the actual space beams and farcaster portals you’ll find in other works of Sci-Fi.

I’m planning to write a big ol' essay (or more) on Science Fiction and my favorite writers, to help introduce my favorite genre to more folks out there.

For now, I’ll just say, give this a read. And let me know your thoughts if you do! 

Maybe I’ll end up organizing a book club.

You Can See Everything

Remember Elizabeth Holmes? 

The disgraced former CEO of Theranos who defrauded investors and once sat at the top of a $9 billion company?

The ever-hilarious Nathan Fielder, along with talented filmmaker Lance Oppenheim, is releasing a documentary that looks like it could be movie-of-the-year territory. 

Even for people who hate documentaries and don’t understand the tech startup scene at all.

Will it clear her name, add insight to the post-mortem, or serve as a cautionary tale? Who knows.

Hasn’t come out yet. It’ll be out on October 16th. I’m definitely going to watch it.

Boysclub.world for All Things AI

No, it’s not as weird as it sounds; safe to click, scout’s honor.

I go on lengthy diatribes quite often and sometimes overexplain to my own detriment. Boysclub and Wetclaude came across my feed, and she’s an absolute legend at breaking down high-level, complex AI thingies and translating them into lay vernacular.

I’ve been spending an uncomfortable amount of time and cerebral real estate as of late, thinking about p(doom). I’ll let them explain what that is. They’re much better at it.

An Unfinished Question

There are already enough people online selling answers. 

I would like to preserve a small habitat for questions I genuinely do not know how to answer.

What are we supposed to use agentic AI for?

I see Meta just came out with Muse; it's the answer to OpenClaw that they basically copied.

Most frontier labs will come out with some version of this.

I’m struggling to find ways to use this to make my life better and more efficient/productive.

I even stumbled upon this wonderful piece, Am I Too Boring to Need an AI Agent, and I find myself wondering something similar.

I also see nearly half the people I know on LinkedIn bragging about how they’ve surpassed me in output and now do the work of an entire team at a fraction of the cost and time.

The other question I have is: How competent is this team?

I don’t want AI to shop for me, handle my comms, set appointments or meetings, or do my research.

What else can I use it for? Genuinely asking for help here!

The Month’s Most Absurd Sentence

Future historians are going to have questions about us. I would like to help by preserving some primary-source material.

Every month, one real sentence from technology, finance, business, academia, politics, or the internet that deserves to survive the collapse.

There are rules.

The sentence must be genuine. Its author should preferably have meant it sincerely. Context should make it more absurd, not rescue it.

Corporate euphemisms receive favorable treatment. Management-consulting language will be considered a performance-enhancing substance.

September's winner

“If you’re evil, you’re at least competent. Um, and if you’re evil, you’re not bad. And therefore, you’re maybe actually kind of good. Because you’re at least getting something done.”

- Peter Thiel - 

Source: Interview with Peter Thiel

To be fair to Thiel, this came in the middle of a longer argument distinguishing moral evil from functional incompetence, and he immediately clarified that he thinks being good is preferable.

None of which makes this sequence of words appreciably any easier to process.

Can anyone else make out what this maniac is saying?

Honestly, I may have to amend the rules and say no Peter Thiel quotes, or else he may dominate this category each month.

When he isn’t trying to purchase politicians or give lectures on the Antichrist, he likes to give cartoonishly bad interviews and soundbites. Can he not afford PR people?!

If you encounter a strong candidate for October, send it to me.

There may need to be a trophy at some point.

Your Field Notes

I don't want this to become a monthly document fired from my inbox into yours while everybody politely remains on their side of the glass.

So every Field Notes will end with one question I actually want answered. 

Something made to provoke conversation and get to know you all better.

I'll bring some of the best responses into the next edition. Anonymous by default, or credited if you want your name or handle included.

For the first one:

What have you become weirdly obsessed with recently?

It can be useful, serious, or completely pointless.

A company, technology, historical event, book, obscure website, scientific question, forgotten civilization, particular kind of chair, insect, machine, map, conspiracy about airport carpet.

The more specific, the better.

I am particularly interested in the subjects you struggle to explain to normal people without sounding slightly unwell.

Hit reply.

I want to know what everybody else has open on their mental desktop.

See you in the margins.

Your weird friend,

- Carmelo -