👋 Howdy to the 1,984 new legends who joined since our last edition! You are now part of a 547,495-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🥅 Ted Lasso: The Lasso Way. A story of empathetic leadership from the Manager of AFC Richmond.
💵 Take The Cap Off Your Salary Bands. An interview with Matt McFarlane, Founder & Director at FNDN.
🥣 Granola: Healthy Breakfast & Killer AI Notepad. The AI notepad that ate a category.

PARTNERS 💫

AI is changing how fraud shows up and how quickly it can scale. Signals that once worked are easier to spoof, forcing teams to choose between higher losses and more friction.

Plaid's new report explores why traditional identity approaches are breaking down in the AI era and how financial behavior and broader context help teams catch more fraud earlier while maintaining seamless user experiences.

Every hiring pitch sounds the same: faster, smarter, cheaper, blah. I know, because I sat through most of them before deciding to just build the thing myself.

Athyna is smart hiring for teams that need the best talent now—not next quarter when the role’s gone cold. The difference is AI-assisted matching paired with proper human vetting, so you get people who fit the role and the team.

Less back-and-forth, less cost, built to scale. That’s it.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

This email was supposed to go out last night. Truth is, I fell asleep—literally so exhausted I couldn’t move—and actually forgot about my newsletter entirely. If you’re reading this, you’ve probably been there. It’s not great, and kinda embarrassing.

_{*I’m aware this video has nothing to do with the paragraph above, but I still wanted to share it because I found it sooooo funny.}

It’s a lot running this newsletter as its own startup (we have a team of four now) along with my startup baby Athyna. I’m also trying to train like a semi-professional athlete at 40. I guess every now and then I’m going to fall in an absolute heap. So, apologies for the late send; trust that I have given myself a firm facepalm. Here’s hoping you enjoy this one!

GUEST POST 💄

The Wikipedia Paradox: Fully Editable, Fully Trusted

Today, trust is not so easily earned. Especially in tech. Facebook, Polymarket, Flock Cameras, Palantir. The public has gone from ‘these guys are building the future’ to ‘these guys are building a surveillance state with a loyalty program’ in about a decade.

Even Google quietly retired ‘Don't be evil’ back in 2015, which is a hell of a thing to have to retire. And yet one tech brand has somehow escaped the whole mess. The seventh most visited website on earth. More trusted than the BBC, most governments, and definitely your uncle's Facebook feed. Yes, we are talking about Wikipedia.

Yet, about once a year, Jimmy Wales's face pops up at the top of Wikipedia asking for three dollars. A sad yellow banner, a headshot from a wedding in 2009, and copy that reads like a hostage note. Every year I close it and feel bad for about a week. Today, my buddy Isaac Peiris is going to break down why Wikipedia does this, what it costs, and what you can take from it. Over to you, Isaac.

How Wikipedia builds trust

Hey, Isaac here! I write Brand Chemistry, a weekly newsletter about how brands earn an audience and turn it into growth. Which means I have been talking about trust for a while now, and Wikipedia always stands out as a brand that, on paper, shouldn’t be trusted at all. But somehow it’s one of the most trusted names in the world.

On 26 May 2005, someone typed a sentence into Wikipedia claiming that John Seigenthaler, a newspaper editor who had once been Robert Kennedy's administrative assistant, was suspected of involvement in both Kennedy assassinations. A stranger invented it as a joke. It sat on the page for 132 days. Seigenthaler wrote a column in USA Today asking how a website like that was allowed to exist. It was the worst possible advertisement for the idea that a public encyclopedia anyone could edit was ever going to work.

Twenty-one years later, anyone on earth can still edit any page on Wikipedia, anonymously, without an account. Wikipedia is also the single most-cited domain in ChatGPT's answers, at 13.15% of all citations across the roughly 600,000 citation events Similarweb tracked in early 2026, ahead of Reddit and ahead of every publisher on earth. Those two facts should not survive in the same paragraph. Wikipedia's answer to being humiliated in public was to leave the doors open and make every future humiliation permanent, timestamped, and searchable by anyone who wants to check.

Jimmy Wales, Wikipedia’s co-founder, spent last year writing down why that works. His book, The Seven Rules of Trust, arrived in October. I went in expecting a list and came out with something more useful: an explanation of why trust behaves differently from every other asset a founder tries to build.

Someone else holds the verdict

Rachel Botsman, Trust Fellow at Oxford's Saïd Business School, defines trust as a confident relationship with the unknown. Her sharpest point is that if you already knew how something would turn out, you wouldn't need trust at all. Trust exists only where risk does. It is what lets someone act before they have proof. Wales arrives at the same place from the other direction.

Put those together and trust is a risk that someone else decides to take on you. You can be trustworthy on your own. Whether you are trusted gets decided outside your control, inside other people's heads, which makes ‘build trust’ a strange thing to put on a quarterly plan. Charles Green and David Maister gave it a shape in The Trusted Advisor. Their trust equation puts self-orientation in the denominator, so the more visibly you are in it for yourself, the smaller the number gets. Credibility and reliability on the top line cannot rescue it.

Bill argued in a guest edition of Brand Chemistry that trust now turns up on a balance sheet. An asset that only appears on your books once someone else has decided it belongs there is a strange form of ownership, and it helps explain why the companies that try to buy it can struggle to reap the full benefits. Trust resists acquisition. When OpenAI paid nine figures for TBPN, a YouTube channel with fewer than 70,000 subscribers, they bought distribution and the credibility of two hosts. The verdict of each individual viewer was the part that was never for sale. You spend years earning your way towards trust, and somebody else decides whether you have arrived.

Not all seven rules pull equal weight

Wales has seven rules. He presents them flat, seven of equal weight. They aren't. Read together, they sort into three questions a person runs through before deciding to risk you, and they arrive in that order. Do you see me? Can I check? And will you hold when it costs you?

Warmth comes first

Make it personal, be positive about people, and be trusting. These lean on Frances Frei's trust triangle from Harvard, her map of the three things a person needs before they will trust you. The version everyone knows is a grieving customer emailing Chewy to return an unopened bag of food after her dog died. Chewy refunded her, told her to donate the food to a shelter, and sent flowers signed by the rep who took the call. Self-orientation at zero, in the equation's terms.

Ritz-Carlton runs the internal version of the same rule. Any employee can spend up to $2,000 per guest to fix a problem without asking a manager first. That is ‘be trusting’ pointed at your own staff instead of your customers, and it costs real money every year to keep the policy honest. This whole cluster is table stakes. Missing any of it ends the conversation. Clearing all of it earns you consideration and nothing more.

💡 Note: If you’re enjoying this one, make sure to check out Isaac’s own newsletter, Brand Chemistry. It’s a good one.

Show the receipts

Create a clear purpose, and be transparent. A purpose is the promise, and without one there is nothing to hold you to and no test you could ever fail. Transparency is the receipt. It lets people watch you keep the promise without having to take your word for it.

Wikipedia does both at once. The promise of free access to the world's knowledge is backed by non-profit status. Every edit, every revert, and every argument between editors sits in public, permanently, for anyone to inspect. English Wikipedia has logged more than 1.3 billion edits, and roughly 291,000 people edited something in the last thirty days. All of it on the record. You trust the article because you can watch it being fought over.

Even the flaws read as proof. Northwestern research on customer reviews found purchase likelihood peaks somewhere between 4.0 and 4.7 stars, then falls as ratings approach a flawless 5.0. A visible flaw that can be checked beats a perfect score that cannot.

Two companies turned that into an operating system. Buffer has published every employee's salary, and the formula that produces it, for over a decade. GitLab runs a public handbook of roughly 2,000 pages covering compensation bands, board material, and how they handle a security incident, with an internal rule that every change at the company gets made in the handbook first. Both handed their competitors a map. Both get candidates who turn up to the first interview already sold. The cost is the point. A policy that costs nothing to publish tells a candidate nothing.

I get asked about this a lot at Pistachio, but the request almost always comes pre-defused. They want the benefits of the transparency page without the disclosure, the values without the number attached. But this only works with the actual receipts.

What did it actually cost you?

Be civil, and be independent. The other five rules get you considered. These two get you believed, because these are the only two that can actually cost you. Start with independence. In February 2002, the Spanish Wikipedia community forked the project and walked out over a rumor that Wales might put ads on the site. By that August, he had committed, permanently, to running none. Some estimates put the forgone revenue at around $2.3 billion a year. The Wikimedia Foundation operates on roughly $185 million a year. They turn down more than ten times their entire annual budget every year, and have done it for twenty-four years.

That invoice is the reason ‘anyone can edit it’ never turned into ‘nobody can trust it.’ Economists call this ‘costly signaling’, after Michael Spence's 1973 work on job-market signaling. A signal carries information only when faking it would cost a liar more than the lie could return. Every cheap claim a company makes is therefore evidence of nothing. The only thing a mission statement on your careers page proves is that somebody can type.

The independence rule has a trap, where independent gets read as a synonym for neutral. Wales's own qualifier is that you stay out of public fights unless they are directly relevant to what you do. Wise clears that bar. Their whole business is built on the promise of fairer foreign exchange fees, so saying loudly that banks were stealing from customers through exchange rates was just naming what Wise existed to fix.

Calling someone else out is itself a promise about your own conduct. It turns the scrutiny straight back onto you. Wise could afford that because they stood on the customer's side and had a product that followed through. Pick the wrong fight and the maths flips quickly.

The most trusted institution left is the one you run

This is where the Barometer gets uncomfortable, and interesting, for anyone running a company. Trust in almost everything is falling. Only 22% of people say they currently trust someone meaningfully different from themselves. In the middle of all that, 78% said they trust their employer to do the right thing, the highest score of any entity Edelman measured. Higher than government, higher than media, higher than business in general, higher than NGOs. The last institution people still believe in is the one that pays them.

That is an enormous line of credit sitting inside your company, and most founders draw down none of it. 73% of people say CEOs are obligated to help bridge trust divides. 44% think they do it well. Twenty-nine points of demand that nobody is serving.

The seven rules read differently once you notice that. Every one of them is cheaper to run inside a company than outside it, and every one of them is verified faster. Your team can see whether you make it personal, whether you are actually trusting, whether the stated purpose survives a bad quarter. They watch you get tested weekly, in real time, with no press release in between.

Employees are the only audience that gets to check the whole equation. They are also the audience most likely to tell everybody else what they found. For a company whose talent is the product, that verdict is the growth strategy.

There's always an invoice

Build is the wrong verb for trust. You make a promise specific enough to be tested, and then you pay the bill when it lands. Most companies never get there. They never get trusted because they are never truly tested, and they are never tested because they never made a promise that could cost them anything to keep. They stayed safe, stayed liked, and stayed forgettable.

Wikipedia got humiliated in public in 2005 and answered by making every future test public too. Twenty-four years of turned-down ad revenue is the invoice. Top of the citation ranking of the machines now answering the world's questions is the receipt.

Nine figures bought OpenAI a microphone. Whether anyone believes what comes out of it is still, as it has always been, a verdict reached one viewer at a time. The bill always arrives. Trust is what you have left once you have paid it.

Playbook

• Write the promise down somewhere it can be tested: A purpose vague enough to survive anything cannot be failed, which means it also cannot be passed. Say the specific thing you will do and the specific thing you will refuse, then put it where a customer or a candidate can find it without asking you.

• Publish the thing that would embarrass you to publish: Salary bands, the incident post-mortem, the churn number, the roadmap you missed. Buffer and GitLab handed competitors a map and got a hiring advantage in return. A disclosure that costs you nothing signals nothing.

• Pick only the fight your business gives you standing to pick: Wise could attack bank FX fees because fixing them was the product. Test any public stand against one question, which is whether losing the argument would also mean losing the company. If the answer is no, you are borrowing a cause rather than holding one.

• Spend the trust you already have inside the building: 78% of people trust their employer more than any other institution, and only 44% think CEOs use that position well. Your team is the fastest, cheapest, and most credible audience you will ever have. They are also the ones who talk.

• Stop measuring trust as sentiment; start measuring it as tests survived: Brand-tracker scores move for reasons you did not cause. Instead, count the moments where keeping your promise cost you money, and you kept it anyway. That number is the only one that compounds.

Future

AI has dropped the cost of a confident claim to zero. Anyone can now generate a values page, a founder manifesto, and a set of testimonials before lunch, which means the market is about to be flooded with signals that cost nothing to send. Costly signaling says every one of those is worth exactly what it cost to produce.

The scarce thing becomes the verifiable receipt. Wikipedia is second in AI citations because its entire argument history is machine-readable, permanent, and free to audit. Companies with a public handbook, a public changelog, published pricing, and a decade of decisions anyone can go back and check will be the ones AI can actually cite.

The rest become unsourceable, which in an answer-engine world is close to invisible. Trust was always decided by other people. Increasingly, some of those people are machines, and machines only trust what they can check.

Extra reading

• The trust economy - July, 2026

• The Future New Media: a16z - June, 2026

• 2026 Trust Barometer - 2026

And that’s it! You can follow Isaac on LinkedIn and X (Twitter), and make sure to subscribe to his newsletter, Brand Chemistry, as well!

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Svix—a webhooks-as-a-service platform that helps developers send secure, reliable, and scalable webhooks.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors and leaders in tech. 👀