WordPress's $8.15m Golden Parachute  🔗 **Source:** [WordPress's $8.15m Golden Parachute](https://www.youtube.com/watch?v=3ONEJZ-Pf8Y) > [!note]+ Document Notes > tags: [automattic-governance, corporate-governance, source-evaluation, ai-literacy, wordpress-ecosystem] > author: Headless | Show (host: Hayden) > source: Headless | Show (YouTube) > reference: https://www.youtube.com/watch?v=3ONEJZ-Pf8Y > date created: 2026-09-21 > > [!info] Generated analysis — not the author's words. Sections below are machine-written > > from the captured source. Only text inside `> [!quote]` blocks is verbatim from the author. > > Do not quote or attribute any other line to the named author. > # 🏷️ Title: WordPress's $8.15m Golden Parachute (Automattic Executive Drama) > ## 🧭 Table of Contents > - [📝 Summary](#-summary) · [📖 Main Content](#-main-content) · [🗒️ Notes](#️-notes) · [📚 References](#-references) · [🗣️ Paraphrase](#️-paraphrase) · [📰 Newsletter](#-newsletter) · [✅ Action Items](#-action-items) · [🧪 Hypotheses](#-hypotheses) · [📊 Methodology](#-methodology) · [❓ Questions](#-questions) · [🔮 Future Research](#-future-research) · [💡 Implications](#-implications) · [🧠 Learning Styles Assessment](#-learning-styles-assessment) · [📊 Accessibility Assessment](#-accessibility-assessment) · [🌍 DEIB Assessment](#-deib-assessment) · [✅ Logical Validity](#-logical-validity) · [🛡️ DARVO Analysis](#️-darvo-analysis) · [🎯 Conclusion](#-conclusion) · [🧱 Refactor Appendix](#-refactor-appendix) · [🧰 Template Library](#-template-library) > ## 🔗 References (in-page) > - Source video: https://www.youtube.com/watch?v=3ONEJZ-Pf8Y — Headless | Show, uploaded 2026-09-21, 17m49s. Host: Hayden (first name given on air). > - Underlying reporting the episode relies on: TechCrunch, "Automattic's interim CEO and legal chief signed reciprocal severance deals during Mullenweg's brief ouster" (2026-09-16) — https://techcrunch.com/2026/09/16/automattics-interim-ceo-and-legal-chief-signed-reciprocal-sev… ; The Repository (Rae Morey), 2026-09-18. > - **The "$8.15m" figure is not attributed on air.** It matches the $8.15 million total in the vault's captured copy of the TechCrunch report on the reciprocal agreements, so it has a source; the episode just doesn't name it. > - Names: the chief legal officer is Andy Missan (captioned "Andy MSAN"), confirmed against the TechCrunch capture. The board members are captioned "Tony Snder," Sue Decker, and "Dun Woody"; the TechCrunch capture doesn't name them, so those spellings are unverified. > ## 📝 Summary > **What happened.** Headless follows up its earlier episode on Automattic CEO Matt Mullenweg's roughly 33-hour removal and return. The new development: during that window, two executives — CFO Mark Davies, who served as interim CEO, and chief legal officer Andy Missan — reportedly signed each other's severance agreements, each providing about a year of salary and additional equity vesting if their departures qualify. > **The episode's discipline.** The host repeatedly separates what was *signed* from what was *authorized*. Two signatures don't show who approved the arrangements or whether the board knew; "it looks suspicious" and "it wasn't authorized" are different claims, and only one can be established by evidence. The validity of the agreements is unresolved. > **Other developments covered.** > - The Repository (2026-09-18) reports, from one anonymous employee, that questions in internal Slack went unanswered and an anonymous discussion space closed. The host flags it as a single source. > - TechCrunch reports two board members resigned and one was removed; on 2026-09-18 Automattic's public board page still listed all three alongside Mullenweg. The page briefly showed Mullenweg alone, which a spokesperson blamed on "an overzealous employee" before reverting it, with no board update. > - A Reddit commenter posted an AI chatbot's explanation of the executives' motives. The host declines to repeat it and explains why a generated account of private motives isn't evidence. > **The practical takeaway.** For site owners and agencies: identify which part of the stack you depend on — WordPress software, WordPress.com, a host, a plugin vendor, a support agreement — and ask providers whether anything has actually changed, instead of treating governance headlines as technical failures. > **Why it matters.** It's a clear example of reporting on a fast-moving, emotionally charged story without inventing what isn't known. The section on AI-generated explanations of motive is especially useful. > **Risks.** The episode's title uses a dollar figure it doesn't attribute on air (it traces to TechCrunch). It relies on secondary reporting and Reddit comments. The comedic "office scenes" are invented, and the host says so. > **Bottom line.** A careful governance explainer whose core lesson — signatures aren't approvals, and a well-formatted theory isn't evidence — is worth more than the details, which are still unresolved. > ## 📖 Main Content > The cold open sets up the joke — "This isn't a birthday card. You sign mine, I sign yours" — against $8.15 million of potential executive compensation. Hayden defines a golden parachute (lose your job, keep getting paid, subject to contract conditions) and states the question: during Mullenweg's 33-hour absence, two Automattic executives reportedly signed each other's severance agreements. "The signatures are one thing. Who could authorize them is a completely different question." > He then fixes the frame. Automattic runs WordPress.com; corporate upheaval isn't evidence that the WordPress software stopped working. The report concerns CFO Mark Davies, who served as interim CEO, and the chief legal officer. The agreements reportedly provide a year of salary plus additional equity vesting if departures qualify. The timing raises an obvious question — were they preparing for the whole thing to blow up? — but Hayden labels that a question about their expectations, "not an answer we have." What he wants: who could approve it, what was approved, and what the documents say. > A comic sketch contrasts a parking-receipt approval with executive paperwork, making the point that ordinary employees experience corporate approval as friction and are naturally curious how it works at higher stakes. "It looks suspicious and it wasn't authorized are two different sentences. Only one can be established by watching me raise an eyebrow." > The Reddit section models how to read comments. One user dislikes Mullenweg *and* doubts the agreements will stand; Hayden notes their legal prediction is opinion and their suggestion of missing board approval isn't established. A reply — "who knows if the board knew or approved" — names the missing step. He criticizes the internet's habit of turning every story into a casting exercise, choosing a villain and fitting new facts to the script: "It's a fact. It doesn't have an agent." You can dislike a CEO and still ask whether other executives' arrangements were properly authorized. > An imagined investor update ("some changes to the changes") gives way to a commenter's sympathy for Automattic employees, and then to The Repository's reporting: one anonymous employee said Slack questions went unanswered and an internal anonymous discussion space closed; Mullenweg answered the reporter initially but not the follow-up about why the board had placed him on leave. Hayden cautions that this is one employee's account and that "Automattic refused to comment on everything" would overstate the article. On sourcing, he agrees with a commenter that one anonymous source can be worth reporting but "isn't the employee census." There's no evidence of a wave of resignations; retention is a question, not an outcome. > He then turns to the board page — "the boring part" that turned out not to be. TechCrunch reports that two board members (captioned "Tony Snder" and Sue Decker) resigned and a third was removed. On September 18, Automattic's public board page still listed all three with Mullenweg. The Repository reports it briefly listed Mullenweg alone, which a spokesperson attributed to an overzealous employee and reverted, with no board update. Hayden is careful: the page shows what the website displayed, not who legally holds a seat. A company closely associated with publishing is handing readers the homework of reconcilin This isn't a birthday card. You sign mine, I sign yours. Sounds very different when you're discussing $8.15 million of potential executive compensation. >> Off with its head. This is Headless, the weekly briefing for people building, buying, breaking, and relentlessly replatforming the modern web. >> A golden parachute. It's the executive dream. lose your job, keep getting paid. Subject to the contract's conditions, of course. For every normal employee begging for a raise, it's a workplace wet dream. But imagine the executive signing your exit package needs you to sign theirs. During Matt Mullenweg's 33-hour absence, two automatic executives reportedly signed each other's severance agreements. The signatures are one thing. Who could authorize them is a completely different question. This isn't a birthday card. You sign mine, I sign yours. Sounds very different when you're discussing $8.15 million of potential executive compensation. I'm Hayden. This is Headless. Last time, a CEO removal, a return, and a boat. This time, the paperwork. Automatic is the company behind WordPress.com. The corporate upheaval isn't evidence that the WordPress software stopped working. Watch the difference between what was signed and what we know was authorized. One Reddit reply puts that distinction better than an entire comment war. The new report concerns Mark Davies, the CFO, who served as interim CEO, and Andy MSAN, the chief legal officer. The linked summary describes reciprocal agreements providing a year of salary and additional equity vesting if the departures qualify. 33 hours. You had 33 hours. The CEO goes on leave. Two executives sign each other's exit deals. Then the CEO comes back. It raises a pretty obvious question. Were they already preparing for the possibility that this whole thing would blow up? That's a question about their expectations, not an answer we have. But the timing is exactly why the paperwork deserves a very close look. The validity of the agreements remains unresolved. I want to know who could approve it, what was approved, and what the relevant documents actually say. The signatures are just part of the story. They don't provide every answer on their own. Imagine explaining a much smaller version of this at a ordinary office. You submit a parking receipt. Finance wants the date, the purpose, and confirmation from your manager that the car was being used in a manner consistent with the company's values. You were in the parking lot. Yes, but whose decision was that? Then the subject changes to executive arrangements and suddenly the conversation requires people with much more expensive email signatures. This is why stories like this get a reaction. Most people experience corporate approval as a thing that prevents them buying, you know, a replacement keyboard. They are understandably interested in how it operates when the stakes are substantially larger. If there's an approval record, let's see it because it looks suspicious and it wasn't authorized are two different sentences. Only one can be established by watching me raise an eyebrow. Now, it's time to jump over the comment section already because here is where two people looking at the same story reach different conclusions. On the subreddit WP drama thread, user Fatbook Air, great name, makes a useful break from the usual team sport. They express strong dislike for Mullenweg while also doubting the agreements will stand. Now, their legal prediction is an opinion. Their suggestion about missing board approval isn't something we've established. User struggle usual replies, "It just says each exec signed, but who knows if the board knew or approved." That is the missing step. Two signatures don't give us the full approval history. Obviously, you can find an arrangement suspicious and still need the document that establishes what actually happened. You can dislike a CEO and still ask whether another executive's arrangement was properly authorized. You can question the arrangement without endorsing everything the CEO has ever done. Corporate governance should survive the shocking possibility that more than one person deserves a difficult question. The internet makes this harder than it needs to be. Every developing story becomes a casting exercise. You know, pick the villain, assign everyone else a supporting role, and then evaluate new information by how well it respects that screenplay you've written. When an inconvenient fact arrives, you ask, "Could we give that line to somebody else?" No, it's a fact. It doesn't have an agent. I don't need a sympathetic protagonist to follow this story. I need a reliable account of who could do what and what happened when they did. Those questions can lead to uncomfortable answers for different people. That's allowed. We are discussing a company not choosing who gets the last rows. And if you'd like more coverage that can enjoy the absurdity without pretending every unanswered question has already been settled, subscribe on YouTube or follow Headless on Apple Podcast or Spotify. Your subscription does not require you to declare allegiance to a certain executive. We've simplified the onboarding. You just press a button. Now, another comment gets to the emotional center of this beautifully. User Alphaax writes, "Imagine being an investor and watching this go down. Just imagine the investor update. To be clear, this meeting is happening in my head right now. I'm clear that it's probably different in real life, but you join hoping to discuss your customers and revenue, maybe a product launch. Before we begin, there have been some changes." Fine. Companies change and some changes to the changes. You close your spreadsheet. We have prepared a timeline. Excellent. It's being updated. You open the spreadsheet again because numbers feel safer than real life now. Somebody suggests moving on to the product road map. Let's discuss 7.11 for instance. You would love to. You have never wanted to hear about a minor interface improvement more in your life. But the reply to that investor comment changes where the camera should point because user hedgehog named Sonic writes, "Imagine working at A8C during all this. Heart goes out to those folks." Now A8C is internet shorthand for automatic. The commenter is expressing sympathy, you know, not giving us an inside account, and the people doing the ordinary work belong in this story, too. And now there's reporting about what that uncertainty feels like inside the company. On September 18th, Ray my at the repository reported that an anonymous automatic employee said questions in Slack had gone unanswered and anonymatic and an internal anonymous discussion space had closed. The source described anger and frustration and I could only imagine. Now that's only one employees account. We haven't independently verified it, but the repository is known for good sources and it doesn't tell us what every employee thinks. Obviously, the article also says Mullenweg initially replied to the reporter but didn't answer the follow-up asking why the board had placed him on leave in the first place. So, Automak refused to comment on everything would overstate what this article establishes. But if you're inside that company, the question isn't whether the internet has picked the correct villain. It's whether you can ask what's happening at work and get a useful answer. Apparently, we're all adults until someone asks a question in the all hands. Then suddenly, it's a spoiler for season 2. And yes, the Reddit thread immediately gets into the sourcing. User Brian Hudson writes, "Source as confirmed by the article, one automatic employee. User friendly web guy replies, you might think that only having one source is unreliable by default, but that's not necessarily true. More is always better. Certainly, both things can fit in the same sentence. One anonymous source can be worth reporting. It still isn't the employee census. Obviously, I want the sourcing label on the story. I don't want to use it as a blindfold. We don't have evidence here of a wave of resonations. Retention is a question this raises, not an outcome we can announce as of right now. And the executive parachute metaphor starts to feel a little different when the people in the cabin are asking for a flight update. I'm imagining an employee trying to just get through last Tuesday. Any questions before we get back to work? Uh, yes. Are we still doing that project? We're excited about the future. Wonderful. Is the uh project in it? People can work for a company without being personally responsible for every executive decision. The boardroom gets the headline. Somebody else still has a support cue. That is the cost. I'm interested beyond the headline number. Attention spent establishing who's responsible before people can discuss the work. We all know people that work for companies. We all know how hard it is right now. I can't give you a measured financial impact from this reporting. Obviously, I I can't tell you what question I'd ask a customer or a partner. And that is where do I get a current authoritative answer about the part of the company my business relies on? Which brings us to a small piece of the internet with an unusually large job this last week, the board page. So, let's try the obvious place to check who's in charge. The company's own website. This should be the boring part. Techrunch reports that Tony Snder and Sue Decker resigned from the board and Dun Woody was removed. When checked on September 18th, Automatic's public board page still listed those three people alongside Mullen Wick. That page tells us what the website displayed, right? It doesn't by itself settle who legally holds a board seat, which is an important distinction and a remarkable sentence to need in a story about a company so closely associated with publishing things on the internet. I'm imagining the person has signed the update, right? Hey, can you fix the board page? Absolutely. Who should be on it? Use the approved list. Uh, who approved the list? The board. Oh man. They put the request in a folder called dependencies and go to lunch. There's now reporting on the page changing and changing back too. The repository says it briefly listed Mullenweg alone. According to the article, an automatic spokesperson attributed that change to an overzealous employee, said it was reversed and offered no board update. Classic. Even the correction needs a correction. Please do not give this page a version number. We cannot afford another release cycle here. We have enough to discuss without making the screenshot carry the entire company constitution. A public web page can be out ofd. A report can require a correction. A document can't answer one question while leaving another open. If we want to criticize a company for confusing communication, our own explanation has to distinguish those things. The practical problem is that people need somewhere to get an answer. they can use. And when the visible sources appear inconsistent, the work moves to the reader. Now they have to decide which date, which update, which description. For a company that helps other people manage content, that is an unusually awkward piece of homework to hand out. Now, this episode has no paying sponsor. Today's imaginary sponsor is boring as a service. Now with a governance edition. Introducing the current list. The current list tells you who currently holds a role. Each entry includes a name, an effective date, and a person who can answer a follow-up question. Does it use AI? It uses someone checking. Could we say AI assisted? You could, but then we'd have to check that, too. Can we have an animated transition when a name disappears? The legal team has asked us to stop this demonstration now. Enterprise customers receive a second feature, an old list that is clearly labeled old. We know two lists you're going to want to sit down for this. Boring as a service, our most disruptive feature is an effective date. If you'd like to support the actual show, subscribe or follow. Our current list of paying sponsors remains very easy to maintain. Now, the strangest source in the discussion, it can explain the whole thing. It wasn't there. On September 17th, we captured a comment by user TD Sizzle introducing an answer from Claude. The comment has since been deleted. Our capture preserves the opening. For what it's worth, ask Claude why they did it. The generated explanation we read at the time went on to describe supposed reasoning and conversations behind the events. I'm not going to perform those imagined conversations as if they were reporting a generated account of someone's motives needs evidence just like an account written by a person. But then user replies, "It's not worth much." And there it is. The peerreview process has arrived wearing a Reddit username. I understand the impulse. The story is confusing. You want somebody to put the pieces in order. The model can help summarize material you give it, identify competing possibilities, and suggest questions that you can investigate. But when the question is why particular people made a private decision, a convincing explanation can exceed what the supplied evidence establishes. Just imagine asking a chatbot why your boss scheduled a meeting. It gives you a five paragraph account of the secession plan, the budget crisis, and the private conversation that changed everything. Your boss wanted to move the meeting to Thursday. The chatbot has already cast the limited series. The problem isn't that every generated explanation is wrong. The problem is that the explanation can sound complete before the evidence is. A source says an agreement exists. A commenter asks what might explain it. A generated answer supplies a motive. Someone else remembers the motive and forgets the question mark. Now we're discussing what everyone supposedly knew in a conversation nobody has shown us. And if somebody repeats it tomorrow, the repetition will feel like corroboration, confirmation bias. It isn't. We have made a photocopy of the uncertainty and given it better formatting. I use tools to help research, but the sentence still needs a source underneath it. If I can't show you where a claim comes from, better formatting doesn't make it evidence. I realize that's a difficult thing to say on a show where my entire business model involves confidently arranging sentences. The office scenes are jokes. I told you I was inventing. They don't fill the gaps in the reporting. Which leaves one very, very practical question. What does any of this change for somebody who just needed a website? They're running a publication, selling a service, and building a store, or helping clients do those things. They didn't expect to develop opinions about executive paperwork. For that person, the useful question is which part of the stack they actually depend on. WordPress software, WordPress.com, a hosting company, a plug-in business, uh a particular support agreement are related subjects with different responsibilities. That distinction doesn't make corporate leadership irrelevant. It makes the consequence something we can investigate instead of just assume. An agency could ask its provider whether any service commitments have changed. A customer could check where account access and support responsibilities sit. A person covering the story can ask for current documentation rather than treating every disturbing headline as proof of a new technical failure. Imagine the client's call, right? Should we be worried? You could give them 40 minutes of boardroom lore, or you could answer the question about their actual site, identify what you're checking, and tell them what could cause you to change your recommendation. The first option demonstrates that you've uh been online, and the second demonstrates why they hired you. That's where I want the coverage to land. Keep the drama. It's extraordinary material for me. But bring the viewer back to something they can understand about authority, communication, and the services that they use. What would change the story from here? a documented explanation of the relevant approvals, a confirmed resolution of the agreements, clear current information about leadership, those would add significant evidence and help out a lot here. Another beautifully formatted theory would add just another paragraph. And we have enough paragraphs. You know, we know who reportedly signed the parachutes. We still need the approval history, and a board page cannot notoriize a damn thing. The reason this follow-up matters is that a return to the CEO role doesn't by itself explain every decision around it. The paperwork has created another question about who could authorize what and the public discussion is racing to supply the answer. We can scrutinize that without inventing a private meeting, deciding the legal outcome or choosing a favorite executive. And we can reasonably expect a company connected to so much publishing infrastructure to make its own current information easier to follow. Now tell me what this has changed in your actual work. Did a client ask a question? Did you seek an assurance from a provider? Has a concrete decision changed or is everyone just spending more time explaining the same uncertainty? Leave a comment below or join us at the subreddits headless podcast. Specific experiences are more useful than guessing who secretly said what, subscribe on YouTube or follow on Apple Podcast or Spotify for the next episode. If you missed me covering the removal return and ownership background, watch my previous WordPress episode, Succession Meets Below Deck, that explains how we got here and why the boat was somehow relevant. I'm going to go ask Claude who's updating the board page. Actually, no. That person is having a difficult enough week. Peace. Off with its head. This is Headless, the weekly briefing for people building, buying, breaking, and relentlessly replatforming the modern web., WordPress's $8.15m Golden Parachute, Headless Show, 01m31yvf71h78nj6b45tzhcg3v.