The Free Feature

Albert Anker, "The Village School" (1896). Kunsthaus Zurich. The students learn what the teacher puts in front of them. What the teacher removes, they pay for elsewhere. Public domain.

Google shipped a free screen recorder with face cam overlay to every Android phone in August 2015.^[2] The YouTube Gaming app let any user tap a single button and begin recording their screen with a circular video of their face superimposed on the footage. It captured audio from the microphone and the speaker. It included basic editing: trim, effects, background music. It could live-stream directly to YouTube. It was free. It had no ads.^[3]

The feature was not buried in a settings menu. It was the app's signature capability — one that no competitor offered in a free, integrated, zero-setup package. Twitch did not have it. No mobile platform matched it. For two years, anyone with an Android phone had a production-ready screen-plus-face recorder in their pocket, at no cost.

Google announced the discontinuation of the YouTube Gaming app in September 2018.^[4] Ryan Wyatt, YouTube's head of gaming, said the app caused "confusion" about what it was for. The features, Google said, would migrate to the main YouTube app.

Most features did migrate. One did not.

The screen recording capability was not ported to the main YouTube app at launch. It was not ported in the months that followed. It was not restored until August 2019, five months after the YouTube Gaming app was fully shut down, buried in YouTube version 14.31 as a quiet update, that required you to have 1000+ subscribers to use.^[5] For five months, a capability that had been free and universal simply did not exist in Google's ecosystem, and when it was brought back, it was only for established YouTuber broadcasters.

That gap is where the money went.

What Grew in the Gap

Three founders launched a startup called OpenTest in 2016, a user testing platform. It failed, bringing in $600 in seven months. But users liked one specific feature: the ability to record their screens. The founders pivoted, built a Chrome extension called Openvid, and relaunched on Product Hunt. They gained 2,500 users on day one. A year later, they rebranded. The new name was Loom.^[6]

Loom's product was a screen recorder with a face cam overlay and instant link sharing. Record your screen. Record your face in a circular overlay. Share the video with a link. That was the product. It was, feature for feature, the YouTube Gaming app's signature capability, repositioned for the workplace.

Kleiner Perkins led Loom's $11 million Series A in February 2019. Partner Ilya Fushman joined Loom's board of directors.^[7] One month later, Google shut down the YouTube Gaming app. By May 2019, YouTube Gaming was fully dead. Google quietly restored screencasting to the main YouTube app in August — five months after killing it.

The capital kept arriving. Loom raised a $30 million Series B led by Sequoia Capital in November 2019.^[8] By 2021, the company was valued at $1.53 billion. Atlassian acquired it in October 2023 for $975 million.^[9]

Today, the feature that Google gave away for free costs $12.50 per user per month on Loom's Business plan. The free tier limits you to 25 videos at five minutes each, in 720p.^[10] The feature that was once free, universal, and unlimited is now metered, gated, and monetized.

The Seating Arrangement

Here is the part that requires careful attention.

Kleiner Perkins invested $12 million in Google in June 1999. It was one of the most consequential venture investments in history. John Doerr, Kleiner Perkins' most prominent partner, joined Google's board of directors. He remains on Alphabet's board today, twenty-six years later, serving on the Leadership Development, Inclusion and Compensation Committee.^[11]

Kleiner Perkins led Loom's Series A in February 2019. Ilya Fushman, a Kleiner Perkins partner, joined Loom's board. He told interviewers that Loom's users were "just using it like crazy" and compared Loom's freemium trajectory to Dropbox's, where he had previously been head of product.^[12]

This means that Kleiner Perkins simultaneously held:

1. A board seat on Alphabet, the parent company of Google, which operated the YouTube Gaming app and decided its fate.
2. The lead investment position and a board seat at Loom, the startup whose entire product category opened up when Google killed that app.

None of this is an allegation of conspiracy. It is a description of a seating arrangement. The same firm sat at both tables. The question is not whether anyone picked up a phone and said "kill the app so our portfolio company can grow." The question is whether the structure requires that phone call at all. When you can see the cards, you do not need to cheat. You just need to bet at the right time.

When a venture firm sits on the board of the company that digs the hole and leads the investment in the company that fills it, the structure is the strategy. No phone call required.

The Precedent They Already Set

Joseph Wright of Derby, "A Philosopher Lecturing on the Orrery" (c. 1766). Derby Museum and Art Gallery. The mechanism, once visible, cannot be unseen. Public domain.

The conflict has been noticed before.

Kleiner Perkins invested $150 million in Twitter in December 2010, at a $3.7 billion valuation. John Doerr expected a board seat. He did not get one. Twitter's leadership explicitly denied Doerr the seat because of his directorship at Google. Google was, at the time, Twitter's most likely acquirer and its most dangerous potential competitor in the social space.^[13]

The conflict was recognized. The governance mechanism worked. The dual position was blocked. Ten years later, when Kleiner Perkins led Loom's Series A while Doerr sat on Alphabet's board, no equivalent mechanism intervened. No one asked whether a firm with board-level visibility into Google's product roadmap should be funding a startup positioned to capture the market that Google was about to vacate.

The difference between 2010 and 2019 is not that the conflict disappeared. It is that nobody was looking.

The Law That Almost Applies

The Clayton Act's Section 8, passed in 1914, prohibits interlocking directorates: a single person serving as a director or officer of two competing corporations.^[14] For a century, the statute was a dead letter, rarely enforced. Then, starting in 2022, the Department of Justice revived it.

Assistant Attorney General Jonathan Kanter announced an enforcement initiative targeting interlocking directorates across multiple industries. By March 2023, the DOJ reported that its investigations had led to more than a dozen directors resigning from boards. Kanter stated explicitly that the department was "continually looking for interlocking directorates imposed by private equity, venture capital, and corporate venture capital firms."^[15]

The DOJ's "deputization theory" extended the statute's reach: a venture firm can violate Section 8 even when different individuals from the same firm sit on competing boards. The firm itself is the "person" creating the interlock.^[16] The FTC, in its 2023 settlement with Quantum Energy Partners, extended enforcement to non-corporate entities like LLCs and partnerships, and imposed a ten-year prohibition on board appointments without prior FTC approval.^[17]

The legal infrastructure exists. But it has a gap precisely where the Kleiner Perkins structure sits, because the statute applies only when both companies are competitors. Google and Loom are not competitors in the traditional sense. Google is a platform; Loom is a tool. The fact that Google once offered a free feature that overlapped with Loom's entire product does not make them competitors under the statute's narrow definition. The law prohibits two companies from sharing a director when they sell the same things. It does not address the subtler case: a director at Company A who knows that Company A is about to stop selling something, and who uses that knowledge to fund Company B into the resulting vacuum.

Securities law prohibits trading on material non-public information. There is no equivalent prohibition on investing on material non-public roadmap knowledge obtained through a board seat. The corporate opportunity doctrine, rooted in Delaware law, theoretically prevents fiduciaries from exploiting information gained through their board positions.^[18] But here, too, the industry has found its workaround: most venture-backed startup charters now include broad corporate opportunity waivers under Delaware General Corporation Law Section 122(17), which allow directors to pursue opportunities discovered through their board service without liability.^[19]

The legal system recognizes that the dual position is a problem. It has built frameworks to address it. The venture capital industry has contracted around every one of them.

The Pattern Repeats

The YouTube Gaming case is not isolated. It belongs to a class. Google killed Google Reader in July 2013 — the dominant RSS reader with millions of users. Google said usage was declining. Within 48 hours of the announcement, Feedly gained 500,000 new users. Within two months, it had 12 million. Feedly, which had been a modest browser extension, became the market leader in a category that Google had just vacated. The RSS reader market, which had been impossible to monetize while Google offered the service for free, suddenly had a viable business model.^[20]

The pattern: a dominant platform offers a feature for free, eliminating the market for paid alternatives. The platform discontinues the feature. A startup, often already funded and positioned, captures the displaced users. The startup monetizes what was previously free. The investors who funded the startup had, at minimum, advance notice that the market was about to open.

Google has discontinued over 290 products and services.^[21] Each discontinuation is a market event. Each market event creates a funding opportunity. The question is not whether venture firms notice these opportunities. The question is whether firms with board seats at Google notice them first.

The AI Acceleration

Gustave Courbet, "The Stone Breakers" (1849). Destroyed 1945. The workers build the road. Someone else drives on it. The tools have changed. The economics have not. Public domain.

Todd Saunders' April 2026 post describes the same mechanism, accelerated by artificial intelligence. The 2019 playbook was: kill the free feature, fund the replacement, let the replacement grow organically over years. The 2026 playbook is faster: clone the incumbent's software with AI, fund the clone, offer it at 90 percent less.

Saunders was precise about what he saw. A seed-stage company. A well-known VC. A board deck that documented the strategy in writing. The strategy: gain access to a large incumbent's software through a customer's licensed account, use Claude Code to clone the entire product, and undercut the incumbent on price. The VC endorsed this as the go-to-market strategy. In writing.^[22]

Using a customer's licensed access to reverse engineer a product and clone it is ethically bankrupt. I don't know how else to put it. It likely violates terms of service. It may violate trade secret law as well. And a reputable VC putting this in writing in a board deck is genuinely insane.
— Todd Saunders

The ethical bankruptcy Saunders identifies is real. But it is not new. It is the logical endpoint of a structure that has been operating, unchecked, for decades. The dual position — the VC firm that sits on the board of the incumbent while funding the replacement — is the genteel version. The AI clone endorsed in a board deck is the version that no longer bothers to be discreet.

The structural question is the same in both cases: who has the information, who has the capital, and who writes the rules that allow the two to combine?

And the pattern is no longer confined to platforms and startups. In February 2026, TechCrunch reported that at least a dozen venture capital firms — including Founders Fund, ICONIQ, Insight Partners, and Sequoia Capital — now invest in both OpenAI and Anthropic, direct competitors in the most consequential technology race of the decade.^[23] Bloomberg's headline: "VCs Break Taboo by Backing Both Anthropic, OpenAI." TechCrunch called it the death of investor loyalty: "While some dual investors are understandable, others were more shocking, and signal the disregard of a longstanding ethical conflict-of-interest rule." The same table. Both sides. At the top of the AI stack. In plain sight.

What Would Have Changed

Sage.is uses the AGPL. That is not incidental to the argument.

If the YouTube Gaming app's screen recording code had been released under the AGPL, anyone who modified it and served it over a network would have been required to share the source. The code could still have been discontinued, but the capability could not have been locked away. Any developer could have taken the open-source code and built a competing product. The market would have remained open — not captured by the startup with the best-connected investors, but available to anyone willing to build.

If the incumbent software that Saunders' VC-backed clone targets had been AGPL-licensed, the clone would have been required to release its source code under the same licence. You can still compete. You can still undercut. But you cannot lock your modifications away from the community. The improvements come back.^[24]

The AGPL does not solve the governance problem. It does not prevent dual positions. It does not reform Section 8 of the Clayton Act. But it removes the most profitable outcome of the dual position: the ability to take what was free, make it proprietary, and charge for it. Under the AGPL, the code stays common even when the platform does not.

The Game Is Fair. The Seating Is Not.

Kleiner Perkins' investment in Loom was legal. The YouTube Gaming shutdown was a legitimate product decision. The timing — Series A in February, shutdown in March, screen recording not migrated for five months, and then only given to users with over a thousand followers — is circumstantial. No regulator has investigated. No lawsuit has been filed.

But the system does not require illegality to be broken. It requires only a structure where the same people profit from both the death and the replacement. Where board-level visibility into one company's roadmap informs investment decisions at another. Where the law recognizes the conflict in theory but the industry has contracted around every protection in practice.

In 2010, Twitter recognized the conflict and denied John Doerr a board seat. That was the governance mechanism working. In 2019, no equivalent mechanism existed for Loom. In 2026, the mechanism has evolved: the VC does not need board-level roadmap visibility when it can hand a founder an AI tool and a customer login.

The cards are the same. The game is the same. The seating arrangement is what changes. And the people who arrange the seats are the same people who deal the cards.

Section 8 of the Clayton Act prohibits interlocking directorates between competitors. It does not prohibit interlocking directorates between a company and the startup funded to replace the feature that company just killed. The law addresses the conflict it can see. The venture capital industry operates in the conflict it cannot.

Disclosure

This article examines the structural conflicts of interest created by venture capital firms holding board positions at companies with overlapping strategic interests. Sage.is AI-UI is AGPL-3 licensed. Todd Saunders is quoted from a public post on X. No individuals named in this article were interviewed. Kleiner Perkins, Alphabet, Loom, and Atlassian were not contacted for comment. The author holds no financial position in any company discussed.

The views expressed are those of the editorial board. Sage.is AI-UI is a product of Startr LLC and uses AGPL-3 licensing. Todd Saunders is quoted from a public post on X. No individuals named in this article were interviewed or contacted for comment. Full disclosure and transparency is a feature, not a bug.