왜 미국에서 많은 사람들이 샘 알트먼을 싫어할까?

The jury stepped into Courtroom 9 of the Oakland Federal Court in California yesterday, with nine people taking their seats as an "advisory jury" to observe a trial expected to last four weeks, ultimately providing a recommendation to Judge Rogers. Today is Tuesday, and opening statements are about to begin.

Just yesterday, on the same day as jury selection, OpenAI announced a newly revised agreement with Microsoft. This agreement eliminated one thing: Microsoft's exclusive license to OpenAI's intellectual property is gone. And this, precisely, was the final lock OpenAI placed on itself when it transitioned to a "capped-profit" structure in 2019.

What Exactly Is Musk Suing For?

Reuters reports and CNBC's trial diary compiled a list of the case two weeks before the trial began. When Elon Musk initially filed his lawsuit in 2024, he listed 26 charges, spanning securities fraud, racketeering (RICO), and antitrust violations. What enters the trial today is only two: unjust enrichment and breach of charitable trust.

The remaining 24 charges were either dismissed by the judge during the motion phase or voluntarily dropped by Musk himself. Just days before the trial started, he proactively withdrew some charges related to "fraud," focusing the case on its core, simplest argument: "OpenAI promised me it would be a non-profit forever," and now it isn't.

For this single statement, Musk is seeking damages of up to $134 billion. According to his complaint, the compensation would all be returned to OpenAI's non-profit arm, but he demands the removal of Sam Altman and Greg Brockman, and the revocation of the entire for-profit conversion process. This is the "true core" of the lawsuit. Its subject is not stock distribution. It's who the "shell" of OpenAI ultimately belongs to.

Judge Gonzalez Rogers has divided the trial into two phases. The first phase determines liability, concluding before mid-May. If liability is established, the second phase determines damages. The jury only participates in the first phase, and only in an advisory capacity. The final ruling lies with the judge. This means that for Musk, winning the "narrative battle" is more crucial than winning "compensation." His goal is to convince the jury that "this company made promises to its donors back then, and then systematically dismantled those promises." If these nine people agree, the judge can piece together the remaining parts.

OpenAI’s strategy is almost a mirror image. Their aim is to convince the jury that Musk's real motive for suing is competitive jealousy, unrelated to any breach of trust. OpenAI's official account fired the first shot on the day of jury selection: "We can't wait to present our evidence in court. The truth and the law are on our side. This lawsuit has always been an unfounded, jealous competitive attack... We finally have the chance to have Musk testify under oath before a California jury."

Note the phrase "have Musk testify under oath." This is a strategy. What OpenAI truly wants is to portray Musk in the public "court" of X as "the founder of xAI who lost to OpenAI." Convincing the judge is secondary. The goal is for ordinary California residents on the jury to enter the courtroom with this lens.

How Were OpenAI's "Locks" Dismantled?

To understand why Musk is so angry, one must first understand the three locks OpenAI placed on itself in 2019, each with a clear design intent.

You'll notice one thing. In 2019, OpenAI was proving to donors: "Even if we make money, there's a limit, and we must stop at a certain point." On April 27, 2026, OpenAI was proving to investors: "We have no brakes."

The explanation for the profit cap is the most direct. In a 2025 letter to employees, Sam Altman wrote, "The 'capped-profit' structure made sense in a world with only one AGI company, but no longer applies when there is competition." In plain English: There are competitors now, so I need to be able to earn more.

The dismantling of the AGI trigger clause is the most subtle. Originally, "achieving AGI terminates Microsoft's commercial license" meant AGI is for the public good, for humanity, and OpenAI wouldn't privatize it. After the rewrite, AGI is to be determined by an "independent expert panel," Microsoft's license is extended to 2032, explicitly "covering post-AGI models," and Microsoft is permitted to independently pursue AGI. This is a version where even the key to "defining who achieves AGI" has had its lock cylinder changed.

The final lock was the exclusive license. Its dismantling happened precisely as Musk's jury was being seated. The decoupling from "OpenAI's technological progress" means that even if OpenAI announced tomorrow that it had achieved AGI, no commercial clauses would be triggered to change it.

Musk's side will argue in court that this was a deliberate dismantling of protective mechanisms. OpenAI's side will argue it was a necessary adjustment in a competitive environment. But there is one thing neither side will dispute: that 2019 "list of self-imposed constraints" no longer has a single item left.

"Scam Altman": Why Do So Many People Dislike Sam Altman?

On the day of jury selection, X was much livelier than the courtroom. Two hours after OpenAI's official account fired its salvo, Musk fired back with seven consecutive tweets. Fast-paced, harsh wording, dense rhythm. A classic Musk rapid-fire barrage. He gave Altman a nickname: Scam Altman.

He also reposted a video clip of former OpenAI board member Helen Toner, where, word by word, she says in the video podcast, "Sam is a liar."

"Sam is a liar." This phrase wasn't first said by Musk. Former OpenAI CTO Mira Murati said it when she left. Ilya Sutskever said it during the "failed coup" to oust Altman. Jan Leike publicly stated it when he resigned along with the entire Superalignment team.

There are actually three groups of people who dislike Sam Altman, each for different reasons.

The first group is the old OpenAI board. Their landmark event was the five-day firing saga in November 2023. The board's stated reason was that Altman was "not consistently candid in his communications with the board."

What exactly did they catch? In May 2024, Helen Toner publicly stated that the board learned from Twitter that their own company had launched a product that would reshape the global AI industry. She also claimed Altman concealed his ownership of the OpenAI Startup Fund, repeatedly stating publicly that he had "no financial interest in the company" until he was forced to admit it in April 2024.

He also allegedly provided inaccurate information to the board regarding safety processes on multiple occasions. Two senior executives reported Altman's "psychological abuse" to the board and provided screenshot evidence of "lying and manipulation." After Toner published a research paper OpenAI disliked, Altman allegedly tried to push her off the board.

The second group is the former safety faction within OpenAI.

In May 2024, OpenAI's "Superalignment team" virtually collapsed overnight. The first to resign was Jan Leike, one of OpenAI's most senior AI safety researchers. His resignation letter posted on X was one of the sharpest exit notes in the English-language AI community that year, stating that "safety culture and processes have taken a backseat to shiny products."

Then came Ilya Sutskever, OpenAI's co-founder and chief scientist, a key instigator of the failed coup. Next, CTO Mira Murati (who briefly took over the company during Altman's firing), Chief Research Officer Bob McGrew, and Vice President of Research Barret Zoph all resigned within the same week. The "non-disparagement agreement" scandal broke out shortly after. Departing employees were allegedly asked to sign such agreements or forfeit their equity.

The third group is the "covenant-keeping faction" of old Silicon Valley. This group is the hardest to define, yet the largest.

They include early contributors like Musk from 2015, early OpenAI employees who genuinely believed in the "non-profit mission," angel investors who bet on early-stage startups in Silicon Valley, and a significant portion of neutral observers who viewed OpenAI as "common human property."

The commonality among this group is that they paid non-monetary costs – reputation, time, trust, social capital – for OpenAI's promises. What they find hardest to forgive Altman for is very specific: every time OpenAI dismantled its own "locks," Altman claimed it was "for the mission."

When the profit cap was removed, he said, "To allow OpenAI to continue pouring resources into AGI research." When the AGI trigger clause was rewritten, he said, "To enable OpenAI to fulfill its mission even after achieving AGI." When the Microsoft exclusive license was canceled, he said, "To allow OpenAI to move towards a broader collaborative ecosystem."

This is why a segment of Silicon Valley is reluctantly siding with Musk in this lawsuit.

The Weight of a Promise in Silicon Valley: Answers in Four Weeks

Having laid this out, you can probably see it clearly now. This dispute is not about money.

Money is OpenAI's business. In 2026, Altman is the CEO of a private AI company with an estimated market cap exceeding $500 billion. He doesn't lack it. Musk, with xAI in 2026, has already reached the Grok 5 era, with Anthropic to chase and OpenAI to surpass. He needs it even less.

What they are fighting over is something only a few long-term participants in Silicon Valley truly care about. Can a non-profit organization that raised funds, accumulated moral capital, recruited talent, and obtained regulatory exemptions under the banner of "humanity's common good" transform itself within a decade into an ordinary, for-profit company jointly controlled by its CEO and VCs?

If this is allowed, then every future AI startup could do the same. "Non-profit" would become a cheap early-stage narrative tool – used to get headlines, pass regulatory hurdles, and attract employees – only to be quietly dismantled once the valuation is large enough.

If Musk wins, Silicon Valley might face a rare sense of embarrassment. It turns out that what you said in 2015 can be dug up word-for-word in 2026, forcing you to testify under oath in a California federal court. If OpenAI wins, the world will continue operating as Silicon Valley has for the past decade: tell a story early, focus on scale later, and in between, systematically dismantle the covenants linking the story and the scale.

Answers will come in four weeks. But the label "Scam Altman" has already been etched into social media, and will persist regardless of the verdict. The root of why so many people dislike Altman is that he made those who believed in him feel deceived. How much money he makes is secondary.

And the feeling of being deceived is not something a verdict can undo.