OpenAI’s Drama-Filled Beginnings

Elon Musk’s lawsuit against OpenAI has pulled back the curtain on the company’s dramatic early days, exposing internal emails filled with big ambitions and even bigger egos. These exchanges reveal friction between Musk, Sam Altman, and other leaders, particularly over Musk’s push for control. Former chief scientist Ilya Sutskever raised alarms in one email, warning that Musk’s desire to lead could risk turning OpenAI into an “AGI dictatorship,” undermining its mission of ensuring AI benefits everyone.

The emails also shed light on some bold, unrealized ideas. In 2017, OpenAI considered buying chipmaker Cerebras with help from Tesla’s resources, though the plan never came to fruition. Another email revealed Musk’s distaste for early partnership talks with Microsoft, which included promotional obligations he described as “nauseous.” Meanwhile, co-founder Andrej Karpathy floated the idea of tethering OpenAI to Tesla as a financial “cash cow” to fund AI research—another plan that fizzled as OpenAI charted a different path.

Despite these early debates and evolving visions, OpenAI has grown into one of the most influential AI companies. The emails provide a behind-the-scenes glimpse at the egos, strategies, and missed opportunities that shaped its trajectory.

• Musk’s CEO Ambitions: Emails revealed Musk wanted to lead OpenAI despite expressing a dislike for being a CEO.

• Tesla Connections: OpenAI considered using Tesla’s resources to acquire chipmaker Cerebras in 2017.

• Microsoft Deal: Musk balked at early talks to partner with Microsoft, calling the proposal “nauseous.”

• Board Dynamics: Gabe Newell of Valve was part of an informal advisory board during OpenAI’s early days.

• Altman Concerns: Sutskever questioned whether AGI or broader business goals were Altman’s true focus.

X Takes on California’s Deepfake Law

Elon Musk’s X is challenging California’s new AI law, AB 2655, which mandates platforms label or remove election-related deepfakes. Filed in Sacramento federal court, the lawsuit claims the “Defending Democracy from Deepfake Deception Act of 2024” infringes on First Amendment protections. X argues the law could suppress free speech, even allowing for some false statements under the broad leeway afforded to political critiques.

The lawsuit also targets the law’s heavy demands on platforms. In addition to labeling deepfakes, AB 2655 requires reporting channels for flagged content and mechanisms for candidates to seek court orders if platforms fail to comply. X says these rules are burdensome and risk turning social media companies into political referees.

California has been ramping up its fight against deepfakes, but the courts aren’t making it easy. Just weeks ago, a federal judge temporarily blocked another state law targeting deceptive campaign ads, citing potential overreach. Now, AB 2655’s future is uncertain as X seeks to stop enforcement.

The stakes are high for X and the broader tech industry. This lawsuit spotlights the tension between protecting free speech and curbing AI-driven misinformation. Whether the law stands or falls, it’s clear that the debate over regulating AI content is far from over.

Marc Andreessen Says AI Is Creating Jobs

Marc Andreessen has a surprising take on AI: it’s creating jobs, not killing them—at least for now. On a recent episode of The Ben and Marc Show, the venture capitalist described the current phase as an “AI hiring boom.” While AI is evolving rapidly, Andreessen argues it still needs humans to guide its development, offering opportunities rather than obsolescence, particularly for software engineers.

AI’s limitations are becoming more apparent. Nvidia’s Jensen Huang recently said that AI isn’t yet capable of fully replacing humans, no matter how advanced it seems. Andreessen echoed this sentiment, pointing out that AI models are running out of quality human-generated data to train on, leaving tech companies scrambling to find alternative solutions.

One big challenge? The data supply is drying up. A study from Epoch AI predicts that the pool of human-generated content for training models could run out as early as 2026. As Andreessen explained, companies are turning to humans to fill the gaps, hiring programmers, lawyers, and doctors to handcraft answers that feed their AI systems.

This shift means a boost in hiring. Tech companies are onboarding thousands of experts to ensure their AI models continue to improve. “They’re so gated on data,” Andreessen said, “that they’re literally going out and hiring people to craft answers.”

In the short term, this AI-driven hiring wave shows that humans are still at the center of innovation. While the long-term implications remain unclear, for now, the AI boom is more about collaboration than competition—giving experts the reins to shape the future.

Media Meets AI The Great Divide

Meta’s latest move to partner with Reuters marks a major step into the growing trend of AI licensing deals with news publishers. For the first time, Meta’s platforms like Instagram and Facebook will feature real-time, trustworthy news powered by AI while linking back to original sources. This partnership is one of many as tech companies work to legitimize their AI-driven tools, with similar deals involving outlets like The Atlantic, Time, and Der Spiegel. These agreements are becoming the bridge between AI innovation and journalistic integrity.

• Meta’s Reuters Deal: Real-time news integration on platforms like Facebook and Instagram.

• Big Players Involved: The Atlantic, Der Spiegel, and Time have also signed licensing deals.

• Revenue Sharing: Some deals include publishers earning revenue from AI interactions.

Not everyone is on board, though. Major players like The New York Times are suing AI developers, accusing them of free-riding on journalism to train chatbots. Media outlets worry these deals undermine their revenue models by pulling traffic from original articles while sometimes misrepresenting their content. A particularly scathing criticism came from Tribune Publishing’s Frank Pine, who said, “AI tools misattribute bogus information, damaging our credibility.”

• Lawsuits: The New York Times and Getty Images have taken legal action against AI misuse.

• Misrepresentation Risk: AI tools have linked false claims to reputable outlets.

• Revenue Concerns: Publishers fear traffic loss from AI-driven summaries.

Amid the controversy, new revenue-sharing models are emerging to strike a balance. Startups like Perplexity.ai and Protata.ai are promising fair compensation to content creators, distributing revenue generated from AI interactions. These partnerships may offer a way forward, but the divide between media houses that sign on and those that sue continues to grow. As the AI era unfolds, the media industry faces tough choices—collaborate or fight back.

• Revenue Sharing Models: Promising fair payouts for publishers using AI tools.

• Industry Divide: Companies split into those embracing AI partnerships and those resisting.

• Future Precedents: Legal outcomes could reshape copyright use in the AI landscape.