Vinod Khosla and Marc Andreessen, both founders turned investors, spent part of their weekend debating whether the quest for artificial general intelligence—the idea that a machine could become as smart as a human—should be open source.
The discussion started with A Post Khosla praised OpenAI and Sam Altman, the company’s CEO.
“We know @Sama from the early days of @OpenAI and fully support him and the company,” Khosla wrote. “These cases are a major distraction from the goals of achieving AGI and its benefits.”
Andreessen responded to Khasla’s message Blaming him “Lobbying to ban open source” research in AI.
Andreessen took issue with Khosla’s support for OpenAI because the firm has moved away from its former open source ethos. Since the advent of AI, Andreessen has emerged as a big proponent of open source AI, presenting it as a means of protecting a select few big tech firms and government agencies against the most cutting-edge AI research. Controls access.
Both in this debate and in the past, Andreessen has rejected concerns raised by some of EA’s biggest critics. Andreessen has previously framed these anxieties as fears of disruption and uncertainty rather than the evils of technology itself and a point he reiterated in his turn at X.
“Every important new technology that advances human well-being is greeted with a moral panic,” Andreessen was posted At X. “This is just the latest.”
Khosla, on the other hand, views AI through a geopolitical and national-security lens rather than a strictly business approach. In the past, Khosla has warned that he believes the AI competition between the US and China will escalate into a “techno-economic war”. at the fateAt the Brainstorm Tech conference in December, Khosla said that the US and China’s AI arms race will ultimately decide which of the two superpowers will have greater political influence in the world.
In response to Andreessen’s claim that he was not in favor of open source, Khosla said the stakes were too high.
“Will you open source the Manhattan Project?,” Khosla asked answered to Andreessen. “This is more serious for national security. We are in a techno-economic war with China and AI that must be won. That’s what patriotism is about, not slogans.
And @pmarca Will you open source the Manhattan Project? It is more serious for national security. We are in a techno-economic war with China and AI that must be won. This is actually patriotism, not slogans. https://t.co/AN18VmgVdk
— Vinod Khosla (@vkhosla) March 2, 2024
The preliminary discussion between Khosla and Andreessen saw two opinions on Sam Altman, OpenAI’s lawsuit, and Elon Musk, who at one point recused himself. The debate also explored whether anyone should be allowed to pursue any kind of AI research, or if the most advanced versions of it should be handed over to the government. So while this may seem like just some online sniping among a group of wildly successful Silicon Valley entrepreneurs, it contains a microcosm of the ongoing and critical debate surrounding open source AI.
Ultimately, neither camp wants to ban open-source or closed-source research entirely. But part of the debate hinges on concerns about limiting open-source research being co-opted as a bad-faith argument to ensure regulatory capture for the biggest companies already making headway on AI. – A point that the legendary AI researcher and former chief of Meta. Created by AI scientist Yann LeCun when he entered the field at X.
“No one is asking to ban closed-source AI,” LeCun wrote. “But some people are lobbying heavily for governments around the world to ban (or limit) open source AI.” Some of them call for military and economic security. Others invoke the concept of existential threat.
Elsewhere in Silicon Valley, famous angel investor Rob Conway asked Leading AI companies are “committed to building AI that improves lives and unlocks a better future for humanity.” So far it has added the likes of Meta, Google, Microsoft, and OpenAI as signatories to the letter.
Andreessen, along with Khosla’s interpretation of the Manhattan Project, raised concerns about OpenAI’s security protocol. He believes that without the same level of security that surrounds the Manhattan Project—such as “rigorous security vetting and clearance processes”, “constant internal surveillance”, and “hard physical facilities” with “24×7 armed guards”. – OpenAI’s most advanced research could be stolen by America’s geopolitical rivals.
“In fact, what we see is the opposite—the security equivalent of Swiss cheese,” Andreessen wrote At X. “Chinese penetration of these labs would be facilitated by using any kind of industrial espionage techniques, such as bribing cleaning staff to stick USB dongles into laptops.” My own hypothesis is that all such American AI labs have been fully infiltrated and that China is currently downloading all American AI research and codes.
OpenAI did not immediately respond to a request for comment.
Let’s assume, for the sake of argument, that AI in 2024 is like nuclear technology in 1943, that AI is handled like the Manhattan Project, and that the specific threat is that the Chinese Communist Party gains access to American AI. And let’s use OpenAI as an example…
— Marc Andreessen 🇺🇸 (@pmarca) March 3, 2024
Andreessen, however, appears to be exercising more thought than arguing a point, to write In response to his own post, “Of course every part of this is bullshit.”
Of course, every part of this is unreasonable.
(1) AI isn’t nukes, it’s math.
(2) Large companies, small companies, independent developers, and open source communities should be free to pursue AI.
(3) AI companies should not be nationalized, legal or de facto (protected from government… https://t.co/2VqL6T5qZe
— Marc Andreessen 🇺🇸 (@pmarca) March 3, 2024
Elon Musk entered the debate to criticize OpenAI’s security
At this point, OpenAI cofounder Elon Musk chimed in.
“It would certainly be easier for a state actor to steal your IP,” Musk said answered To Andreessen’s post about security on OpenAI.
Kosli also mentioned musk. to call His decision to sue OpenAI “sour grapes.” Last week, Musk filed a lawsuit against OpenAI, alleging that it violated the startup’s founding agreement. According to Musk, OpenAI’s close relationship with Microsoft and its decision to stop open-sourcing its work is a violation of the organization’s mission. According to a memo obtained by Bloomberg, OpenAI took a similar tack to Khosla, accusing Musk of “regretting not joining the company today.”
musk answered Regarding his departure from OpenAI in 2019, Khosla said he “didn’t know what he was talking about”.
Khosla’s venture capital firm Khosla Ventures is a long-time supporter of OpenAI. In 2019, Khosla Ventures invested $50 million in OpenAI. As such, he did not take kindly to Musk’s trial. “As they say if you can’t innovate, sue and that’s what we have,” Khosla wrote. Xtagging both Musk and OpenAI.
Join Musk now, the discussion continues. Khosla has maintained that AI is more important than the invention of the nuclear bomb and therefore cannot afford to be completely open source – although he agreed with Musk and Andreessen to require its top companies to take more stringent security measures. Needs, even relying on it. To help the government.
“Agreed that national cyber support and protection should be provided and required by all [state of the art] AI, “Kosla wrote. “AI is not just about cyber defense but also about winning globally economically and politically. The future of the world’s values and political systems depends on it.
Despite his reservations about making AI research open source, Khosla said he doesn’t want development to stop. “[State of the art] AI should not slow down because enemy nation states are orders of magnitude more dangerous in my view,” Khosla said in response to Andreessen.
But Khosla and Andreessen found some common ground on the question of AI alignment, which refers to the set of theories, principles, and ethics that inform the model on which AI technologies are developed. Khosla wondered which groups would determine how AI is built, before Andreessen came up with his proposal.
Or the crazy activist mad at a small set of giant Northern California and Washington state tech companies? https://t.co/CD95ChVmY2
— Marc Andreessen 🇺🇸 (@pmarca) March 4, 2024