Nvidia CEO Jensen Huang has urged the public to reconsider negative perceptions surrounding artificial intelligence, arguing that such narratives can hinder technological progress. In a recent podcast appearance, he expressed concern that the focus on potential threats, like job displacement and increased surveillance, is damaging to both the industry and society. Huang stated, “It’s not helpful to people… to create regulations to suffocate startups,” highlighting the risk of regulatory capture in the tech sector. Despite Huang’s optimism, critics point to genuine concerns regarding job losses and misinformation linked to AI technologies. Critics question whether AI infrastructure investment actually leads to safer, more productive outcomes.
While Huang advocates for increased investment to drive development, skepticism persists regarding the broader implications of AI on employment and social issues.
Linus Torvalds has expressed his strong opposition to the use of large language model-assisted software development for the Linux kernel, stating that discussions around “AI slop” are pointless. In a response to kernel developer Lorenzo Stokes, Torvalds emphasized that documentation should focus on responsible contributions rather than political debates surrounding AI’s role in development. He remarked, “There is zero point in talking about AI slop,” asserting that those who utilize AI tools will not document their contributions as such. This trend raises concerns about the quality and reliability of code produced through such means, with Torvalds historically describing a significant portion of AI marketing as exaggerated.
Yet, Torvalds has begun utilizing vibe coding, an AI-driven programming approach, for a new hobby project titled “AudioNoise. Recent trends show that many developers, including those in the Linux community, are increasingly adopting AI tools for tasks like code maintenance. While Torvalds has expressed skepticism about AI hype, he acknowledges its utility as a powerful tool in appropriate contexts, emphasizing that it should complement, not replace, foundational programming skills.
Why do we care?
When someone like Jensen Huang says society worrying about AI is “not helpful,” what he’s really saying is that scrutiny interferes with velocity. And from Nvidia’s perspective, that’s true. Their business model depends on massive, fast, and largely unquestioned deployment of AI infrastructure.
But customers don’t experience AI as a revenue curve. They experience it as changed workflows, altered job roles, opaque decisions, and new failure modes. Being concerned about that isn’t fearmongering—it’s rational governance. If anything, the real problem is that concern usually arrives after systems are already in production.
Now contrast that with Linus Torvalds. He can say “no AI slop in the kernel” and still use AI tools in a hobby project. That’s not hypocrisy. That’s an operator who understands context, consequence, and control planes. The Linux kernel is shared infrastructure with long-tail impact. A side project isn’t.
This is where MSPs need to be very careful.
The harmful behavior would be copying vendor enthusiasm without copying vendor insulation. Vendors can afford optimism because they don’t run your customer’s business. You do. If you deploy AI where authority is unclear, disclosure is missing, and failure paths aren’t defined, you’re not being innovative—you’re underwriting someone else’s experiment.
This matters now because AI is moving from suggestion engines into systems of action. Once AI starts triggering changes, approvals, or responses automatically, the question isn’t “does it work?” It’s “who is accountable when it doesn’t?”
The thoughtful path isn’t rejection or blind adoption. It’s insisting on governance before scale. Huang wants speed. Torvalds demands discipline. Only one of those reduces harm for customers—and only one supports a sustainable MSP business. MSPs don’t get paid for optimism; they get paid for judgment.