Author: Kyle Hooten
Topic: AI Fear-Mongering Benefits
Silicon Valley trades on optimism. Even with venture activity relatively depressed since its early-Covid peak, investors are still leaning hard into seed rounds, infusing the least stable startups with capital that values them far beyond an ordinary revenue multiple. For evidence of this, look no further than Y Combinator’s Winter 2025 batch, which included quirky ideas like “Roomba for weeds on a farm,” “GPS satellites for lunar navigation,” and “fusion reactors for [commercial] ships.” Ideas like this that attract investment show that the San Francisco spirit of optimism and a willingness to experiment with unorthodox concepts is still alive and well.
However, over the past few years, a new sentiment has gained popularity in the Bay Area and beyond. Rather than marketing novel technologies as forces for good, driving the human experience to new heights, the current frontier is frequently represented as dangerous. Conversations about artificial intelligence often sound more like discussions on nuclear proliferation than the latest chatbot. Ironically, this discourse is driven in large part by men and women at the forefront of AI development, whose companies stand to benefit the most from a rise in public awareness of their work.
Genuine Concern, or Marketing Tactic?
Things that are dangerous are powerful. Anything that’s powerful can be profitable. Therefore, representing a technology as dangerous can do as much (if not more) than optimism to describe its transformative potential, attracting new users and investment.
This all begs the question: How genuine are concerns over AI’s danger, and where do these concerns come from? Since founders, their backers, and leaders of established companies have so much to gain by driving “AI hype,” as it’s often described, can the public trust their “10-minutes-to-midnight”-style assessments of the current market? Is it possible that the oft-heard warnings about AI’s apocalyptic power are issued, at least in part, for marketing purposes?
“AI doomers,” as legendary venture capitalist and technologist Marc Andreessen describes them, have long been the target of critique in tech discourse. “There is a whole profession of ‘AI safety expert,’ ‘AI ethicist,’ ‘AI risk researcher,’” he observed via a widely circulated post on his Substack in 2023. His view is that these types of figures “are paid to be doomers, [pessimists],” and function as scare merchants for media outlets to cash in on.
Andreessen also identifies another, more insidious, faction he calls the “bootleggers;” people who spread AI fear in hopes they can “financially profit by the imposition of new restrictions, regulations, and laws that insulate them from competitors.”
While widely celebrated by both tech and popular audiences, Andreessen’s years-old assessment of the situation falls short of considering a third, marketing-driven, motive for AI pessimism, which has seemingly emerged recently, since he penned his 7,000 word takedown of artificial intelligence skeptics a few years ago.
Less than a week before Anthropic closed a $3.5 billion raise earlier this month, valuing the company at $61.5 billion, its CEO Dario Amodei literally compared generative algorithms as “easier to build nuclear weapon[s].”
“A new breed of danger has come into play,” he warned, ominously foreshadowing that the public will experience “shock” as the magnitude of this danger comes into focus within the next two years.
Anthropic makes money selling access to its generative chatbot, Claude, which just released a new model (Claude 3.7 Sonnet). What could possibly underscore the potential impact – and therefore value – of Amodei’s growing product line more than comparing it to nuclear weapons, one of the most impactful and civilization altering technologies ever developed? Further, the CEO’s choice to offer a time frame (less than two years) provides a signal to investors and analysts that he sees large advancements or paradigm shifts on the near horizon. Who wouldn’t want exposure to that?
Amodei is not alone. The CEO of Google DeepMind, Demis Hassabis, issued a similar warning in recent weeks. Though not as dramatic as Amodei’s remarks, Hassabis’s recent interview with Axios centered around AI related danger, also concluding that these foreseen harms may manifest on “relatively short timelines.”
Perhaps the most famous living technologist, Elon Musk, is yet another figure who has recently issued dire warnings about AI danger – while profiting from AI development. “We are on the event horizon of the singularity,” he posted on X in late February, shortly after the release of his Grok 3 model. While taken on its own, Musk’s post doesn’t convey much about his feelings regarding the “singularity,” this statement has been interpreted as a warning by many who remember his 2024 statement that a “singularity … [poses] some chance it will end humanity.”
Describing a product as an existential threat to civilization goes directly against traditional ideas of salesmanship. Typically, companies extol only the positive aspects of their creations. Instagram, for example, was sold as ‘a revolutionary technology that stands to change the way people connect with their friends, sharing their photography and important moments from their lives.’ It was not billed as ‘a dangerous network that will forever disrupt the way people socialize, playing a role in millions of young people developing depression, anxiety, eating disorders, and social image issues,’ although this potential was well understood very early in social media’s rise.
However, it seems an inverse dynamic has evolved around AI salesmanship. AI is pitched to consumers and investors in a darkly futuristic ‘join or die’ fashion that runs contrary to the usual Silicon Valley ethos. The heads of these companies aren’t selling the hottest new tech, they’re selling the end of the world, but with the chance for you to be on the team that enters the final launch codes – or at least invest in the team that knows them. Whether or not these claims are accurate, they’re certainly compelling, which warrants a second, slow and sober look at why they’re being leveled.