← Accueil actu | Salle du marché
2026-09-16 · Source：TechRadar
Surfshark Lead System Engineer Karolis Kaciulis argues that sci-fi doomsday scenarios distract from the concrete problems generative AI is creating today, dismissing extinction rhetoric as a marketing tactic.
Karolis Kaciulis, Lead System Engineer at VPN provider Surfshark, has pushed back against fears that artificial intelligence could drive humanity to extinction, characterizing such apocalyptic warnings as a 'marketing move' rather than a serious assessment of the technology's trajectory. In remarks reported by TechRadar, Kaciulis contends that cinematic doomsday scenarios are pulling attention away from the tangible problems generative AI is already producing in the present day.
His argument lands in the middle of an ongoing dispute over how the industry should talk about AI risk. High-profile figures and laboratories have repeatedly warned that advanced systems could one day pose existential dangers, while critics counter that such rhetoric functions as publicity — inflating the perceived power of the technology and, by extension, the companies selling it. Kaciulis belongs firmly to the second camp, suggesting that treating generative AI as a sci-fi menace obscures the concrete, present-tense harms that deserve regulatory and engineering attention now.
The viewpoint carries some weight given its source. As a systems engineer at a consumer security company, Kaciulis works on infrastructure rather than frontier model development, and his remarks reflect a pragmatic school of thought that favors addressing demonstrable, current damages over speculating about far-future catastrophe. TechRadar's coverage frames his position as a call to refocus: less existential dread, more practical problem-solving around the tools already in users' hands.
For the compute industry, the debate is more than philosophical. Narratives about AI risk shape policy debates, procurement decisions and public tolerance for data-center expansion — forces that ultimately influence demand for GPUs, accelerators and the power infrastructure behind them. If voices like Kaciulis's gain traction, scrutiny could shift from speculative long-run scenarios toward immediate questions such as present-day misuse, carrying its own implications for how AI infrastructure is built and sold. For now, his message is straightforward: the apocalypse can wait; today's problems cannot.
Mots-clés: Surfshark · Karolis Kaciulis · AI Safety · Generative AI · AI Policy
萬安算交所 AIXX · AIXX · Bourse compute vérifiée · Accéder à la salle