The world’s biggest tech companies are betting big on computers that control themselves

For over a decade, leading technology firms have been striving to create computers capable of managing complex tasks independently. Yet, previous attempts have largely missed the mark, with virtual assistants like Alexa and Siri primarily employed for basic functions such as setting alarms or playing music. Now, major players in the tech industry, including Nvidia, Microsoft, and Google, are optimistic that a significant breakthrough is on the horizon. Recently unveiled innovations—including new chips, laptops, and software—aim to pave the way for a future where keyboards and mice are less essential. These advancements are geared towards enabling AI agents to autonomously execute intricate tasks. "The ultimate aim is to figure out how to simply tell the computer what I want, and have it accomplish that," explained Bob O’Donnell, founder and chief analyst at Technalysis. Nvidia and Microsoft are enhancing Windows to support this vision. Nvidia recently introduced the RTX Spark chip, designed specifically for Windows laptops to run AI agents without needing cloud connectivity. This new chip merges Nvidia's graphics, computing, and networking technologies and features greater memory than typical laptops. Major brands like Dell, HP, and Lenovo are set to launch computers powered by this chip this fall. Google is also making strides with its upcoming Googlebooks, which will be able to suggest actions based on mouse movements, such as scheduling meetings when hovering over a date in an email. For years, tech companies have sought to design digital assistants that simplify mundane tasks like creating shopping carts, managing online orders, and organizing trips. While earlier versions could handle single tasks, they struggled with multi-step processes or personal preferences. The landscape began to shift dramatically with the emergence of large language models, notably following the release of ChatGPT in late 2022. OpenClaw, a burgeoning AI assistant that gained traction among developers this year, exemplifies how AI is reshaping computing habits. This agent can execute programs and fulfill requests with minimal user intervention. Developers have indicated to CNN that they utilize OpenClaw for research on dedicated home computers, checking in on its progress through messaging platforms like WhatsApp or Telegram. Reports suggest that some tech employees are now opting to voice commands to AI agents instead of typing, as noted by Bloomberg and The Wall Street Journal. "The landscape has shifted significantly as more individuals have become accustomed to using ChatGPT, Gemini, or Anthropic," remarked David Naranjo, associate director at Counterpoint Research. At a recent press conference, Nvidia CEO Jensen Huang demonstrated how a laptop equipped with its new chip could assist in architectural design by utilizing AI agents to interact with 3D modeling applications. Moreover, Microsoft is developing a new AI agent for Microsoft 365 named Scout, leveraging OpenClaw’s technology. Scout is designed to operate across various platforms, including cloud storage, local computers, and web applications like Outlook and Teams, with the potential to monitor emails and work communications consistently. Despite these advancements, experts caution that widespread control of computers through simple prompts remains a longer-term goal. The high cost of new laptops may deter many users from adopting this technology, and while companies like Anthropic and OpenAI are rapidly upgrading their AI agents, the justification for purchasing expensive new devices may not yet be compelling. "It hasn't reached a point of being indispensable yet, which presents a challenge for Nvidia, Microsoft, and others," Naranjo stated. While AI applications tend to be more beneficial for businesses than individual consumers, conducting AI tasks locally enhances security and reduces costs for companies. However, trust in AI to manage critical tasks continues to be a significant hurdle. For instance, if an AI agent purchases concert tickets, how can users be sure it won’t misinterpret their budget? "Numerous challenges need to be addressed before this technology becomes mainstream," noted Jitesh Ubrani, research manager at the International Data Corporation. "But are we headed in the right direction? Absolutely."