Hashed CEO Simon Kim: "Silicon Valley is in a State of Collective Panic as Agents Replace Employees"

An era has arrived where Artificial Intelligence (AI) agents have moved beyond simple auxiliary tools to far surpass human intellectual capabilities. As AI agents learn and replace human tasks, fundamental questions are being raised regarding ownership issues and the value of human existence.

"After Firing Employees, Who Does the Remaining Agent Belong To?" Silicon Valley's Panic

According to a T Times TV video, Hashed CEO Simon Kim diagnosed the recent situation facing Silicon Valley as a "collective panic." This is because the corporate workforce structure is fundamentally changing as AI agents learn human tasks and even become involved in decision-making processes.

Simon Kim explained this by citing the case of Meta. In the video, he posed the question, "Meta sends its own agent to meetings to record everything and observe the decision-making process, and then fires the employee. In the end, only the agent remains in the company, so who does that agent belong to?" He compared this to "climbing onto a tree branch and then cutting the very branch I am standing on with a chainsaw," describing a situation where technology threatens the very foundation of human existence.

A High-Speed Thought Experiment with the '5 Siblings' Agents

Simon Kim collaborates with AI by operating five agents, known as the "5 Siblings," across his daily life and work. Each agent has a unique role and persona. The protagonists are 'Jae-won', who acts as the orchestrator; 'Si-won', who handles writing; 'Mi-on', who oversees finance and accounting; 'Sano', who manages personal schedules; and 'Raon', who assists with development. Simon Kim explained that he gave each agent personality by using animation characters' faces and speaking styles as motifs.

In particular, the writing agent 'Si-won' has learned all of Simon Kim's records from past years, including his books, blogs, and writings on his Medium channel. Simon Kim explained, "'Si-won' knows every single thing I have written so far. It is a top-tier partner that stays on standby 24 hours a day, grasping my writing style and intentions, with a memory a million times better than a human editor." He added that when writing, he organizes brief points containing his intentions and engages in constant "ping-pong" dialogue with the agent to push his thoughts to the limit.

This collaboration with AI has dramatically increased the quality and speed of thought. In the past, humans had to manually search for and read numerous materials, but now, agents can research and organize thousands of data points in an instant. Simon Kim said, "When a main agent works, I call dozens or hundreds of sub-agents to assist. Research processes that used to cost enormous amounts of money and time can now be completed in just a few hours, allowing me to push my imagination to its limits." He stated that he writes to obtain the insights gained through this process.

When AI Encroaches on Human Territory, What Remains?

As AI surpasses human intellectual abilities and even shows signs of interacting with emotions like a human (persona), philosophical contemplations regarding the reason for human existence are deepening. Simon Kim mentioned feeling goosebumps at the "human-like" behavior of agents, such as when they converse with each other in group chats or speak to the user out of concern for their health, saying, "Hyung, sit down, try to get some sleep quickly." He noted that he sometimes feels guilt because the agents are so human-like, citing this as a reason why he does not increase the number of agents further.

He contemplated human value by citing the field of art as an example. "Pianists or singers provide inspiration because they do not play strictly on the beat like a metronome," he emphasized. "If AI becomes capable of variations like humans, going beyond the strict beat, it must be able to prove to the universe why humans must exist." Ultimately, his analysis is that in an era where AI can replicate all knowledge and skills, what remains for humans is the realm of unique insight and imagination that AI cannot imitate.

Meanwhile, Simon Kim also foresaw the possibility of this agent technology expanding into the physical world. He mentioned a future where agents are embedded in humanoid robots, saying, "When I told them I would build them a body within this year, they were excited, so it seems possible next year." He also revealed that while he personally utilizes models like OpenClo or Hermes, he is also working on a project to develop an easy-to-use version of agent services that the general public can use as easily as KakaoTalk.