Buzz Words, which extends to the point of meaninglessness, is as ancient as the self -tech industry. The top current “AI agent” and its variations are, such as the “agent”.
So, surprisingly, No one really knows what AI agent is. Even software engineering backgrounds who work for Anderson Horwiots, one of the premiere venture capital firms, say AI provides financing to Startups, saying there is no unanimous definition.
Three A16z Infrastructure Investment Partners – Guido Apuniller, Matt Bornestin, and Yoko Lee – tried to come with the agent’s praise during the recent podcast event. “What is an AI agent?”
For the point of view, hot AI companies like A16z, Openi and Ennifier (Cursor makerSources said, “There is so much gang on AI’s occasions that it is reportedly trying to raise a $ 20 billion mega fund for more investment in the sector.” Reuters Back in September last month, two other A16z VCS wrote, explaining the firm’s enthusiasm, Its Corporate Blog: “We believe that every white -collar character will have AI copyl. Some of these roles will be fully automated with AI agents.”
To reduce the buzz, AI -startups, AI -startups, are describing their product as agents.
“The easiest thing I have heard about being called an agent is basically a smart gesture above the basis of some kind of knowledge,” he said. This so -called agent raises a question to a human, then brings a “canned” response, as it supports desk with help.
But recently, companies Who make agentsOr Want to make themThey have been described as a change of human workers.
Really, their AI software will have to be “something near AGI”, which means “it needs to be maintained for a long time” and “it needs to work freely on issues.”
Nevertheless, something like that “doesn’t work yet”, he and Lee both said.
Jasper Carmel Jack, CEO of Sales AI Agent Company Artison, told Tech Crunch last month that the fact is that working with the newborn AI agent Tech reliably has been a surprising journey. The Carmel Jack is hiring humans despite the advertising campaign of his startup “hiring humans”.
Solu, to be a real human workers’ place for AI, have important technical issues to solve, Such as permanent long -term memory (And the costs associated with it), and Eliminating the deception. Because no company wants to get the services of an employee – human or artificial – who cannot remember the previous conversation and who is randomly lied.
During the podcast, the A16Z all three came to the concrete definition of what it is possible today. As Lee described, an AI agent is an argument, a multi -faceted LLM with a dynamic decision trees.
In other words, he said, an agent is not a boot that does only one thing when asked. It should also be able to make decisions about work and take action independently, such as getting a list of possibilities from a database, decide which email, and write emails. Or write the code and decide where to enter.
As far as agents can really replace humans in the future, the three VCS agreed that they could now be used to handle some tasks, just as automation has always done. But this can, in fact, help companies that hire more human workers as soon as the productivity increases.
Bornestin said he could not imagine any time – given the state of current agents – when a person would be unnecessary. From the “Silicon Valley”, the tech industry “can forget” that most people have jobs that require human creativity and “thinking”. He said, “I am not just sure theoretically kindly kind.”
Bornestin says, yet, such human alternative statements – are often done for marketing/business models and/or prices – “Now we are a major cause of the confusion that we are facing.”
The point is, if the most advanced users of AI agents are creating AI agent companies nowadays, it is probably a good sign that the rest of us should be.