Shulou(Shulou.com)11/24 Report--

After the popularity of generative AI, there is another AI who can order his own pizza? Here comes HyperWrite!

Imagine that on a sunny afternoon, an AI had just finished outputting a lot of user requests and was bored to open the takeout APP to order something to eat.

Wait, AI? Why does AI order takeout?

Never mind why you order takeout, the important thing is, you can order it.

Today, we are going to introduce such a start-up company-HyperWrite.

Since the explosion of ChatGPT, companies engaged in different types of AI have sprung up like bamboo shoots after a spring rain.

Recently, HyperWrite released an experimental AI agent that can browse the web like a human and even interact with a website.

In the just-concluded demonstration, HyperWrite's CEO Matt Shumer shows how the agent visits the Domino Pizza website through Chrome's extension.

Even, the AI can find its own address, fill in the zip code, and pay to complete the order.

But then again, will AI be hungry?

Tuyuan Pexels, the AI will order pizza, although the demonstration ended before users entered their credit card numbers, but Mr. Schumer said the feature is currently only available to some test users. The function of the model is to automatically handle basic network tasks according to instructions as the user's personal assistant.

If you follow the artificial intelligence section on Twitter, you can see the amazing potential of AI.

The AI session is full of AutoGPT and BabyAGI. These AI are driven by other models, such as OpenAI's GPT-4, and can automatically generate instructions, further develop and execute instructions based on initial input, and further generate new instructions.

Currently, applications based on GPT API are basically limited to single use, which means that the model cannot obtain information from previous interactions.

AI, HyperWrite's personal assistant, is the first user-friendly AI that can operate the browser like a human, Schumer said.

But experts warn that such AI may also have a variety of vulnerabilities, such as phishing, hacker attacks or fraud.

Safety issues can not be ignored, Schumer said, "our team has realized the importance of AI to complete all kinds of tasks on its own." However, this ability can be abused to hack into websites, manipulate communications, etc., and may even spread spam on social platforms. "

Schumer says his team is very concerned about security.

They want to find the right way to deploy HyperWrite on the premise of ensuring the security of the whole society.

For individuals, since HyperWrite can order pizza on its own, will it be out of control to order something else tomorrow?

There is a precedent for such a thing. So Schumer's team is committed to ensuring the security of HyperWrite on the basis of its application.

He explained that agents such as AutoGPT and BabyAGI might be thought of as simple tools to accept text input and provide output.

Users can integrate them into a system that can adapt to new information and improve over time, which is basically cyclical until the desired goal is achieved.

Powerful as it is, GPT is not so advanced that people think it is dangerous.

At present, the application of GPT in real life is still limited.

While it can perform basic research tasks, Schumer's team focuses on using GPT to improve business productivity.

Reference:

Https://venturebeat.com/ai/hyperwrite-unveils-breakthrough-ai-agent-that-can-surf-the-web-like-a-human/

This article comes from the official account of Wechat: Xin Zhiyuan (ID:AI_era)

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.