Opera’s Neon shows how confusing AI browsers can be

The trick to understanding Opera’s Neon browser is to realize that it’s not just a browser with an AI bot added to it, but a browser with three AI bots all living side by side. This is both a strength and a weakness, because even though you don’t have to leave to do all the AI ​​things you want, knowing where to go for which AI tasks can be really confusing.

Opera has started taking people off its waiting list The artificial intelligence browser, Neon, was released last month. It enters an increasingly crowded market for AI-powered browsers including Google’s Chrome filled gemini, Guilty of confusionand Dia Browser Company. Unlike many of these competitors, Opera charges $19.90 per month, which sets a high bar for a product that most people get for free.

Neon has a similar setup to other Opera browsers. It has features like a built-in ad blocker, a VPN, and a sidebar that you can launch with apps like WhatsApp and Facebook Messenger. AI features are placed front and center on the Home and New Tab screens. Below the search bar is a toggle switch with four settings: regular Internet search; A chatbot called Chat; A browser control agent called Do; and an AI-powered construction agent called Make. (Opera says Neon is built on AI models from… OpenAI and Googlebut does not specify the form used for any part of the neon.)

Chat is the most obvious feature, an AI assistant built into the browser that will likely be familiar to users of any AI chatbot by now. It can be accessed at the beginning of the search query but is also located in the top right corner of the browser for access at any time. Chat easily handled our quick research queries and could answer specific questions about the page we were browsing, such as our request for a summary of the latest research in quantum computing. But it wasn’t perfect. The chat answers were so verbose that we had to make a real effort to get through them. When we asked Chat to count and summarize comments from five of the most recent comments edge Stories, he responded with nearly 400 words to tell us there was nothing.

It was this exercise that showed us that Chat couldn’t always read the Internet, but it gave the impression that it certainly could. Another time, she said there were no comments on three edge Stories when there were actually four. We found it odd that Chat instead offered a guess about what “early comments on tech news sites” often included.

Christian Colondra, executive vice president of browsers at Opera, said the bot failed because we chose the wrong tool. The chat succeeded in observing the number of comments listed at the top of the page edge When Oprah tried it, Colondra said the articles. However, summarizing comments requires expanding the comments section on the web page, which takes one click. Instead, we should have sent our queries to the AI ​​agent Do, Chat’s more active cousin that takes over the browser and completes tasks for you.

We’ve tried a variety of tasks: booking a CrossFit class, booking an under-$50 massage at a nearby spa, and finding PDFs of baby romper sewing patterns. While Do was working, we couldn’t return to the chat in the same window if we wanted to ask follow-up questions about the task at hand. There is also no way to course correct while working. We watched in horror as the robot scrolled through absolutely gorgeous flower arrangements, tasked with finding a friend just to add a giant funeral wreath to our basket, even as we clicked on the better options. Another time, Doe announced that there were no theater tickets for a January show when a quick check revealed plenty of them. It is difficult to trust anything after such constant but misplaced trust.

As with other AI browsers, doing things with Do was also slower than doing it ourselves, though it hints at what crowdsourcing web browsing in general could look like. Using Do does not mean that you can complete the entire payment process yet. Sometimes I faced obstacles that only a human could handle. When this happened, the Action tab at the top of the screen flashed an easy-to-miss shade of red to let us know we needed to step in and help the robot on its way.

In addition to Chat and Do, there’s also an AI Make agent, which can create small web widgets for you. Make resides in a virtual PC where it downloads the required programs, scripts or, in our case, images for your creativity without cluttering your personal computer. We asked for a simple memory matching game with introductory Spanish vocabulary. It worked, and within a few minutes we were able to match the word “libro” to a picture of a book. The game was old, but it was convenient to close the tab and know that all the book images were gone with it.

The final selling point of Opera’s Neon is the cards, which are effectively pre-written prompts that can be used on any of the AI ​​agents that Opera says act like “augmentation modules” for your AI interactions. Cards could, in theory, save us from writing a prompt, but right now, it’s hard to see ourselves needing the option to reuse a set of instructions. The app store-like interface is filled with content from the Neon team ranging from eye-catching prompts that rewrite sites as if Yoda were speaking them to more serious prompts from news aggregators. Opera hopes the platform will be filled with useful user creations as more people use the platform, but there are only a few of them today.

At times, using neon felt a bit like working with an unlucky intern we never asked for, rather than a cutting-edge, time-saving piece of technology. Often, one of its AI systems will ask for feedback, then begin a task without waiting for a response. Given its browser-based capabilities, it’s very easy to imagine where this proactive activity could go wrong, such as sending a large number of LinkedIn requests to people you only wanted to anonymously pursue in a professional capacity. One time, we responded and told her everything looked great and to move on, and Neon said “I’m glad you think so!” It stopped working immediately. Colondra told us that Neon will pause for needed feedback in the future, but the feature is disabled because it’s not ready yet.

Colondra admitted that neon is a work in progress at the moment. “In general, Opera Neon is in an early access release phase, available to people who want to participate in the development journey of this product,” Colondra said. Edge.

But Neon is also a paid subscription product, asking $20 a month, for something that’s largely available for free elsewhere. Right now, that’s a tall order — especially since Neon feels more like an AI browser we need to adapt to than one that’s smart enough to adapt to us.