🚨 breaking newsOpenai launched the operator, an agent of Amnesty International who can use his browser to perform tasks for you. Currently, it is only available for professional users in the United States, but it will soon come in the world. 🌍
Wonderful, right? But steadfastness – Are we sure that web sites will not retreat? 🤔 Will he do the current anti -bot’s technology such as IP ban, browser fingerprints, TLS fingerprints, and of course, keep up with CAPTCAS with the new Openai?
Therefore, who really wins this battle between complex automatic robots and anti -bot defenses? Reading to find out that! 🔥
LLM models and online data: rocky relationship
When LLM models hit the market for the first time, this was not less than the revolution. The way we deal with the daily tasks at work changed forever, the stock market reaction with excitement 🚀, and everyone jumped on the artificial intelligence train (even if there is no TRUE Artificial intelligence behind most of the online products so far).
As always, the first noise faded at the end, and some important questions began to appear. You do not need to be an engineer for automated learning or Kaggle Grandmaster (see for work, we can find us there too! 😉) to know that LLM Tons of data To train.
So, where does all this data come from? Easy answer: Web! 🌍
The web is the largest source of data on this planet, so it is not surprising that companies like Openai collect the Internet for years to collect data needed to train their leading technology. As long as the web is morally scraped, there is nothing wrong with that 🤷.
Professional advice: Take a deep diving on this topic by reading our article on how to stay morally and legally in the era of web scraping from artificial intelligence.
But this is hunting: Most site owners are not happy with artificial intelligence companies using their data! 😠
After all, the data is equal to money 💰. It has passed several years Economic The article “The most valuable resources in the world are no longer oil, but data.” So, frankly, there is no need to explain it.
In short, abandoning your data for free is the same basis as the money distribution 💸. No wonder that the owners of the site – especially the major companies – are not completely happy with that. 😅
Now that the scene develops and the new artificial intelligence operators and tools enter the scene, web sites may start getting truly Not satisfied with that. 😬
Artificial intelligence operators opposite web sites: the next stage of this troubled relationship
In his article on how the operator works, Openai Share:
“The operator works with a new model called a computer use agent (CUA). Combining the capabilities of the GPT-4 and advanced thinking through reinforcement learning, CUA is trained to interact with the user’s graphic interface .
It is clear that although AI, such as Openai, previously built robots that collect data from famous sources to train their models, they now give users to a tool that can interact “in a magically” way with web sites and move them. This is exciting and frightening! 😱
See Openai operator in the presentation video:
Again, from the official presentation article:
The operator can “see” (through screenshots) and “reaction” (using all measures allowed by the mouse and keyboard) with a browser, allowing him to take a web action without the need for customized API integrations.
If you encounter challenges or make mistakes, the operator can take advantage of his self -correction capabilities. When you stumble and need help, it simply controls the user to the user, ensuring a smooth and cooperative experience. “
This is incredibly promising, but it also raises some dangerous concerns. 🤔 What if users began to misuse the operator for harmful purposes? We all had enough robots (such unwanted comments that overwhelm YouTube), and this may quickly lead to a big problem. ⚠
Assuming that OpenAI managed to prevent the operator from performing harmful or unwanted procedures-just as if they were working to prevent Chatgpt from answering dangerous questions-can we really welcome that most web sites will welcome this type of new or automatic interaction that It works with the same AI? 🤖
How artificial intelligence operators work
Before diving in the big question, we left open, let’s first explain the type of reactions we deal with. At the end of the day, if the new artificial intelligence operators are not effective as we think, then why should we get upset even protection against them in the first place? 👀
Bot control is not a joke. Companies such as Cloudflare- WAF (WOF Application Fire), known for their strong anti-BOS-Millions of dollars every year, offer research and development to stay in the foreground. 🤑
currently, American users who pay 200 dollars per month to the highest subscription level in Chatgpt Pro can reach Openai’s OperatorTherefore, everyone has no opportunity to test it. But for those who have? The results are impressive! 🤯
The first users and technology auditors found amazing in automating daily tasks such as:
- Requesting food (yes, he can make decisions automatically like choosing the restaurants you ask from 🍔)
- Responding to users to some social media platforms
- Complete small tasks online, such as filling out polls for rewards
How is this possible? The operator opens a miniature browser window and completes the tasks based on your text claims – just like the ordinary user:
Certainly, the product is still in the “search examination” stage and not perfect. Sometimes, you will need to give it a boost or save it from a failed episode.
Although some Reddit users have expressed complaints – especially given the high price point – it does not deny that this technology is already. rare Even at this stage. Watch it a trip reservation, for example!
➡ The real question now: Will web sites welcomes automation that works with artificial intelligence materials, or will they fight? And if they did, how? ⚔
How to fight web sites against artificial intelligence
Bot and control control solutions are not a new thing-regular sites have been used for years to protect from automated text programs that embody data and interact with their pages. 🚫
If you are interested in these methods, check our web symposium on advanced Bot control techniques:
As you may already know – especially if you have followed our chain on the advanced web scraping – we are talking about:
-
Prices limitations: Tools that restrict the number of requests from the user at a certain time to prevent excess download. They work by blocking IPS.
-
TLS fingerprints: How to track the unique properties of the encrypted browser connection to identify robots. Explore the role of the TLS fingerprint in the web scrolling.
-
Browser: A technique for detecting the features of the device or the unique browser to discover automatic tools.
These initial defenses focus on prohibiting applications from automated tools (such as artificial intelligence operators) before you get a chance to reach the site 🛡.
If these defenses fail, other technologies come into effect. Some examples? User behavior analysis, Javascript and Captcha challenges!
Captchas is especially effective because it is designed to be easy for humans to solve, but it is difficult for robots.
But as artificial intelligence increased and they began to think more like humans, recognition has become more difficult. This is why some wild ideas, such as using video games such as captchas, are presented around them. 🎮
But the real question is – is the final solution against artificial intelligence operators? Let’s dive and discover! 💡
Captchaas: Can artificial intelligence operators really overcome the system?
TL; D: No, not really … 🙅
Since Openai Corporaator has reached the market for testing, users have pushed it to complete the tasks involved on Captcha – moving to social media, filling models, and more.
But as it is mentioned in the OpenAI computer viewer display page, human intervention is still required:
“Although he deals with most steps automatically, CUA seeks to confirm the user for sensitive procedures, such as entering login details or responding to CAPTCHA models.”
Certainly, the thinking engine of artificial intelligence may sneak at times, but more often It fails to fail– With the results that are joyful and frustrated. When you take the test Reddit, Google Maps, Amazon and G2It is closed again and again by anti -bot protection.
Watching artificial intelligence operators has become their crash and burning against Captchas a viral trend. These artificial intelligence tools are floundering on their way through entry attempts, Reddit and X:
Other technology auditors confirm the same frustration: Openai operator is banned by most Captchas.
On the one hand, this is reassuring – Captchas does their work and prevents automatic robots from enhancing chaos. On the other side, We are in a cat and mouse game 🐁 🐈. Anti -Bot technology operators and artificial intelligence operators will continue to develop, and take turns one step forward.
The real losers? Ordinary users! It is possible that more Captchase sites, which makes browsing more painful for everyone. And let us be honest – we all hate Captain. 😩
This battle not only affects artificial intelligence operators, and moral web shows are also discovered in Crossfire. While the sites intensify the anti -bot’s measures, evil textual programs will be blocked unfairly, Make data extract more difficult for researchers, companies and developers.
Fortunately, there is a better way to interact with software sites without Dealing with Captchas and other anti -bots: Drgsing Browser!
The real winner? Browse browser browser!
Openai Operator works to automate regular browsers just like other browser automation tools. But this is the thing-most anti-bot techniques, including cappuers, do not appear because From automation itself. Show Because of how the browser is formed!
Most of the browser automation libraries prepare browsers with ways to display it as a mechanism, and completely defeat the purpose of using a “regular” browser. This is the place where the anti -fun and arrival systems enter. 🚫
Instead of focusing on whether Amnesty International can cross Captchas, the real game change uses the right browser-one Implementation of bulldozing and automation. This is exactly where the Data Data Bright browser comes, packed with:
-
TLS fingerprints reliable To avoid detection
-
Unlimited expansion capacity To extract data on a large scale
-
IP rotation is compact Satisfied with the IP 72 million agent network
-
Automatic re -simulation To deal with failed requests
-
Great powers that solve Captcha This surpasses artificial intelligence operators 🧠
It is not surprising here-Captcha Solever is more effective than the Openai operator. Why? Since it is supported by years of development from the same team that dealt with the interruption of the data of senior economic officials in minutes. ⚡
Captcha Solver from Bright Data has proven successful against:
- Recaptcha ✔ (Yes, the Openai operator could not solve it in a tweet above)
- HCAPTCHA ✔
- PX_CAptcha ✔
- Simplecaptcha ✔
- Geetest Captcha ✔
- … and many others!
Not only Reducing the chances of CaptchasBut when you appear, that He solves them without trouble. 🔥
Dragting Browser works with all the main browser automation frameworks – including theatrical writer, brides and selenium. So, whether you want full control of the program or Even to add the logic of artificial intelligence on topYou are covered.
See a browser browser browser at work:
So … should we continue to force AI to solve Captchase, or just use a tool? The choice is clear. Scrape the ftw browser. 🏆
Final ideas
The Openai operator is working here to revolutionize the web interaction-but it is not strong. Although it is impressive, it is still struggling against Captchas and is banned.
Avoid trouble with a scraping browser, and is characterized by the solid CAPTCHA solution for uncompromising automation. The start of our endeavor to give the democratic character to the web, while making sure that it is still available to everyone, everywhere, even through automatic textual programs!
Even next time, keep the Internet freely and without Captchas!
