Starting today, ChatGPT can be used without registration.
Just now, OpenAI opened it hard:
Starting from today, ChatGPT is ready to use, no need to register an account and log in anymore!
Like this, log in directly to the website and start a conversation:
OpenAI’s explanation for the release of this “ultra-free version of ChatGPT” is:
100 million people are using ChatGPT every week. Our purpose of doing this isLet more people who are curious about AI use AI.
However, faced with the sudden opening of OpenAI, many netizens still find it unbelievable:
People say “it’s April Fools’ Day” (US time) and then go to the website to see what’s going on.
But this time OpenAI is really not joking.
Ultra-free version of ChatGPT
Regarding this “ultra-free version of ChatGPT”, it needs to be emphasized that only version 3.5 is open, and the use of GPT-4 is still the old method.
The problem is that if you use it directly,EffectIs it guaranteed?
Let’s take a practical test and answer the questions——The relationship between |ln(1+x)| and |x|.
First, let’s take a look at the answer given by the “Super Free Edition ChatGPT”:
The second is the answer given by ChatGPT after logging in to the account:
Regardless of whether the result is right or wrongjudging from the form of the answer, the “Super Free Version ChatGPT” will give a conclusion and give a simple explanation of the reasons, while the login account version will give a more detailed explanation of the understanding process of the question.
For more differences between the two, OpenAI officials also provided further elaboration.
For example, “Ultra-free version of ChatGPT” willIntroduce additional content protection measuresincluding blocking prompts and spawns in a broader category.
However, regardless of whether you are logged in or not, OpenAI frankly admits that ChatGPT will still use the content of the conversation to train its own model.
But if the user doesn’t like it, he can cancel this function in the settings. The specific operation is to click “?” and then turn off the switch.
Of course, OpenAI officials still encourage users to log in to their accounts to experience ChatGPT, because many functions are not available in the “ultra-free version of ChatGPT”, such as saving or sharing chat records, using custom commands, etc.
Even some very basic functions, such as the multi-language interface and dark/light mode switching that are still in the alpha stage (although it is not ruled out that OpenAI forgot to add them), can only be set after logging in.
Although OpenAI generously offered the wool to users this time, netizens seemed not to be satisfied with this:
When will GPT-4 be like this (manual dog head)?
Why did OpenAI suddenly open?
According to the official statement, it is to “let more people access AIGC.”
We are gradually rolling out this feature to make it accessible to everyone interested in AI.
There is nothing wrong with saying nice things, but it cannot be ruled out that OpenAI noticed a decrease in ChatGPT traffic.I want to use this to save a wave of users, as well as the model iteration brought about by the feedback mechanism behind user interaction.
The latest statistics from Similarweb show that from December last year to February this year, ChatGPT had about 1.6 billion monthly visits.
Although there has been a growth of about 1% in the past three months, the growth rate has slowed down significantly compared with before May, and the absolute number is still nearly 200 million short of the peak period from April to May.
While the growth rate of ChatGPT traffic is weak, the entire AIGC industry has taken off.
Overseas, there is the advent of Google Gemini and the major upgrade of Claude 3. Domestic models are also fighting a “long text war” under the leadership of Kimi, a subsidiary of Dark Side of the Moon. The maximum amount of text that can be read directly reaches one thousand. Million words.
At the same time, various AI search assistants based on large models, such as Kunlun Wanwei’s Tiangong and MiniMax’s Conch, have also sprung up.
Formally, the application methods of large models have become more diverse. For example, Kimi already has a Chrome plug-in (although it is unofficial).
Jia Yangqing’s team also launched a Chrmoe browser AI plug-in called Elmo, which supports web page/PDF summary, can analyze YouTube videos, and can be used for free without logging in.
Therefore, even though ChatGPT is still the traffic king of AIGC, OpenAI still needs to remain competitive as competition becomes increasingly fierce.Without users and feedback data, no matter how excellent the AI model is, it will gradually lose its competitiveness.
OpenAI has to embark on such an Open path.
And with OpenAI’s move like this, it is estimated that there will be another wave of chaos for other large model players. ChatGPT can be used online, and other large AI models must also show more sincerity, right?
One More Thing
In addition to ChatGPT, there are also new trends on the side of OpenAI CEO Altman.
According to the latest information disclosed by relevant U.S. departments,Altman no longer owns or controls the OpenAI Startup Fund.
Control of the fund has been handed over to Ian Hathaway, who has helped manage the fund since its inception in 2021.
The total asset value of the fund now exceeds $325 million, but OpenAI stated that the fund does not involve Altman’s personal investment or financial interests.
This change is probably also related to OpenAI’s controversial accusations of “profit-making” tendencies, especially the accusations against Ultraman.
Although the OpenAI internal strife has ended, there has been no fundamental reconciliation between the idealist faction and the accelerationist faction inside and outside OpenAI. Ultraman’s surrender of control of the fund may be a compromise by the accelerationists. After all, in the internal strife, the accelerationists had a great victory.
