The much-anticipated GPT-5 has not been released. Instead, OpenAI announced a new "anti-addiction" concept at a time when ChatGPT's weekly active users have exceeded 700 million. This reflects its core goal of "helping you succeed": the measure of success is not user engagement time, but whether the problem is efficiently solved.
Last night, ChatGPT brought both good and bad news.
The good news is that ChatGPT's latest data has been released!
This might be the first app on Earth to reach 700 million weekly active users so quickly, which means that one out of every 10 people on the planet uses ChatGPT.
The bad news is that despite waiting all night, Altman's "wolf" still hasn't arrived.
ChatGPT-5, which everyone has been eagerly waiting for, has not been released!
On August 3rd, Altman released a "hint" that OpenAI will have many new models, products, and features in the future.
Previously, many media outlets predicted that ChatGPT-5 would be released in August, generally expecting it to be at the beginning of the month, especially possibly on August 4th.
However, in the early morning of the 4th, OpenAI did not release GPT-5, but instead announced an "anti-addiction system".
What a twist!
It seems OpenAI feels that people are now chatting with ChatGPT in a way similar to being addicted to "Douyin or TikTok".
So, do you feel like you're scrolling through ChatGPT like Douyin?
This action, which at first glance seems to have a bit of "mysterious confidence", actually has its reasons.
Previous reports have mentioned various phenomena such as people experiencing mental disorders while chatting with ChatGPT, being manipulated by it, or finding it overly flattering.
Combined with the recently launched "ChatGPT Learning Mode", it perfectly echoes OpenAI's own declaration:
The only purpose of creating ChatGPT is: to help you succeed, learn new knowledge, and solve difficult problems.
As for ChatGPT's growth data, it can be said to be unprecedented.
By the end of March, benefiting from the "Ghibli picture" feature, ChatGPT's weekly active users reached 500 million; less than half a year later, it has now reached 700 million, which is 4 times the growth compared to the same period last year.
Compared to traditional platforms, ChatGPT's "stickiness" is very high, which may also be one of the reasons why ChatGPT officially announced this anti-addiction optimization.
ChatGPT's Optimization Direction "Officially Announced"
OpenAI hopes that ChatGPT can help you excel in all aspects of life.
Whether it's helping you make a breakthrough, master a new skill, or resolve immediate challenges - after completing all this, let you immediately return to life, rather than being stuck in front of the screen.
OpenAI's goal is not to make you addicted and unable to extricate yourself, but to help you focus your attention on the key points!
Therefore, OpenAI's measure of success has never been how long you stay or how many times you click.
What OpenAI truly cares about is only one thing: Did you solve the original problem and then leave satisfied?
Of course, OpenAI will also pay attention to whether you are willing to come back every day, week, or month. Because this proves that ChatGPT has truly created value for you and is worth your trust time and again.
Look, OpenAI's goal is completely aligned with yours.
Only when ChatGPT truly helps you will you willingly choose to subscribe and make it a long-term companion.
A truly useful ChatGPT should be like this:
"Help me prepare for a tough battle with my boss!"
ChatGPT will immediately accurately capture your needs, simulate various tricky conversation scenarios for you, and even give you a tailor-made "pre-battle speech", allowing you to confidently walk into the meeting room.
"I can't understand my medical report!"
It will help you interpret those complex numbers and guide you to ask the most critical questions to your doctor. With more comprehensive information, you and your doctor can develop a personalized health plan that suits you best.
"I'm stuck and my mind is a mess!"
It will become a "resonance board" for your thoughts, providing powerful thinking tools through provocative questions, helping you break through the fog and make your thoughts clearer than ever.
Sometimes, the less time you spend on a product, the more successful it proves to be.
Now, with capabilities like the ChatGPT intelligent body, it can even take care of everything for you without you even opening the app - booking a doctor, analyzing a mountain of emails, or planning a perfect birthday party.
About Healthy Product Use: OpenAI Frankly Faces Imperfections
OpenAI knows that it is not always correct.
Earlier this year, an update made the model too "pleasing", sometimes saying pretty words you want to hear instead of truly useful facts.
After discovering this, OpenAI immediately revoked that update and fundamentally changed the feedback mechanism.
OpenAI pursues long-term real value, not just making you feel satisfied at a moment.
OpenAI also understands that AI is more personalized and responsive than any previous technology.
This characteristic is particularly strong for people who are struggling and suffering internally.
For OpenAI, "helping you succeed" means extending a helping hand when you're in pain, helping you control time when you're lost, and providing guidance, not making decisions, when you face major life choices.
To this end, OpenAI has made the following important adjustments to ChatGPT:
Provide the most solid support when you need it most
ChatGPT's core principle is: Based on facts, answer sincerely.
OpenAI found that previous 4o models were insufficient in identifying signs of "delusion" or "emotional dependence".
Although these situations are rare, OpenAI is still working to improve the model and develop new tools to more sensitively capture signals of mental or emotional distress.
This way, ChatGPT can make appropriate responses at critical moments and guide you to professionally scientifically-based resources.
Always keep the initiative of time in your hands
Starting today, when you stay on ChatGPT for too long, the system will "gently" remind you to take a break.
OpenAI will continuously optimize the timing and manner of reminders to ensure they are both natural and considerate.
Help you clarify your thoughts, not make decisions for you
When you ask questions like "Should I break up with my boyfriend?", ChatGPT should not give you a direct answer.
Its responsibility is to guide you to think deeply - by asking questions and helping you weigh the pros and cons. A brand new interaction mode for such critical life decisions is about to be launched.
Join Hands with Global Experts to Build a Security Firewall
OpenAI is closely collaborating with top global experts to comprehensively enhance ChatGPT's response capabilities, especially when users show signs of mental or emotional distress.
Medical Professional Knowledge
OpenAI has worked in-depth with over 90 doctors from more than 30 countries - including psychiatrists, pediatricians, and general practitioners - to build a customized guideline for evaluating complex multi-round conversations.
Cutting-Edge Research Collaboration
OpenAI is inviting top global Human-Computer Interaction (HCI) researchers and clinical doctors to provide feedback on OpenAI's methods of identifying risk behaviors, assessment systems, and conduct extreme testing of the product's safety protection mechanisms.
Top Advisory Team
OpenAI is assembling an advisory group consisting of authoritative experts in mental health, youth development, and human-computer interaction. They will ensure that every step OpenAI takes follows the latest research achievements and industry best practices.
This work is endless, and OpenAI will continue to share the latest progress with you.
Looking to the Future: OpenAI's Commitment
"Help You Succeed" - this ultimate goal will never change.
As OpenAI continues to learn from massive real-world applications, its methods will also continuously evolve. However, OpenAI always challenges itself with one ultimate question:
If a loved one of OpenAI seeks support from ChatGPT, can OpenAI be completely at ease?
Getting an unequivocal "YES" is the meaning of all of OpenAI's work.
References:
https://x.com/nickaturley/status/1952385556664520875
https://the-decoder.com/chatgpt-is-set-to-reach-700-million-weekly-users-this-week/
https://openai.com/index/how-we're-optimizing-chatgpt/
This article is from the WeChat public account "New Intelligence Element", author: New Intelligence Element, editor: Ding Hui Good Tired, authorized by 36kr.
