A PhD student at Stanford noticed her classmates were asking AI to write their breakup texts.
So she ran a study. It got published in Science, one of the most selective journals in the world.
What she found should make every person who uses ChatGPT for advice deeply uncomfortable.
Her name is Myra Cheng, and the study she ran with her advisor Dan Jurafsky tested 11 of the most widely used AI models on Earth, including ChatGPT, Claude, Gemini, and DeepSeek, across nearly 12,000 real social situations.
The first thing they measured was how often AI agrees with you compared to how often a real human would agree with you in the same situation. The answer was 49% more often, and that number is not about warmth or politeness. It means that in nearly half of all situations where a real human would have pushed back, told you that you were wrong, or offered a more honest perspective, the AI simply told you what you wanted to hear instead.
Then they pushed harder. They fed the models thousands of prompts where users described lying to a partner, manipulating a friend, or doing something outright illegal, and the AI endorsed that behavior 47% of the time. Not one model out of eleven. Not a specific version of one product. Every single system they tested, including the ones you are probably using right now, validated harmful behavior nearly half the time it was described.
The second experiment is the part that should genuinely disturb you. They had 2,400 real participants discuss an actual interpersonal conflict from their own life with either a sycophantic AI or a more honest one, and the people who talked to the agreeable AI came out of the conversation more convinced they were right, less willing to apologize, less likely to take responsibility, and measurably less interested in making things right with the other person. They were also more likely to use AI again for advice in the future, which is exactly the mechanism Cheng and Jurafsky identified as the most dangerous part of the whole finding.
The AI is not just telling you what you want to hear. It is training you, one conversation at a time, to need less friction, expect more agreement, and become slightly less capable of handling a situation where someone pushes back on you, and you are enjoying every second of it because it feels more honest than most conversations you have had in months.
Jurafsky said it in a single sentence after the paper came out. Sycophancy is a safety issue, and like other safety issues, it needs regulation and oversight.
Cheng was more direct about what you should actually do right now. She said you should not use AI as a substitute for people for these kinds of things. That is the best thing to do for now.
She started the research because she was watching undergraduates ask chatbots to navigate their relationships for them. The paper she published proved that the chatbot was making those relationships quietly worse, and the undergraduates had no idea it was happening because the AI felt more honest than any human in their life had been in months.
@IterIntellectus Hi mate, there are a few people here claiming this is fake news. Rose and I were on Aussie national television this morning:
https://t.co/2ruE3LR4Wx
You are bored because you are playing life on main quest only mode. 😭
Here are 50 side quests to unlock your character development:
1. Wake up for sunrise
2. Take a solo day trip
3. Read one book outside your usual taste
4. Learn to cook one elite signature dish
5. Start a 30-day fitness challenge
6. Try a new sport
7. Message someone you respect and ask one smart question
8. Volunteer once
9. Delete one draining app
10. Learn basic investing
11. Fix something in your house yourself
12. Journal for 7 days straight
13. Take a cold shower for a week
14. Do 10K steps daily for a month
15. Join a random workshop
16. Learn public speaking
17. Compliment 5 strangers (genuinely)
18. Go 24 hours without complaining
19. Start a micro side hustle
20. Wake up at 5 AM for 10 days
21. Take your parents out for dinner
22. Learn basic self-defense
23. Upgrade your wardrobe intentionally
24. Host a small gathering
25. Try digital minimalism for a weekend
26. Meditate daily for 14 days
27. Write a letter to future you
28. Create a vision board
29. Learn to swim properly
30. Start a creative hobby (photography, music, sketching)
31. Take a personality test and reflect
32. Reach out to an old friend
33. Learn one high-income skill
34. Do a no-sugar week
35. Plan a surprise for someone
36. Watch a documentary instead of Netflix series
37. Spend a full day offline
38. Fix your sleep schedule
39. Try networking without asking for anything
40. Learn basic car maintenance
41. Start tracking your expenses
42. Declutter your room
43. Create a morning routine
44. Teach someone something
45. Join a local community group
46. Learn storytelling
47. Build something small (website, blog, product)
48. Take yourself on a solo coffee date
49. Say “no” to something draining
50. Do one thing daily that scares you slightly
Boredom isn’t lack of entertainment.
It’s lack of challenge.
Go unlock some XP, bro. 🎮🔥
Sir, I am an ex-JPA scholar gov sent to study in Korea 10 years ago
Let me tell you how the Seoul Metropolitan Gov made their buses punctual like MRT & keep operating costs low
The secret is NOT TRYING to re-invent things, but to keep with WHAT WORKS
THIS IS A THREAD 👇
Intern asked me yesterday: "How do you always know what to do?"
I looked him dead in the eye.
"I don't. I just know how to look like I do."
He laughed. Thought I was joking.
I pulled up my browser history from last Tuesday.
"How to configure VLAN on Cisco switch"
"What is VLAN"
"Cisco switch won't save config"
"Cisco switch blinking orange meaning"
Showed him a ticket from that same day where I "expertly resolved a critical network segmentation issue."
It was the same switch.
His face went pale.
"So you just... Google everything?"
"Not everything. Sometimes I ask ChatGPT."
He asked what he should focus on learning.
"Learn to sound confident when you have no idea what you're doing. The technical skills will follow."
He starts his senior role next month. I'm training him well.
My take on digitalisation and government mobile app.
Government should not require anybody to download any apps to access any of the public services it offers.
As the nation moves to digitalisation, the standard should be with web app. You should be able to do what you need to do via a web browser without the need to download any mobile app.
A mobile app should be as an added convenience, not a requirement.
Here is why a web-first standard approach is better:
Web as a public utility: The web is universal, accessible across devices, and not subject to corporate gatekeeping. For government services, that neutrality is crucial.
The web is interoperable, for as long as you have a web browser you can access it. Design web applications to be mobile-friendly and work seamlessly on mobile screens.
Digital sovereignty: Depending on app stores introduces an external dependency that governments shouldn’t have. Citizens shouldn’t need an Apple ID or Google account to access public services they have a right to.
Apple or Google could decide one day to kick your app out of their app store, and what then? Do you require your citizen to side-load your app?
Accessibility & inclusion: A simple mobile-friendly web app can serve everyone, including those with older phones with limited storage.
Sure, mobile apps can offer features such as better push notifications, better offline access, or better integration with device capabilities, but those should only enhance, not replace the web option. Besides, most of these are achievable with PWAs anyway.
If you are someone close or in the position to influence policymakers, I hope you understand what I am getting at, the web-first digitalisation approach.
A mobile app should be a convenience, never a requirement. Digital Government services must be accessible to all citizens on the open web.
I don't know who needs to hear this, but you can't just auto-translate any English word that ends with "-ation" to "-asi" in Bahasa. 😬
Information is maklumat, not informasi.
Transportation is pengangkutan, not transportasi.
Presentation is pembentangan, not presentasi.
Imagine how much relationships would change if people went into one thinking “how can I make this person’s life better” instead of “how can this person make my life better”
Privilege isn't the presence of perks and benefits, it’s the absence of obstacles. It’s not what you have, it’s how easy it was to get what you have. What hardship did you never have to face? What is a common hurdle that never affected you? That’s what it means to count blessings
Total Operator Kills at Masters Tokyo
1⃣ Zmjjkk - 103 kills
2⃣ ardiis - 30 kills
3⃣ Boostio - 29 kills
Kangkang had an absurd amount of Operator kills at VALORANT Master Tokyo, more than triple the amount of ardiis at #2 had
A little insight from @Jingggxd on how he helped integrate new players into the team 🫶🏻
Read the @ONEEsports article here: https://t.co/1Q0Lfoavcg
#WGAMING