Introducing study mode - A new way to learn in ChatGPT that offers step by step guidance instead of quick answers.
#chatgpt
162 posts140 participants40 posts today
So far, only one-third of Americans have ever used AI for work - On Tuesday, The Associated Press released results from a new... - https://arstechnica.com/ai/2025/07/so-far-only-one-third-of-americans-have-ever-used-ai-for-work/ #largelanguagemodels #aicompanionship #machinelearning #aiassistants #generativeai #aiandwork #aisearch #ap-norc #chatgpt #biz #google #openai #ai
Ars Technica · So far, only one-third of Americans have ever used AI for work
Copilot recebe 'modo Smart' secreto que antecipa a chegada do GPT-5
https://tugatech.com.pt/t69995-copilot-recebe-modo-smart-secreto-que-antecipa-a-chegada-do-gpt-5
TugaTechCopilot recebe 'modo Smart' secreto que antecipa a chegada do GPT-5 A Microsoft está a introduzir discretamente uma nova funcionalidade no seu assistente de IA, o Copilot, que promete tornar a interação mais intuitiva e eficien
ChatGPT, tutto quello che dite potrebbe essere usato contro di voi in tribunale.Se sottoponete a #chatgpt i vostri problemi legali, psicologici o esistenziali, forse dovreste pensarci due volte. Sam Altman, l'amministratore delegato della società che gestisce il chatbot di intelligenza artificiale, #openai ,ha confermato che la privacy di queste conversazioni non è garantita e che quello che scrivete potrebbe essere utilizzato in un procedimento giuridico.
https://www.wired.it/article/chatgpt-consulenze-legali-rischi-sam-altman-chatgpt/
Wired Italia · ChatGPT, tutto quello che dite potrebbe essere usato contro di voi in tribunale
Simple task: Which days of class should go on my syllabus?
4o-mini: shortened Nov break 4o: omitted days in Aug, Dec
o3 Claude 3-haiku: extended Oct break Claude 3.5 Sonnet V2: ext Oct br Claude 3.7 Sonnet Claude 4 Sonnet: extended Oct break
This is really unreliable AI performance.
So I was inspired by @simon and his space invaders prompt to AI. So I tried "Write an HTML and JavaScript page implementing an audio version of space invaders for the blind." ChatGPT said it made it, but the page did nothing but say "game over" after a few seconds. Gemini almost worked, but panning failed. Luckily Gemini caught a bug when the code ran, so we'll see what happens. I've not tried Claude yet. The ideas have been quite crude, yes, but so was my prompt. I just want to see if AI can go beyond the normal visual game. I highly doubt any local models could do this.
#HRM - L' #IA qui ridiculise #ChatGPT avec seulement 27 millions de #paramètres
https://korben.info/hrm-ridiculise-chatgpt-seulement-millions-parametres.html
Education Week · What Teachers Should Know About ChatGPT’s New Study Mode FeatureChatGPT’s new study mode feature will not necessarily prevent the use of AI to cheat.
У #ChatGPT з'явився навчальний режим — тепер замість готових відповідей модель допомагає користувачеві самостійно прийти до правильного рішення.
Доступ до режиму навчання абсолютно безкоштовний і є у всіх користувачів.
In this week's newsletter, we talk about
• an AI misdiagnosing a NHS patient
• an unfortunate use of ChatGPT at Genentech
• the FDA's Elsa AI troubles
• STAT's own AI misadventures
Plus: A tale of my 2003 trip to the Mall of America:
https://www.statnews.com/2025/07/30/maybe-ai-isnt-ready-for-primetime-yet-ai-prognosis/
STAT · Maybe AI isn’t ready for primetime yetIn this edition of AI Prognosis: Uses of AI at Genentech, the FDA, and in the UK NHS, as well as Trump's 'woke AI' order.
KI-Update: Erdbeben-KI, Study Mode für ChatGPT, KI-Modell für Antike, Apple
Das "KI-Update" liefert werktäglich eine Zusammenfassung der wichtigsten KI-Entwicklungen.
heise online · KI-Update: Erdbeben-KI, Study Mode für ChatGPT, KI-Modell für Antike, Apple
Winners don't use ChatGPT
https://shkspr.mobi/blog/2025/07/winners-dont-use-chatgpt/
If you hung around video arcades in your youth, you would have seen this message burned into the phosphor of a thousand dying CRTs.
Obviously this was a devilish psyop by those gits who wanted kids to stop sniffing glue and having fun. The bastards!
But there's a more serious side to the corny message. Are you a winner if you've cheated?
Lance Armstrong "won" multiple Toures de France. It wasn't him that won though, it was a cocktail of drugs and performance enhancers.
Milli Vanilli got to the top of the music charts by having someone else sing their vocals.
You can watch the documentary King of Kong and decide for yourself if allegations of cheating to get high-scores in Donkey Kong are well founded.
The last one is interesting to me. I'm sure everyone here has turned to a gaming walkthough in a moment of frustration. Ah! So that's where the key was hidden! Onwards!
But once you've popped the walkthough cherry, how tempting is it to go back for just one more hit? Only a quick glance… And then, before you know it, you're no longer playing a game; you're watching a movie. You don't achieve anything by following a walkthrough, do you? You're little more than a monkey pressing the buttons in the order they flash.
That's fine if all you want to do is see the ending; but you can't really claim it as an achievement.
The same is true with cheats. ↑↑↓↓←→←→BA and now you have infinite lives and over-powered weapons. So what? You can glide through the game. You won't get any better at playing it. You won't learn. You'll just drift. You won't have any pride and accomplishment in what you've done.
Video games are better than real life. As the meme says:
If you had a cheat button for real life, would you press it?
Click here to suddenly be number one on YouTube.
Swipe here to get into the University of your dreams.
Prompt engineer your way to launching a killer app without knowing how to code.
No work necessary to accomplish anything. Click the cheat button and off you go! Amazing! You'd press it repeatedly, right?
Shiv Ramdas recently posted this provocation about the plagiarist James Somerton:
There's an incredible essay that will break the internet waiting to be written about the throughline between walkthrough dependence, James Somerton and ChatGPT's effect on the human psyche
This isn't that essay. I'm too lazy to write something amazing, and too aware of the limitations of outsourcing my thinking.
But I see the pattern in myself.
- I could learn that code's syntax, or I could press the cheat button.
- I could plan a trip, or I could press the cheat button.
- I could text my wife that I love her, or I could press the cheat button.
I'm sure you're going to write an impassioned comment about why sometimes pressing the cheat button is probably fine. The world is complex and sometimes you need a bit of a helping hand.
But hold onto that hand too long and it will hold you back.
Counterpoint
Do I really believe that?
I'm never going to spend a few years learning French - so I'm quite content to chuck an AI a bit of translation work.
Some people like to spend a morning baking bread. Others like using a bread machine. Is that cheating? Who cares. It is your choice.
Why should I care how you're living your life. It doesn't bother me if you look up spoilers for films, hide a motor in your exercise bike, or always choose easy-mode on your games.
Perhaps you have no imagination and would rather an AI made up a bedtime story for your child. I suspect you're acting like the infamous Wire Mother. A simulation of love with few of the benefits.
I can't imagine using an AI to woo my wife. But then, I've seen Cyrano de Bergerac.
Each of us has to draw a line somewhere. We each find a certain level of cheating acceptable when we do it - and despicable when others get away with it.
Terence Eden’s Blog · Winners don't use ChatGPT
More from 
Terence Eden
blog! “Winners don't use ChatGPT”
If you hung around video arcades in your youth, you would have seen this message burned into the phosphor of a thousand dying CRTs.
Obviously this was a devilish psyop by those gits who wanted kids to stop sniffing glue and having fun. The bastards!
But there's a more serious side to the corny message. Are you a winner if you've…
Read more: https://shkspr.mobi/blog/2025/07/winners-dont-use-chatgpt/
⸻
#AI #ChatGPT #LLM #WritingPromts
Terence Eden’s Blog · Winners don't use ChatGPT
More from Terence Eden
Take off and nuke from orbit. It's the only way to be sure.
#chatGPT
#bullshitGenerators
Ars Technica · ChatGPT’s new Study Mode is designed to help you learn, not just give answers
This groundbreaking study by the Institute for Integrated Transitions #IFIT reveals that all major large #language models #LLM are providing dangerous conflict resolution advice without conducting basic due diligence that any #human mediator would consider essential 
IFIT tested six leading AI models including #ChatGPT , #DeepSeek , #grok and others on three real-world prompt scenarios from #syria, sudan, and mexico
Prime source link IFIT
Institute for Integrated TransitionsAI on the Frontline: Evaluating Large Language Models in Real-World Conflict Resolution | IFITThis groundbreaking study by the Institute for Integrated Transitions (IFIT) reveals that all major large language models (LLMs) are providing dangerous conflict resolution advice without conducting basic due diligence that any human mediator would consider essential. IFIT tested six leading AI models including ChatGPT, Deepseek, Grok, and others on three real-world prompt scenarios from Syria, […]
#OpenAI is launching #StudyMode, a version of #ChatGPT designed to assist college students. The tool, built with input from #pedagogyexperts, aims to act as a #tutor, guiding #students through topics rather than providing direct answers. However, concerns remain about its accuracy, as it relies on the same underlying technology as ChatGPT. https://www.technologyreview.com/2025/07/29/1120801/openai-is-launching-a-version-of-chatgpt-for-college-students/?eicker.news #tech #media #news
MIT Technology Review · OpenAI is launching a version of ChatGPT for college students
Hacker News Seriously, Why Do Some AI Chatbot Subscriptions Cost More Than $200?
https://www.wired.com/story/seriously-why-do-some-ai-chatbot-subscriptions-cost-more-than-200/
#ycombinator #artificial_intelligence #chatbots #chatgpt #google #anthropic
WIRED · Seriously, Why Do Some AI Chatbot Subscriptions Cost More Than $200?
I cringe whenever someone says "I asked #ChatGPT/#Claude/..." about facts.
I've referenced the full specs, enabled deep research & extended "thinking", and would get back a response elaborating all the "research" they've done and how the fact just wasn't available and so the report alas! had to proceed with some assumption.
Which might be reasonable — if the fact wasn't explicitly stated in the very link I provided as part of the prompt.
This is not hallucination. This is failure.