 | Bananas glow blue under black light. | Forget neon parties, your fruit bowl is low-key radioactive rave décor. | What's in store: | ChatGPT still explains medical and legal info, despite the rumours. Japan wants AI to ask first, not apologise later. Mindstream Picks: Coca-Cola is doubling down on AI, dropping new Christmas ads with bizarre animated animals. 10 GPT-5 prompts that change everything!
| Read Time: 6 minutes |
|
| | | OPENAI | | | OpenAI says ChatGPT's behaviour has not changed, after posts online claimed it would no longer give legal or medical information. | Karan Singhal, OpenAI's head of health AI, said this is incorrect. | He added that ChatGPT has never been a replacement for real lawyers or doctors, but can still help people understand these topics. | The confusion came after an October policy update that listed activities needing a licensed professional, such as personalised legal or medical advice. | Quick facts | Claims about ChatGPT "banning" legal/medical guidance are false Rules about expert involvement existed before OpenAI simply merged its policies into one document
| Panic for what | OpenAI says this language is not new. Previous rules already required expert involvement for tailored advice in legal, health, or financial areas. | The only real update is that OpenAI combined several separate policy documents into one shared version. The rules themselves stay the same. | In short: ChatGPT can still explain legal and medical topics, but personalised advice must involve a qualified professional. | ICYMI: Policies didn't change, people just can't read. - MG | Should OpenAI make policy updates clearer next time? | | Vote for live results and see results + opinions from yesterday at the bottom of the email. |
|
| | | AI STRATEGY | | | ChatGPT's latest model, GPT-5, just arrived with PhD-level reasoning that makes previous AI look like toys. While most professionals are still using basic prompts, early adopters are leveraging GPT-5's unprecedented capabilities for competitive intelligence and complex analysis. This guide provides 10 advanced prompts engineered specifically for GPT-5's unique strengths. | Get instant access to: | PhD-level analysis prompts that surpass human analysts Unlimited data processing for complex journey mapping Executive-ready insights not just raw analytics
| First-mover advantage starts now. | |
|
| | What book gets lighter the more you use it? | Find the answer at the bottom of the email! |
|
| | I use AI as a copilot to think faster and build cleaner work—never as a shortcut. I design prompts like experiments, verify with sources, and leave an audit trail. Purpose. Learn deeper, code quicker, and stress-test ideas in STEM (math, stats, physics, data/AI). Method (loop). 1. Frame the task with constraints, success criteria, and data I already have. 2. Decompose into subtasks; prompt for each; ask for definitions and edge cases. 3. Cross-check: compare models, verify with textbooks/FRED/Docs, and run unit tests. 4. Document prompts, versions, and decisions in a changelog for reproducibility. 5. Refine: convert results into code, proofs, or write-ups; cite sources. | | | | David, Canada |
|
| Tell us how you're using AI! → |
|
| | | AI ENTERTAINMENT | | | A major Japanese trade group representing studios such as Studio Ghibli has asked OpenAI to stop training AI models on their copyrighted content without permission. | The request follows the rise of "Ghibli-style" images and videos made using OpenAI tools. | When ChatGPT's image generator launched, it became common for users to create artwork in Ghibli's style, even Sam Altman briefly used a "Ghiblified" profile photo. | With OpenAI's new video tool Sora, concerns have grown further. | Japan's Content Overseas Distribution Association (CODA) says material from its members should not be used for training unless permission is granted. | The group argues that, under Japanese law, copying protected works during training could count as infringement. | Unlike the US, Japan does not allow a "use it first, object later" approach. | Legal clarity remains limited. In the US, there is no settled position on whether training AI on copyrighted works is allowed. | In short | A Japanese trade group has asked OpenAI to stop training on its members' work without permission. Japan's copyright rules require approval before use; there is no opt-out system. Legal standards around AI training remain unclear globally.
| Sora stirred the pot | A recent ruling found that AI training on copyrighted books did not violate US law, though the company involved was penalised for how it sourced the material. | Studio Ghibli co-founder Hayao Miyazaki has not commented on this current dispute. | However, in 2016 he criticised AI-generated animation, calling it "an insult to life itself." | I think we can all collectively agree to leave Ghibli alone. Or else. - MV |
|
| | Recommended Reading: As AI accelerates change across industries, understanding the scale of its impact has never been more vital. Semafor Technology is a twice-weekly email briefing that does exactly that. | Space: The Southern Taurid meteor shower peaks Nov 4–5, with up to ~10 bright fireballs per hour, though a near-full Moon may wash out all but the brightest streaks.
Business: Primark's UK sales fell 3.1% as shoppers cut back, prompting owner ABF to explore a potential spin-off of the retail chain.
Music: TRNSMT Festival will move to an earlier June weekend for its 2026 return at Glasgow Green. | Don't Miss: Coca-Cola has released new AI-made Christmas ads featuring odd, cartoonish animals after last year's backlash, doubling down on generative video to cut costs and speed production. |
|
| | Image of the Day |  | Artwork submitted by Mindstream reader Alejandro: "Henry VIII eating a marshmallow" |
| Daily Image Prompt | 1980s retro synthwave scene of a futuristic sports car |
|
| Submit your artwork to Mindstream → |
|
| | "What's the bigger issue here?" | AI hallucinations - 55% ✅ | People using dev tools like consumer apps - 45% | Your Views: | "Users need expertise, both in their specialist discipline and in the application of AI models - and should not hallucinate themselves ;)" - george | "dev apps are just that. Apps for developers to reveal situations like Senator Blackburn found before the application is accepted for public use. It is well known fact that AI's make stuff up "Hallucinate" when forced to answer. Killing the AI application is not an accpeptable fix." - dubeb | "Hallucinations can impact anyone using any tool (e.g. like a consumer car with a problem). People who use dev tools as if they are consumer apps are using specialized tools not designed for them and shouldn't be using them (e.g. like a consumer trying to use a Formula 1 car" - eschr | Submit your opinions in our polls to be featured! |
|
| Riddle Me This Answer: A matchbook. | #889.
01100110 01110010 01100101 01100101 01100100 01101111 01101101 00100001 |  | | | |
|
VHAVENDA I.T SOLUTIONS
0 Comments
VHAVENDA IT SOLUTIONS AND SERVICES WOULD LIKE TO HEAR FROM YOU🫵🏼🫵🏼🫵🏼🫵🏼