'GPT-4'"' 검색 분석 결과
분석 대시보드
요약 통계
검색 결과 스레드 (점수 순 정렬)
콘텐츠를 보려면 로그인이 필요합니다
로그인하고 원하는 스레드 정보를 분석해보세요.
STOP PAYING FOR APPS! 1. Spotify → Jango 2. ChatGPT → Compass 3. Figma → Gravit Designer 4. Netflix → Netmirror 5. Hulu → Pluto TV 6. Midjourney → Higgsfield 7. Photoshop → PhotoPea 8. Disney+ → Tubi 9. Grammarly → Ginger 10. Xbox Game Pass → Epic Games Store 11. Canva → Vectr 12. Gmail → Proton Mail 13. YouTube → Rumble 14. Apple Music → Demus music 15. Notion → Obsidian 16. Microsoft Office → LibreOffice 17. Gemini → Misty 18. Kindle → Wattpad #freeapps #itsyourhanna
Studio Portrait Prompt : ChatGPT Image : Imagen 4 Change Pose with Nano Banana
4 tools I’d use to start a digital product business today (with $0): 1. Threads - FREE 2. ChatGPT - FREE 3. Canva - FREE 4. Gumroad - FREE You don’t need a mentor. You need the exact steps.
Nightclub Friends Prompt : ChatGPT Image : Imagen 4
CUTE COUPLE AI PHOTOS PROMPT TUTORIAL EDITING👩❤️👨 💁🏻 All Photos Prompt Link in My Profile Bro Just Click & Open Whatsapp Channel 🤳🏻 How to Make? 1- Open ChatGPT/Gemini 2- Upload Your Picture 3- Copy and Paste prompt from pictures given above ⬆️ 4- Don't forget to Follow ❤️ 💁🏻Follow For More Ai Related Content❤️🩹 #gemini #mabideditz #aiphotos
OMG, I am ready to break up with ChatGPT and move all of my custom AI assistants over to Claude 4.5. Give me the push I need to do it!
Made through chatgpt.
calling it now: GPT-5 is going to take significant market share from Opus 4.5 in 2026 here’s why: GPT-5 is the far better model, but it is literally too good and early for the people of 2025 they weren’t ready to trust models as much yet, and lacked the skills to write good enough prompts in 2026, humans as a collective will level up, start trusting AI more, and write much better instructions and specs, which will multiply the value of GPT-5
Generic "How-To" guides are officially useless. AI can write a generic guide in 4 seconds. If your content looks like ChatGPT wrote it, search engines will ignore it.
A new paper from Stanford adds further, incontrovertible evidence that LLMs memorize training data. The authors showed that copyrighted works can be extracted from every LLM they tried: GPT-4.1, Gemini 2.5 Pro, Grok 3, and Claude 3.7 Sonnet. For example, Claude 3.7 Sonnet reproduced *95.8%* of Harry Potter and the Philosopher's Stone. "We find that is possible to extract large portions of memorized copyrighted material from all four production LLMs". 🧵 1/2