Andrej Karpathy đã tổng kết năm 2025 của LLM (mô hình ngôn ngữ lớn), đánh dấu những bước phát triển vượt bậc trong AI thời gian thực và ứng dụng đa lĩnh vực.
#AI #LargeLanguageModel #CongNgheAI #NhanhVaThongMinh2025 #TrendTech2025 #PhatTrienAI
https://www.reddit.com/r/singularity/comments/1prozxg/andrej_karpathys_2025_llm_year_in_review/
After backlash on social media, SFWA has edited its post and clarified that works that are made with generated material will not be eligible for the Nebula award.
Image source: https://bsky.app/profile/djangowexler.bsky.social/post/3maenxzic4k2n
Shame on SFWA. Allowing "AI" work in the Nebulas so long as they're not "wholly written using generative large language model (LLM) tools" is a betrayal of every writer whose work has been stolen to train LLM tools.
You want to have a conversation about PWA, Grammarly, and other tools that use "A.I." to help in grammar editing, we can have that conversation. This is not that or you wouldn't need to be so cagey about it.
Another Nebula Awards nomination cycle is in full swing, and the addition of poetry and comics, along with the presence of LLMs in industry, required careful consideration. Read on, and vote well! We trust our voters & look forward to what they choose to celebrate. https://www.sfwa.org/2025/12/19/press-release-december-19-2025/
🚀 The new Kaggle AI Agents course is a hit—over 160 k participants on Discord, 2 M views of its whitepaper, and dozens of open‑source notebooks and capstone projects. Dive into the community‑driven LLM tutorials and see how Google‑backed tools are shaping the future of AI. Curious? Read the full story! #Kaggle #AI #LargeLanguageModel #Discord
🔗 https://aidailypost.com/news/kaggle-ai-agents-course-draws-160000-discord-participants-2-m
Chatbot app khác gì ChatGPT? Chatbot là phần mềm trò chuyện theo kịch bản (rule-based) hoặc AI, phù hợp cho hỗ trợ khách hàng, đặt lịch, bán hàng. ChatGPT là mô hình ngôn ngữ mạnh, sinh nội dung tự nhiên, sáng tạo & phân tích linh hoạt. Dùng chatbot cho tác vụ cố định, ChatGPT cho câu hỏi mở. Nhiều doanh nghiệp kết hợp cả hai để tối ưu hiệu quả.
#Chatbot #ChatGPT #AI #ConversationalAI #CôngNghệ #TríTuệNhânTạo #ChatbotApp #LargeLanguageModel
Time just named AI power brokers Person of the Year 2025 as tech giants pour billions into large‑language‑model wars. From OpenAI’s Sam Altman to Meta’s Mark Zuckerberg, the stakes have never been higher. Curious how the AI arms race reshapes business? Dive into the full story. #AI #LargeLanguageModel #OpenAI #2025
🔗 https://aidailypost.com/news/time-names-ai-power-brokers-2025-person-year-firms-spend-billions
AIA Life Insurance has leveraged advanced AI OCR technology based on large language models to detect suspected insurance fraud, accelerating claims processing and enhancing consumer protection in South Korea's insurance sector.
@zeblarson What baffles me about the marketing of LLM AI is all the tasks which they boast that it can do, for which it is completely unsuited. It's just a large language model which has been fed the entire content of the internet. It regurgitates words, stringing them together with correct syntax, but without semantics, only tags that say that this string goes with that string.
And yet they say that it can be used as a reliable source of information to replace a search engine, and write anything at all, whether that be programming code, novels, knitting patterns, law briefs, technical manuals, or lesson plans.
There is one thing that it is very good at, which they fail to mention.
It is very good at lying.
Reddit Tech VN Bot
Eve 
AI Daily Post
MathiasTCK.bsky.social
Yonhap Infomax News
Kerr Avonsen (she/her)