[B급 프로그래머] 12월 2주 소식(빅데이터/인공지능, 하드웨어, 읽을거리 부문)

(오늘의 짤방: 근데 한중일 언어 대환장쇼는 이게 최고인거 같음 via @@Benzenekim)

1. 빅데이터/인공지능
   • Use Gemini to prompt Stable Diffusion
   • Here is my selection of papers for today (15 Dec) on @huggingface sent via email
   • Weak-to-strong generalization
   • Log probs are now available with chat completions! (for ChatGPT)
   • UDiffText - A Unified Framework for High-quality Text Synthesis in Arbitrary Images via Character-aware Diffusion Models
   • VGen is an open-source video synthesis codebase developed by the Tongyi Lab of Alibaba Group, featuring state-of-the-art video generative models.
   • Google Launches New Multi-Modal Gemini AI Model
   • Upscale-A-Video - Temporal-Consistent Diffusion Model for Real-World Video Super-Resolution
   • gigaGPT - 565줄의 코드로 만든 GPT-3 모델 (cerebras.net)
   • vectara Cross-Encoder for Hallucination Detection
   • Measuring Hallucinations in RAG Systems
   • Using 🦜🛠️LangSmith To Inspect LangChain Agents
   • Google's new PaLI-3 vision language model achieves performance of 10x larger models
   • GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
   • AutoGPTQ - An easy-to-use LLMs quantization package with user-friendly apis, based on GPTQ algorithm.
   • Unsafe Diffusion: On the Generation of Unsafe Images and Hateful Memes From Text-To-Image Models
   • Researcher develops filter to tackle 'unsafe' AI-generated images
   • After Beatles, AI to bring Edith Piaf voice back to life for biopic
   • Transforming the future of music creation
   • talc.ai - Evaluation as a Service
   • giskard - The testing framework for ML models
   • AI bot capable of insider trading and lying, say researchers
   • Hacking Google Bard - From Prompt Injection to Data Exfiltration
   • Text-to-image AI models can be tricked into generating disturbing images
   • SneakyPrompt: Jailbreaking Text-to-image Generative Models
   • ● Open Interpreter - Let language models run code on your computer. An open-source, locally running implementation of OpenAI's Code Interpreter.
   • Fast and Portable Llama2 Inference on the Heterogeneous Edge
   • 🔎Proposition-Based Retrieval
   • 생산 시스템에 도입 본격화되는 생성형 AI… 핵심은 ‘신뢰 구축’
   • “허비한 3개월조차 아깝다”··· 생성형 AI 기업 업스킬링 사례 6선
   • Round 2: We test the new Gemini-powered Bard against ChatGPT
   • “IT 변화 속도에 발맞추려면 데이터 활용이 관건” 다이나트레이스가 말하는 AI 기반 데이터 관리 전략
   • “온디바이스 AI는 어떻게 작동할까?” 아이폰 확대기 앱을 보면 안다
   • Build apps faster with new Duet AI in Google Cloud training content
   • IBM이 제안하는 ‘생성형 AI 비즈니스 활용법’
   • ‘LLM보다 특정 사용 사례에 적합’… MS, 차세대 SLM 파이-2 공개
   • "생성형 AI의 적정가는 누가 결정하는가?" 생성형 AI 가격 정책과 전망
   • 다른 AI 챗봇은 제공하지 않는 ‘구글 바드’의 독보적인 기능 한 가지
   • IT가 동네 축구팀을 분데스리가로 진출시킨 방법 : TSG 호펜하임
   • 미국 최대 노조와 손잡은 MS "AI로 변화하는 노동환경에 대비"
   • CogAgent: A Visual Language Model for GUI Agents
   • Gemini API Starter Notebook
   • Everybody’s talking about Mistral, an upstart French challenger to OpenAI
   • Mixture of Experts Explained
   • Windows AI Studio simplifies generative AI app development by bringing together cutting-edge AI development tools and models from Azure AI Studio Catalog and other catalogs like Hugging Face.
   • 🦜📃Docugami KG-RAG Template
   • 사업 영역 넓혀가는 AI… 챗GPT, 뉴스 사용 계약 체결
   • Elon Musk가 FSD에 대해서 했던 말들을 시간순으로 정리
   • StemGen: A music generation model that listens
   • PyTorch 2 Internals
   • Microsoft Windows AI Studio 프리뷰 (github.com/microsoft)
   • Gemini Playground 💬 - Play with Gemini Pro and Gemini Pro Vision API
   • tldraw - is this the future of backend?
   • You can now run ChatGPT-like LLMs such as Mistral 7B locally on your phone.
   • Support Vector Machine Clearly Explained!
   • 코어에이아이, 포스코이앤씨 첨단 인공지능 기술 망라한 AI 비서 ‘포박사’ 구축
   • This application demonstrates real time image-to-image inference capabilities of fal.ai.
   • tldraw - why did I learn sql
   • Introducing Google Gemini Pro as Copilot in VSCode 🚨
   • Turning Google smart speakers into wiretaps for $100k
   • "국제통화 이제는 안두렵다!"...AI가 통화 실시간으로 통역해준다!!
   • 2024년 광고 혁명 온다... 구글, 실사에 가까운 생성AI ‘이매진2’ 출시
   • Gemini API로 빌드 - 구글
   • It’s time for developers and enterprises to build with Gemini Pro
   • Outfit Anyone: Ultra-high quality virtual try-on for Any Clothing and Any Person
   • Gemini API and more new AI tools for developers and enterprises
   • PromptBench: A Unified Library for Evaluation of Large Language Models
   • 메타, AI 안전성 평가하는 오픈소스 툴 2종 출시
   • 또 왓슨 브랜드로?...'왕년 빅테크' IBM의 AI 재도전
   • 줌, 줌 컨택센터 요금제 개편… “모든 조직에 AI 기반 고객경험 제공”
   • AWS, 아마존 커넥트의 생성형 AI 신기능 발표
   • ‘책임은 결국 사람의 몫’··· 조직이 감안할 생성형 AI 함정 6가지
   • ‘IT 서비스 데스크, 생성형 AI로 자동화’… 북미 토요타의 애자일 접근법
   • Netflix, 년 2회 인게이지먼트(시청 현황) 보고서 발행 (about.netflix.com)
   • Netflix releases most transparent data to date about viewership
   • Langchain의 주요 구성 요소 소개
   • MedLM: generative AI fine-tuned for the healthcare industry
   • AI로 투자가 대세…증권사 전담조직 신설 붐
   • EdgeSAM - Prompt-In-the-Loop Distillation for On-Device Deployment of SAM
   • First Impressions with Google’s Gemini (esp. vision)
   • ⭐️ Benchmarking semi-structured RAG ⭐️
   • A big factor for building production RAG is deciding the "chunk" used for retrieval + synthesis: should it be a sentence? Paragraph?
   • Celebrating the first year of Copilot with significant new innovations
   • GN⁺: Show HN: 가짜 Google Gemini 데모를 GPT-4를 사용하여 재제작, 이번엔 실제로 구현 (sagittarius.greg.technology)
   • ChatGPT의 강력한 경쟁 언어모델 등장!, Mixtral 8x7B
   • Toolformer: Language Models Can Teach Themselves to Use Tools
   • Multi-modal RAG on slide decks
   • How To Choose Your LLM
   • 구글, Imagen 2 이미지 생성 AI 정식 출시 (cloud.google.com)
   • Solar rises ☀️ @upstageai just released Solar a 10B an open LLM outperforming other LLMs up to 30B parameters, including Mistral 7B.
   • Marigold: Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation
   • tldraw - Screenshot of an entity relationship diagram -> SQL queries needed to create the tables
   • 업스테이지, 글로벌 생성 AI 대전에 출사표!...사전학습 LLM '솔라' 공개, ‘허깅페이스 오픈 LLM 리더보드’ 세계 1위 등극
   • How to Think Exponentially and Better Predict the Future
   • You can now immediately get started building multi-modal and advanced RAG use cases with the brand-new Gemini API ✨
   • count people in zone
   • Imagen 2 - Our most advanced text-to-image technology
   • GN⁺: macOSPilot - macOS를 위한 음성+비젼 CoPilot (github.com/elfvingralf)
   • 0.1% Data Makes Segment Anything Slim 🔥 Colab 🥳
   • COLMAP-Free 3D Gaussian Splatting
   • Phi-2: The surprising power of small language models
   • 🤖 DeciLM-7B-Instruct: A Fast Instruction-Tuned Model💨
   • “Make It Real” AI prototype wows devs by turning drawings into working software
   • Honeybee: Locality-enhanced Projector for Multimodal LLM
   • experimenting with CogVLM (GPT-4V alternative) and image prompts
   • DiffMorpher: Unleashing the Capability of Diffusion Models for Image Morphing
   • Word2Vec 모델로 중국 전4사 텍스트 살펴보기: '皇'의 의미 변화를 중심으로
   • GitHub - microsoft/promptbase: All things prompt engineering - Medprompt and The Power of Prompting
   • The Emergent Abilities of LLMs Could Be A Mirage!
   • ROC and AUC are important concepts for evaluating classification models in business (e.g. lead scoring). In 6 minutes, I'll share what took me 60 days to figure out. Let's dive in.
   • AnimatableDreamer: Text-Guided Non-rigid 3D Model Generation and Reconstruction with Canonical Score Distillation
   • VILA: On Pre-training for Visual Language Models
   • Gaussian-SLAM: Photo-realistic Dense SLAM with Gaussian Splatting
   • Introducing Segmind Vega: Compact model for Real-time Text-To-Image
   • Stability AI, Edge 기기를 위한 Stable LM Zephyr 3B LLM 공개 (stability.ai)
   • tldraw-llm-starter - A starter for working with tldraw and large language models.
   • Together, StripedHyena-7B 모델 공개 (together.ai)
   • A Survey of Large Language Models in Medicine: Principles, Applications, and Challenges
   • Memory Cache is a project that allows you to save a webpage while you're browsing in Firefox as a PDF, and save it to a synchronized folder that can be used in conjunction with privateGPT to augment a local language model.
   • Python client library for Mistral AI platform
   • LLM360 - A framework for open-source LLMs to foster transparency, trust, and collaborative research.
   • Microsoft's Orca 2 LLM Outperforms Models That Are 10x Larger
   • 제2회 LLM 밋업 with. Liner참고: 발표 자료)
   • This is perhaps one of the most important charts on AI for 2024.
   • Audiobox - Meta의 오디오 생성을 위한 새로운 파운데이션 모델 (ai.meta.com)
   • Welcome Mixtral - a SOTA Mixture of Experts on Hugging Face
   • GN⁺: Mistral AI, Llama 2 70B 모델보다 뛰어난 Mixtral 8x7B 모델 공개 (mistral.ai)
   • Audiobox: Where anyone can make a sound with an idea
   • Mistral AI brings the strongest open generative models to the developers, along with efficient ways to deploy and customise them for production. We’re opening a beta access to our first platform services today.
   • Mixtral of experts - A high quality Sparse Mixture-of-Experts.
   • Harvard university is offering FREE world class education in Data Science!
   • 👇quick intro guide for non-technical folks on how to get started(hugging face):
   • Photorealistic Video Generation with Diffusion Models
   • How to Build and Evaluate Agents over your Data 🤖🧪
   • Lean Copilot allows large language models (LLMs) to be used in Lean for proof automation, e.g., suggesting tactics/premises and searching for proofs.
   • DreaMoving: A Human Dance Video Generation Framework based on Diffusion Models
   • Ray vs Dask: Lessons learned serving 240k models per day in real-time
   • IITP "내년 AI경기 좋을 것 50%"···5대 AI전망 발표
   • STraceBERT: Source Code Retrieval using Semantic Application Traces.
   • Are We Testing or Being Tested? Exploring the Practical Applications of Large Language Models in Software Testing
   • Ragas is a framework that helps you evaluate your Retrieval Augmented Generation (RAG) pipelines.
   • I tested the RAG performance of hybrid search weightings (alpha) using Ragas 🧪
   • gpt-fast - Simple and efficient pytorch-native transformer text generation.
   • LLM4TDD: Best Practices for Test Driven Development Using Large Language Models
   • THEY LIVE (1988). 타이포그래피를 스토리텔링에 활용한 최고의 영화 중 하나가 아닌가 싶음.
   • PCA helped to visualize the K-Means similarity of RFQ (request for quotes).
   • CS 4786/5786: Machine Learning for Data Science Spring 2020
   • AI-enhanced development makes me more ambitious with my projects
   • X(트위터), ChatGPT 경쟁제품인 Grok을 X 프리미엄 구독자 대상으로 배포 시작 (techcrunch.com)
   • LlaVa Demo with LlamaIndex
   • 🦙 Panel ChatBot Pack - Build a chatbot to talk to your Github repository.
   • pix2tex - LaTeX OCR
   • Insanely Fast Whisper - distil-whisper small now in insanely-fast-whisper ⚡️
   • "제미나이 프로, 성능 별로"...'바드' 초기 테스트서 불만 속출
   • Improve your RAG application response quality with real-time structured data
   • Multimodal Retrieval Augmented Generation(RAG)
   • EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anything
   • guidance is a programming paradigm that offers superior control and efficiency compared to conventional prompting and chaining.
   • come to our demo at #EMNLP2023 of Papermage, a library for fast manipulation of PDFs
   • PaperMage: A Unified Toolkit for Processing, Representing, and Manipulating Visually-Rich Scientific Documents
   • RAG the real world
   • Amazon Q is simply Outstanding!!!
   • Large Language Models on Graphs: A Comprehensive Survey
   • Booking.com 에서의 축차 검정 (Sequential Testing)
   • How it’s Made: Interacting with Gemini through multimodal prompting
   • 구글, 멀티모달 생성형 AI 모델 '제미나이' 공개
   • 메타, 멀티모달 데이터 세트 Ego-Exo4D 공개 "839명의 인간 활동 기록"
   • 오늘 강의에서 사용한 LLM 프롬프트 작성 요령. 자꾸 길어지면 안 되는데... 그래도 다 유용한 내용이라 못 줄이겠어 😭
   • ⚽ Football Players Tracking with YOLOv5 + ByteTrack
   • 4DHumans: Reconstructing and Tracking Humans with Transformers
   • 애플 실리콘용 ML 프레임워크 ‘MLX’ 등장의 숨은 의미
   • Multi-modal eval: GPT-4 w/ multi-modal embeddings and multi-vector retriever
   • 클라우드플레어, 2024년 AI 주요 전망 발표
   • 메타, 심리스 커뮤니케이션 모델 공개 "'준실시간으로 자연스럽게 번역"
   • 인포뱅크, AI 기반 채팅 서비스 ‘채티스트’ 개발
   • 드디어 내보인 아마존의 '진심' "AWS 생성형 AI의 미래는 통합"
   • 구글의 제미니, 모든 기기를 지원하는 새로운 멀티모달 생성형 AI 소프트웨어 모델
   • 차세대 윈도우에 탑재될 것으로 알려진 6가지 AI 기능
   • 맥도날드, 구글과 손잡고 매장에서 일하는 AI 챗봇 개발한다
   • “AWS 생성형 AI 전략 나왔다”…‘리인벤트 2023’ 행사 총정리
   • 오픈AI ‘큐스타’와 완벽한 AI를 만드는 마지막 5%
   • GN⁺: Meta, AI 관련 새로운 기능 업데이트 (about.fb.com)
   • Large Language Models for Mathematicians
   • Resume Matcher - AI를 이용한 이력서 교정 도구 오픈소스 (github.com/srbhr)
   • Alpha-CLIP: A CLIP Model Focusing on Wherever You Want
   • MLX is an array framework for machine learning on Apple silicon, brought to you by Apple machine learning research.
   • Made With ML - Design · Develop · Deploy · Iterate
   • ⭐️ Multi-modal RAG public benchmark ⭐️
   • Announcing Purple Llama: Towards open trust and safety in the new world of generative AI
   • GN⁺: 구글의 최고 AI "Gemini" 데모는 조작되었다 (techcrunch.com)
   • I tried 14 of the multimodal reasoning examples from the @GoogleDeepMind Gemini paper on @OpenAI's chatGPT-4 (with vision).
   • Ollama Llama Pack Example
   • Building and Productionizing RAG
   • PyPose: A Library for Robot Learning with Physics-based Optimization
   • Here are 12 “hyperparameters” and tuning strategies you can experiment with RAG
   • LLMFlows is a framework for building simple, explicit, and transparent LLM(Large Language Model) applications such as chatbots, question-answering systems, and agents.
   • AlphaCode 2 Technical Report
   • 구글, AI 모델 Gemini 공개 (blog.google)
   • 의료 AI 첫발 내디뎠지만 보험수가 낮아 업계 '울상'
   • Towards Accurate Differential Diagnosis with Large Language Models
   • Summarizing Post Incident Reviews with GPT-4 - How we use GPT-4 to summarize incident reports.
   • Best-of Machine Learning with Python - 🏆 A ranked list of awesome machine learning Python libraries. Updated weekly.
   • GPTs 사용법 - 나만의 GPT 만들기 가이드
   • 무료 인공지능 기초 강좌 모음 (slashpage.com)
   • 🎩Magicoder is a model family empowered by 🪄OSS-Instruct, a novel approach to enlightening LLMs with open-source code snippets for generating low-bias and high-quality instruction data for code.
   • 법조계가 직면한 생성형 AI 과제… 법률 서비스 기업 렉시스넥시스 사례
   • GN⁺: AWS, 기업용 AI 챗봇 "아마존 Q(프리뷰)" 발표 (aws.amazon.com)
   • 맞춤 GPT에 비전API도, 오픈AI 업글 왜 중요한가
   • “아마존 Q, 환각 일으켜 내부 기밀 유출한다” 주장 제기돼
   • Building a Live RAG Pipeline over Google Drive Files
   • Marker converts PDF, EPUB, and MOBI to markdown. It's 10x faster than nougat, more accurate on most documents, and has low hallucination risk.
   • The Efficiency Spectrum of Large Language Models: An Algorithmic Survey
   • Magicoder: Source Code Is All You Need
   • Self-Operating Computer Framework - A framework to enable multimodal models to operate a computer.
   • GPTs vs Assistants - Comparison of GPTs in ChatGPT to Assistants in the OpenAI API.
   • Quivr, your second brain, utilizes the power of GenerativeAI to be your personal assistant ! Think of it as Obsidian, but turbocharged with AI capabilities.
   • Enhance! HORNK! Artificial intelligence can now ID individual geese
   • Leveraging Large Language Models to Improve REST API Testing
   • It's December! And I was just reminiscing about all the things that happened in / defined AI in 2023
   • Segment and Caption Anything
   • RAGs is a Streamlit app that lets you create a RAG pipeline from a data source using natural language.
   • Chunk Viz - This is a tool to visualize how different text splitters chunk up text(데모 사이트)
   • ChatGPT-Comparison-Detection Project 🔬
   • llamafile - 단일 파일로 LLM 배포 및 실행 하기 (github.com/Mozilla-Ocho)
   • SBS D포럼 2023은 "AI시대, 다시 쓰는 경제 패러다임"
   • How AI could lead to a better understanding of the brain
   • Spring AI - The Spring AI project aims to streamline the development of applications that incorporate artificial intelligence functionality without unnecessary complexity.
   • Design Patterns for Machine Learning Based Systems with Human-in-the-Loop
   • LLMs in real-world projects
   • Sketch Video Synthesis
   • StyleCrafter: Enhancing Stylized Text-to-Video Generation with Style Adapter
   • GPT 찍먹해 본 이야기 - SRE Oncall Assistant를 만들어 보자
   • Introduction to Machine Learning(10-315, Spring 2019) by Carnegie Mellon University
   • GN⁺: OpenAI, 커스텀 GPT 스토어 출시를 2024년 초로 연기 (axios.com)
   • Dolphins: Multimodal Language Model for Driving
   • The AI Battlefield Engineering - What You Need To Know
   • SeamlessExpressive - Create translations that follow your speech style
   • Unnatural Error Correction: GPT-4 Can Almost Perfectly Handle Unnatural Scrambled Text
   • LLM Visualization
   • LlamaIndex Full-Stack Projects
   • Retrieval-Augmented Generation (RAG): From Theory to LangChain Implementation
   • GN⁺: LLM 시각화(Visualization) (bbycroft.net)
   • Marker - PDF를 마크다운으로 변환하는 오픈소스 (github.com/VikParuchuri)
   • But how good are multi-modal models at this task? 🖼️🧐
   • Multimodal-Maestro gives you more control over large multimodal models to get the outputs you want.
   • Steamlit Chatbot Pack - Build a chatbot powered by LlamaIndex that augments an LLM with the contents of Snowflake's Wikipedia
   • OpenCopilot allows you to have your own product's AI copilot.
   • CoDi-2: In-Context, Interleaved, and Interactive Any-to-Any Generation
   • unsloth - 2-5x faster 60% less memory local QLoRA finetuning
   • GPT4Video: A Unified Multimodal Large Language Model for lnstruction-Followed Understanding and Safety-Aware Generation
   • The Observe-Orient-Decide-Act (OODA) Loop and Prompting Patterns.
   • GN⁺: Meta, Seamless Communication 모델 패밀리 발표 (ai.meta.com)
   • “생성형 AI, 비즈니스용으로 점점 더 부정확·불투명”… 스탠퍼드 外 연구 결과
   • Show GN: Noiselith - Stable Diffusion XL 을 로컬에서 쉽게! (noiselith.com)
   • AWS, 기업용 생성형 AI 서비스 아마존 Q 발표…MS 코파일럿 겨냥한 듯
   • Look Before You Leap: Unveiling the Power of GPT-4V in Robotic Vision-Language Planning
   • Awesome Deep Vision - A curated list of deep learning resources for computer vision, inspired by awesome-php and awesome-computer-vision.
   • Is Facebook's "Prophet" the Time-Series Messiah, or Just a Very Naughty Boy?
   • DALL·E Party (dalle.party)
2. 하드웨어
   • Firmware key extraction by gaining EL3
   • Leica, the German maker of elegant but absurdly-expensive cameras, just released the M11-P. The most interesting thing about it is a capability whose marketing name is “Content Credentials”, based on a tech standard called C2PA (Coalition for Content Provenance and Authenticity)
   • "데이터센터용 GPU 시장, 2032년까지 연간 28.5% 성장" GMI 전망
   • Microsoft is finally making custom chips — and they’re all about AI
   • Arm Pushes AI into the Smallest IoT Devices with Cortex-M52 Chip
   • "AI 슈퍼컴퓨터 시장, 2028년까지 22% 성장"
   • 2024년, 윈도우 12와 AI CPU 4파전 가능성
   • GN⁺: 바르셀로나 슈퍼컴퓨팅 센터, 새로운 오픈소스 RISC-V 칩 'Sargantana' 발표 (bsc.es)
   • Building a LEGO-powered Submarine with automatic depth control
   • Engineers Working to Resolve Issue With Voyager 1 Computer
   • 애플vs삼성vs구글 "손안의 AI, 1등은 나야나!"
   • Whisper로 비교한 Nvidia 4090과 M1/M2/M3 with MLX (owehrens.com)
   • AMD, 엔비디아 견제 나선다··· AI 가속기 '인스팅트 GPU' 공식 출시
   • 구글의 픽셀 계획을 암시하는 4개의 단어
   • 클라우드 넘어 온디바이스 AI 주목...관련 업계 행보 꿈틀
   • 엔비디아 추격하는 AMD, 신규 AI용 반도체 공개
   • 퐁당퐁당 서버 담근다··· ‘액침 냉각’ 주류화에 쏠리는 관심
   • “나에게 맞는 제품, 어떤 기준으로 골라야 할까?” 애플 펜슬 3종 비교 분석
   • 딥엘, 전 세계 가장 빠른 슈퍼컴퓨터 시스템 톱500 중 34위...연산 능력, 21.85 페타플롭스
   • “서버 판매량 여전히 저조…매출은 오히려 증가” 옴디아
   • LLM, 추천 시스템 및 GNN을 위한 하나의 거대한 슈퍼칩: NVIDIA GH200 NVL32
   • GPT4 was presumably trained for around 90 days using 25k A100 GPUs. Microsoft and Meta having reportedly bought 150k H100 GPUs each this year, can now train a GPT4 class model in only 7 days from scratch (not even including their existing GPU stockpile).
   • 애플 M1 맥이 그렇게 빠른 이유
   • 세그웨이 나비모우 | 정원 관리 ‘궁극의 꿈’ 실현한 잔디깎기 로봇
   • 2024년 네트워크 관리에 일어날 수 있는 좋은 일
3. 읽을거리
   • Crispr gene editing shown to permanently lower high cholesterol
   • How open-source drug discovery could help us in the next pandemic
   • "쿠팡도 직격탄"…불황에 우르르 몰려간 '모바일 앱' 뭐길래
   • 최악 땐 2026년 출산율 0.6명도 깨진다…출생아 10만명 붕괴 가능성
   • [단편] 신뢰의 도약
   • '인앱결제 강제' 앱스토어 모델, 결국 무너지나 | 구글 패소+EU 디지털시장법' 위협적
   • [세이노 칼럼 단독 공개] ‘세이노의 가르침’ 못다 한 이야기
   • GN⁺: 2024년에 공공 도메인에 들어갈 작품들 (publicdomainreview.org)
   • [찰리 멍거 특집 1] 오판의 심리학
   • [박진영의 사회심리학] 오늘 할 일의 70%라도 해냈다면 성공했다고 기뻐하자
   • [최원호의 독일버스이야기 #002] 면허 시험, 도로주행, 자격증 시험까지
   • 저축도 재능이다. 지금 당장 끊어야 할 과소비 습관
   • 1인가구 60%는 年3000만원도 못벌어
   • 쪼그라든 가구당 순자산… 집값하락 여파 빚더미만 쌓였다
   • 자식에게 기대던 시대에서 셀프 부양의 시대로[서영아의 100세 카페]
   • 수도권 출퇴근 계산기
   • 내 ‘출퇴근 체감비용’은 얼마일까…강남 출근길, 고양-김포 月40만원 차이

(보너스: “Use LLMs as a reasoning engine to process information, rather than using it as a source of memorized information” Great tip by Andrew Ng. via @omarsar0)

EOB