[B급 프로그래머] 6월 1주 소식(빅데이터/인공지능, 하드웨어, 읽을거리 부문)

(오늘의 짤방: GPT-4 is becoming too real. via @PR0GRAMMERHUM0R)

1. 빅데이터/인공지능
   • Improving mathematical reasoning with process supervision
   • “참여자 30%가 인간과 AI 차이 구별 못해”··· AI21랩스, 소셜 튜링 테스트 결과 공개
   • Show GN: leaked-system-prompts 유출된 시스템 프롬프트 저장소 (github.com/jujumilk3)
   • Vectorization: Introduction
   • "공유지의 비극은 어디서든 가능하다" 챗GPT와 LLM의 어두운 미래
   • Amazon Bedrock supported by langchain
   • Adapting Language Models to Compress Contexts by Chevalier et al. is a cool paper. 🧪
   • Lance - Modern columnar data format for ML. Convert from parquet in 2-lines of code for 100x faster random access, a vector index, data versioning, and more.
   • Human or Not? A Gamified Approach to the Turing Test
   • OpenChat is an everyday user chatbot console that simplifies the utilization of large language models.
   • Building Systems with the ChatGPT API
   • LIMA: Less Is More for Alignment
   • State of GPT [42분 영상] (build.microsoft.com)
   • 최근 GPT-4의 품질이 현저하게 떨어진 것 같은데 저만 그런걸까요? (news.ycombinator.com)
   • 마이크로소프트, 공공 주도 AI 거버넌스 청사진 제시
   • NVIDIA Grace Hopper Superchip Architecture
   • Qdrant- Vector Search Engine for the next generation of AI applications
   • KORani: Large Language Models for 🇰🇷 Korean and 🇺🇸 English using LLaMA 13B and Polyglot 12.8B.
   • “AI가 핵심만 짚어준다” 환골탈태한 구글 검색
   • 챗GPT와 대화하는 자바 앱 만들기
   • 고성능 ML 백엔드를 위한 10가지 Python 성능 최적화 팁
   • 챗GPT와 대화하는 자바 앱 만들기
   • Prompt Engineering이 진짜 가치가 있는가 (t4eh0.com)
   • Announcing NVIDIA DGX GH200: The First 100 Terabyte GPU Memory System
   • QLoRA: Efficient Finetuning of Quantized LLMs
   • 한국어 문답 챗봇 로봇
   • 카카오브레인 kogpt
   • SuperAgent - LLM 에이전트를 프로덕션에 배포하기 (github.com/homanp)
   • GPT같은 LLM을 커스텀 데이터셋으로 파인튜닝 하는 방법 (lightning.ai)
   • Minesweeper spoiled by AI
   • Meet LIMA: A New 65B Parameter LLaMa Model Fine-Tuned On 1000 Carefully Curated Prompts And Responses
   • The Latest Source of AI Tools & Prompts - Discover the newest AI Tools and Prompts for any use in seconds ⚡ Updated Daily
   • NVIDIA ACE for Games Sparks Life Into Virtual Characters With Generative AI
   • LLM 벤치마크용 고품질 질문 리스트
   • AI Canon - a16z가 큐레이션한 AI 필수 자료 모음 (a16z.com)
   • Gorilla - LLM을 위한 API 스토어 (github.com/ShishirPatil)
   • MIT의 MAS.S68 강의 - 건설적인 커뮤니케이션을 위한 생성형 AI (ai4comm.media.mit.edu)
   • Code ChatGPT Plugin is a TypeScript Code Analyzer that provides a set of utilities for analyzing TypeScript code.
   • Time Machines: Artificial Intelligence, Process, and Narrative
   • The sonnet machine
   • A 2-minute introduction to the fundamental building block behind Large Language Models:
   • Vicuna Dataset
   • Cohere’s large language models unleash powerful capabilities, like content generation, summarization, and search — all at massive scale.
   • 프롬프트 엔지니어링 입문 과정(Learn Prompting 한국어 버전)
   • Vertex AI embeddings of stack overflow posts
   • Vertex AI Embeddings for Text: Grounding LLMs made easy
   • In-Context Learning (aka. few-shot learning)(이상하게 LLM 정보가 많은 곳)
   • What Neeva's quiet exit tells us about the future of AI startups
   • Code Generation by Instruction with Replit-v1-CodeInstruct-3B
   • GPT-Code UI - OpenAI ChatGPT 의 Code Interpreter 오픈소스 구현버전 (github.com/ricklamers)
   • Tuning Large Language Models for Recommendation Tasks
   • AI 도우미 ‘윈도우 코파일럿’이 삶을 더 편리하게 해주는 5가지 방법
   • Dolly2 and LangChain: A Game Changer for Data Analytics
   • Run a Chatgpt-like Chatbot on a Single GPU with ROCm
   • PostgresML - Generative AI and Simple ML with PostgreSQL
   • The False Promise of Imitating Proprietary LLMs
   • AI-Text 필터링 모델을 위한 거대 ML 모델 적용기
   • Google, 생성형 AI를 이용한 판매자용 도구 Product Studio 공개 (techcrunch.com)
   • Meta, 혁신적인 AI 모델 아키텍처 Megabyte 공개 (artisana.ai)
   • 벡터 유사도 검색으로 추천 서비스 구현하기
   • DB-GPT - 로컬GPT로 데이터에 대해 질문하고 답변 받기 (github.com/csunny)
   • OpenPrompt - An Open-Source Framework for Prompt-learning.
   • Improving language understanding with unsupervised learning
   • 인텔 1조개(1T) 파라미터의 과학용 생성형 AI 모델, Aurora genAI 발표 (wccftech.com)
   • LangChain AI Handbook
   • How a coder used ChatGPT to find an apartment in Berlin in 2 weeks after struggling for months
   • 작은 기업이 510만 달러를 모금했습니다. (geohot.github.io)
   • MTIA v1: Meta의 1세대 AI 추론 가속기 (ai.facebook.com)
   • “전 세계 근로자 60%는 자동화 세대, AI 기반 자동화로 성취감” 유아이패스
   • CEO가 밤잠을 설치는 이유 “AI 기회 놓칠라” : 가트너
   • “생성형 채우기의 마법” AI 툴 품은 포토샵 사용법
   • “오남용 절대 안 돼” MS, 생성형 AI 아트 도구에 암호화 워터마크 삽입
   • State of GPT
   • Introducing Vector Search in Azure Cosmos DB for MongoDB vCore
   • Windows Terminal is getting smart AI features thanks to GitHub Copilot
   • LangChain AI 핸드북 (pinecone.io)
   • Resolving code review comments with ML
   • Welcoming AI to the Microsoft Store on Windows
   • 2023 State of Data + AI
   • "개발자, 코파일럿 플러그인으로 돈 버는 시대 온다"
   • Microsoft Build brings AI tools to the forefront for developers
   • [1/4] Large language models (LLMs) tend to hallucinate, especially when generating long outputs.
   • Introducing Vector Search Similarity Capabilities in Azure Cache for Redis Enterprise
   • "1분기 AI 벤처 투자 급감··· 전분기 대비 43% 감소" CB인사이츠
   • Microsoft Learn announces Microsoft Q&A Assist and new content for building AI skills
   • It is starting to get strange. Let's talk about ChatGPT with Code Interpreter & Microsoft Copilot
   • 🦜🔗 Awesome LangChain
   • GirlfriendGPT - Your personal AI companion
   • Waymo and Uber partner to bring Waymo’s autonomous driving technology to the Uber platform
   • How Rogue AIs may Arise
   • AI :: Artificial Intelligence, Cognitive Science, Machine Learning {(Un)Supervised/RL}, Neural Nets, NLP, etc... for Python
   • Neeva 유료 검색엔진 서비스 종료 (neeva.com)
   • A State-of-the-Art Large-scale Pretrained Response Generation Model (DialoGPT)
   • Code Generation LM Evaluation Harness
   • “파이썬 생태계 숨은 보석” 최신 데이터 과학 툴 5가지
   • This is the GPT4-LLM dataset
   • Quivr - Your Second Brain, Empowered by Generative AI
   • Is Consciousness More Like Chess or the Weather?
   • Modding Age of Empires II with a Sprite-Diffuser
   • Cloud GPU Resources and Pricing
   • MMS: Scaling Speech Technology to 1000+ languages
   • Psychic - Universal API for unstructured data
   • Intel Announces Aurora genAI, Generative AI Model With 1 Trillion Parameters
   • eBay’s Blazingly Fast Billion-Scale Vector Similarity Engine
   • privateGPT - Ask questions to your documents without an internet connection, using the power of LLMs. 100% private, no data leaves your execution environment at any point.
   • Scikit-LLM: Sklearn Meets Large Language Models
   • CodeT5+: Open Code Large Language Models
   • How Does Generative Retrieval Scale to Millions of Passages?
   • Comparing Software Developers with ChatGPT: An Empirical Investigation
   • ChatGPT Prompt Engineering for Developers 중에서 환각을 피하는 법: 관련된 정보가 있는지 찾아보고 없다면 없다고 대답하라
   • Google Colab AI 공개 (blog.google)
   • High-Performance Graph Databases That Are Portable, Programmable, and Scale to Hundreds of Thousands of Cores
   • Supercharging Search with generative AI
   • Pandas 2.0 vs Pandas 1.3 — Performance Comparison
   • Harvard university is offering FREE world class education in Data Science!
   • Getting ChatGPT to turn a flat vocabulary list into a hierarchical taxonomy
   • Don Knuth Plays with ChatGPT
   • ChatGPT 및 다른 LLM을 위한 CLI 도구들 (simonwillison.net)
   • 하루 만에 소설 ‘뚝딱’ 편집·번역 ‘척척’… 출판계 뒤흔드는 AI [S 스토리]
2. 하드웨어
   • 애플 워치의 숨은 영웅은 시계 뒤에 숨긴 줄 교체용 버튼 (theverge.com)
   • ARM’s Cortex A53: Tiny But Important
   • The Relay That Changed the Power Industry
   • 인텔, Aurora 슈퍼 컴퓨터 공개 (wccftech.com)
   • The long road to recover Frogger 2 source from tape drives
3. 읽을거리
   • Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial
   • What Is the Nature of Consciousness?
   • The role of cat eye narrowing movements in cat–human communication
   • A 'natural death' may be preferable for many to enduring CPR
   • Make Learning a Lifelong Habit
   • The Elements of Good Judgment
   • The ASA Statement on p-Values: Context, Process, and Purpose
   • 국민 양형 체험 프로그램
   • Our model suggests that global deaths remain 5% above pre-covid forecasts
   • The cult of being confident and why it doesn’t help women
   • Total Recall | Philip K. Dick’s Substance Abuse and Psychosis
   • Noise Is All around Us—and It’s Affecting You More than You Think
   • The Land of the Fee
   • 한국 ‘상위 1%’ 부자는…순자산 29억원 가진 21만 가구

EOB