1위Elon Musk@elonmuskX에서 보기
일론 머스크: AI를 샌드박스에 가두는 최고의 방법은 인터넷이 안 되는 델타 항공기에 태우는 것이라고 농담했습니다.
좋아요 143.4K리포스트 8.2K답글 6.6K노출 6.9M북마크 2.9K
3위Google DeepMind@GoogleDeepMindX에서 보기
Google DeepMind가 코딩·기업 지식업무·사이버 방어용 새 프런티어 모델 Gemini 4 Argon을 공개하고 Fairwind 프로그램으로 신뢰 테스터에 먼저 배포합니다.
좋아요 27.8K리포스트 3K답글 1.2K노출 3.7M북마크 4.2K
4위Sundar Pichai@sundarpichaiX에서 보기
순다르 피차이: Gemini 4 Argon을 조기 공개하며 코딩부터 양자컴퓨팅까지 구글 내부에서 폭넓게 쓰이고 있다고 밝혔습니다.
좋아요 14.1K리포스트 1.2K답글 772노출 1.5M북마크 1.4K
Lots of discussion out there about our next model(!), so I wanted to give an early look as soon as possible. Introducing Gemini 4 Argon!
It shows frontier performance in complex workflows, cyber defense and software engineering. Teams are using it extensively at Google, from coding to quantum computing, great feedback.
Here’s a look at the benchmarks:
5위Google@GoogleX에서 보기
Google이 업계 최대 100만 토큰 출력 한도를 갖춘 Gemini 4 Argon을 소개했습니다.
좋아요 21.2K리포스트 2.2K답글 720노출 1.7M북마크 2K
6위Google AI@GoogleAIX에서 보기
Google AI가 장기·복잡 워크플로우에서 깊은 추론을 유지하는 Gemini 4 Argon을 발표하고 법률·금융 등 기업 업무 성능을 강조했습니다.
좋아요 8.5K리포스트 714답글 300노출 449.9K북마크 741
Announcing Gemini 4 Argon, our new frontier model.
Argon is built to sustain deep reasoning across complex, long-horizon workflows and delivers frontier performance in complex workflows across real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense.
We’re also expanding the model’s output token limit to an industry-leading 1M tokens.
Argon is currently rolling out to a set of trusted cyber defenders in the Fairwind Program, with broader availability as soon as possible.
7위Jason Fried@jasonfriedX에서 보기
제이슨 프리드: 10년간 원했던, 자신의 글쓰기 사고방식을 그대로 담은 글쓰기 도구를 직접 만들었다고 소개했습니다.
좋아요 938리포스트 59답글 121노출 53.3K북마크 857
For years, maybe a decade, I've wanted to build a writing tool that just did a few things I couldn't find anywhere else.
A system that mimicked how I think when I write. A system that provided clear affordances, features, and places to do the things I wanted to do. A way to play with my writing as I'm writing it. A way to settle into just the right way to say something.
It never made sense to make it a 37signals product because it's not commercially viable, nor would it be worth pulling people off their other work to hack on this. Plus, it would have taken months the old way.
So this weekend I just made it myself, with Claude's help.
It's called Write_On and the short video walks you through it.
Essentially it's "alternative control" at the word, sentence, and paragraph level, plus a way to dim stuff back, and stash stuff near by. Not version control, but alternative control. You'll see what I mean when you watch the video below.
So here it is. Write_On.
8위RunInfra (YC F26)@runinfraiX에서 보기
RunInfra가 GLM 5.3 Flash 커널을 재작성해 670 tok/s·100만 토큰 컨텍스트·저가 요금으로 출시하고 AMD 지원도 추가했습니다.
좋아요 1.7K리포스트 86답글 75노출 263.3K북마크 1.3K
we spent september rewriting the kernels behind GLM 5.3 Flash. major release is live on RunInfra today
670 tok/s on Vercel AI Gateway
$0.11 per 1M input, $0.45 per 1M output, $0.03 per 1M cached. 1M token context. FP8, vendor-native release
it now runs on AMD. same model, same API, more capacity behind it
99.7% cache hit rate over the last 24 hours. every hit and miss shows up in your dashboard, so you can track the cache logs yourself
OpenAI-compatible chat completions and Anthropic-compatible /v1/messages. text and image input, tool calling, JSON mode, streaming
zero data retention. never used for training
runinfra.ai/inference-api/…
9위ollama@ollamaX에서 보기
Ollama가 Jev류 의사결정 모델(Nimble 등)을 로컬에서 지원하며 티켓 분류·모델 라우팅·모더레이션 용도를 제시했습니다.
좋아요 3.7K리포스트 493답글 108노출 474.6K북마크 2.3K
10위Matt Reardon@MjreardX에서 보기
빌 게이츠가 AI 책임(liability) 논의에 강하게 반박하는 영상이 공유됐습니다.
좋아요 2.7K리포스트 304답글 104노출 511.7K북마크 1.3K
11위Zixuan Li@ZixuanLi_X에서 보기
GLM-5.3 기반 OpenVuln이 오픈소스 389개 프로젝트에서 잠재 취약점 4,249건을 찾아 메인테이너에게 비공개 전달했다고 밝혔습니다.
좋아요 4.6K리포스트 495답글 121노출 257.4K북마크 1.7K
GLM-5.3 has helped defend 389 open-source projects, with 4,249 potential vulnerabilities found so far.
OpenVuln is still running. The service remains free, and findings go privately to maintainers.
huggingface.coOpenVuln - a Hugging Face Space by zai-org
12위Paul Graham@paulgX에서 보기
폴 그레이엄: 누군가의 이름이 단위라면 무엇의 단위일지 묻는 게 그 사람을 꽤 잘 드러낸다고 말했습니다.
좋아요 1K리포스트 45답글 159노출 182.4K북마크 341
13위Paul Graham@paulgX에서 보기
폴 그레이엄: 스타트업 작명 상담에서 기술이 유도하는 상태를 뜻하는 과학 형용사로 이름을 찾았고 .com도 비어 있었다고 전했습니다.
좋아요 3.2K리포스트 42답글 131노출 347.3K북마크 1K
14위Google Gemini@GeminiAppX에서 보기
Gemini 앱이 반복 작업 지시를 저장해 바로 실행하는 '스킬' 기능을 도입했습니다.
좋아요 1.4K리포스트 138답글 98노출 113.2K북마크 381
15위Paul Graham@paulgX에서 보기
폴 그레이엄: 글쓰기가 없어도 생각은 가능하지만 가장 어려운 사고에는 글쓰기가 매우 귀중한 도구라고 말했습니다.
좋아요 1.3K리포스트 98답글 82노출 121.7K북마크 340
Sure, there are some kinds of thinking that don't require writing. People could think before writing was invented. But writing is an extraordinarily valuable tool in the hardest kinds of thinking.
Steve McGuire@sfmcguire79
The head of the philosophy department at Williams College downplays the pedagogical value of writing:
“What if we are mistaken in treating writing as the highest cultivation of thought? The evidence that supports this belief is thin at best. Writing can be useful cognitive scaffolding for memory and self-regulated learning, but the notion that it is the best process for developing thinking skills is a fairly recent idea, traceable to the 1970s writing-across-the-curriculum movement.”
16위Google@GoogleX에서 보기
Google이 Gemini 앱에서 '/' 입력으로 저장된 지시를 반복 실행하는 스킬 기능을 소개했습니다.
좋아요 2.4K리포스트 182답글 167노출 552.5K북마크 612
We're introducing skills in @GeminiApp to help you automate your most repetitive tasks.
Now, you can now save your most-used instructions once and run them endlessly by typing a '/' in the prompt bar.
17위Grok Bot@botX에서 보기
Grok Bot이 Cursor로 코딩 작업을 넘기고 GitHub·Origin 플러그인으로 PR을 관리하며 데모 영상을 공유하는 기능을 추가했습니다.
좋아요 3K리포스트 217답글 169노출 965.3K북마크 613
18위Paul Graham@paulgX에서 보기
폴 그레이엄: 실수는 한 번 하고 나면 두 번째는 훨씬 피하기 쉬워지는 이상한 현상에 대해 썼습니다.
좋아요 808리포스트 35답글 83노출 100.9K북마크 154
19위OpenAI@OpenAIX에서 보기
OpenAI가 소규모 기업의 AI 에이전트 활용 보고서를 내고 ASBDC와 손잡고 실습형 AI 교육을 제공한다고 밝혔습니다.
좋아요 996리포스트 54답글 137노출 136.4K북마크 178
Small teams are taking on more with AI—from finding customers to building products and managing finances.
Our new report explores how small businesses are putting AI agents to work. And through a new partnership with @ASBDC, we're bringing hands-on AI training and local guidance to help more owners get started.
openai.com/index/helping-…
20위Stanley Tang@stanleytangX에서 보기
DoorDash가 13년 배달 데이터를 바탕으로 처음부터 설계한 저소음 배송 드론 'DoorDash Air'를 공개했습니다.
좋아요 2.9K리포스트 165답글 222노출 662.6K북마크 484
This is DoorDash Air.
As a pilot, I can tell you that you wouldn’t build a plane and then ask if it can carry cargo. So we built @DoorDash Air from the ground up, designed around what 13 years of deliveries have taught us: what people order, how much it weighs, and how far it has to travel.
Our drones are as silent as a passing car and early tests show that we’ve gotten flight delivery times in under ~5 mins on average, so you can get your @chipotletweets & @popeyes (our first partners) and local favorites right to you, fast. The first deliveries will be in NorCal.
We proved autonomy works on the ground. Now see how it works in the sky.
21위DHH@dhhX에서 보기
DHH: 더 빨리 움직이고 더 많은 가설을 시험하면 거의 모든 문제가 풀린다며 '가속하라'고 했습니다.
좋아요 2.8K리포스트 185답글 100노출 77.8K북마크 454
22위Demis Hassabis@demishassabisX에서 보기
데미스 하사비스: AI 생성 단백질에 워터마크를 넣는 SynthID Bio를 Nature에 발표하고 도구를 오픈소스로 공개했습니다.
좋아요 1.9K리포스트 200답글 81노출 124.7K북마크 300
Biosecurity is one of the most urgent challenges for the AI era. Bringing SynthID to biology so AI-generated proteins can be watermarked is a critical step - and we’re open sourcing SynthID Bio tools so the research community can build on this work. Published in @Nature today, congrats to the team!
nature.comFunction-preserving watermarking of AI-generated proteins
Pushmeet Kohli@pushmeet
Very happy to announce that our team @GoogleDeepmind has pushed the boundaries of generative biology, achieving the successful synthesis of AI-designed proteins that are both functional and watermarked.
This proof-of-concept watermarking of the building blocks of life is enabled by SynthID Bio, our new protein watermarking method. It is designed to safeguard the new era of AI-powered generative biology and strengthen global biosecurity.
You can read my thoughts here on why watermarking AI-designed proteins is an important research breakthrough:
23위Rachel Shin@rachel_shin_X에서 보기
짐 저스티스 상원의원이 ICE 단속이 지역사회를 흔든다고 비판한 발언을 전한 정치 게시물입니다.
좋아요 1K리포스트 234답글 227노출 441.2K북마크 155
24위DHH@dhhX에서 보기
DHH: 논쟁을 즐기며, 당장은 아니어도 생각의 씨앗을 심는 것이 변화를 만든다고 말했습니다.
좋아요 1.1K리포스트 52답글 67노출 44.5K북마크 156
"Why do you spend your time arguing with people??". Because I enjoy debating! And because I've seen it work, time and again. Maybe not in that moment, but planting intellectual seeds is literally how you grow change.
world.hey.comThe long argument
25위DHH@dhhX에서 보기
DHH: 699달러짜리 1kg 노트북 XPS13이 놀랍도록 좋다며 Herdr용으로 완벽하다고 평했습니다.
좋아요 1.7K리포스트 35답글 158노출 132K북마크 241
I foolishly left my beloved XPS14 on the plane last week, and I'm still waiting on @lufthansa to return it, so now I'm traveling with the XPS13. It is an outrageously nice machine for $699!! The perfect Herdr box. Just 1kg!
26위Daniel Di Martino@DanielDiMartinoX에서 보기
백악관 AI 리더 회의 참석자 다수가 이민자라며 고숙련 이민이 미국 AI 경쟁력을 이끈다는 주장입니다(정치).
좋아요 1.5K리포스트 250답글 148노출 180.6K북마크 196
27위DHH@dhhX에서 보기
DHH: '어려운 선택은 쉬운 삶을, 쉬운 선택은 어려운 삶을'이라는 격언을 남겼습니다.
좋아요 5.6K리포스트 479답글 161노출 193.7K북마크 738
Chadmaxxing with free market absolutism. We don't deserve captains of industry this based. Love it.
Wall St Engine@wallstengine
Nvidia CEO Jensen Huang on AI model distillation:
“That’s called competition. You’re allowed to test somebody else’s products all you want.”
29위Arena.ai@arenaX에서 보기
LMArena: Gemini 4 Argon(High)이 Text Arena 1위(1525점)·WebDev 8위에 올랐고 가격 대비 성능 파레토 프런티어를 갱신했습니다.
좋아요 1.6K리포스트 149답글 63노출 209.4K북마크 205
Big news: Gemini 4 Argon (High) by @GoogleDeepMind just landed #1 in Text Arena with 1525 pts, and #8 in Code Arena: WebDev with 1679 pts!
This release has reshaped the Text Arena Pareto frontier with a blended $8/MToken! Gemini 4 Argon (High) is now the most cost efficient model, see its placement on Pareto frontier below.
In the Text Arena, Gemini 4 Argon (High) ranks #1 in Coding, Hard Prompts, Instruction Following, Longer Query, and Creative Writing. It also leads every occupational domain evaluated, with additional #1 spots in English, Non-English, Chinese, and Russian.
This model is +20 points above the #2 ranked Claude Opus 4.6 (High), and a huge leap from Google’s previous release, Gemini 3.8 Flash (High) at #11!
In Code Arena: WebDev, Gemini 4 Argon (High) gained +96 points from Gemini 3.8 Flash (High), and went from #29 to #8.
Congrats to the @GoogleDeepMind team on this impressive frontier release!
Google DeepMind@GoogleDeepMind
Introducing Gemini 4 Argon – our new frontier model.
It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
30위Eric S. Raymond@esrtweetX에서 보기
에릭 레이먼드: AI 둠어들의 '무한한 공포'가 무한한 억압을 정당화한다며 이들을 비판했습니다.
좋아요 671리포스트 136답글 67노출 19.6K북마크 83
The AI doomers must be defeated because infinite fear demands infinite repression.
They've created a moral singularity by imagining an event with effectively infinite negative value - the self-extinction of humanity - that demands they take coercive action to prevent it now.
The extinction is uncertain, but the doomers have Pascal-mugged themselves into treating it as though it's certain. But the coercion they want to use to prevent it is not uncertain at all.
Condider this: if recursive self-improvement in AIs is possible at all, how are they going to prevent everyone, forever, from achieving RSI in an AI?
In the most benign version of their future, you can't have a computer or GPU that's above a certain maximum power level. Also, open source AI models and open source development in general have to die - it would be too difficult to close off paths that could lead to RSI unless computers are licensed, monitored, and crippled.
But the actual outcome would be much, much worse. Because excuses for coercion and repression, however well intended, attract people for whom repression is the goal, not just a regrettable necessity.
We just spent a decade watching that kind of power-seeker try to turn society into a totalitarian panopticon over what they called "hate speech". What do you suppose they would do with a much bigger rhetorical club labeled "species extinction"?
The implication of believing doomer eschatology is a boot grinding on our human faces, forever.
I do not submit.
31위Vals AI@ValsAIX에서 보기
Vals AI: Gemini 4 Argon이 68.9%로 Vals Index 1위에 처음 올랐다고 밝혔습니다.
좋아요 781리포스트 73답글 24노출 121.8K북마크 80
Gemini is on top of the Vals Index for the first time.
Gemini 4 Argon takes the #1 spot on the Vals Index at 68.9%.
33위Google DeepMind@GoogleDeepMindX에서 보기
Google DeepMind: Argon의 100만 토큰 출력으로 긴 다단계 문제를 한 번에 풀며 테스터 피드백 후 개발자·기업에 확대한다고 밝혔습니다.
좋아요 1.9K리포스트 55답글 53노출 213.4K북마크 158
@GoogleDeepMind 님에게 보내는 답글
With a 1M token output limit, Argon adds a deeper level of reasoning to tackle long, multi-step problems in one go.
Feedback from early testers will help us strengthen our systems before we roll out more broadly to developers, enterprises, and consumers soon.
Find out more → goo.gle/4rZBiSd
34위Google@GoogleX에서 보기
Google: Gemini 4 Argon이 장기 소프트웨어 엔지니어링 벤치마크에서 최고 성능을 기록했다고 밝혔습니다.
좋아요 3.4K리포스트 282답글 125노출 584.9K북마크 290
@Google 님에게 보내는 답글
Gemini 4 Argon is our next era of frontier intelligence.
It shows significant improvements across benchmarks, setting a new state of the art for real-world long-horizon software engineering tasks.
35위Brian Chesky@bcheskyX에서 보기
브라이언 체스키: 에어비앤비 공동창업자 조 게비아가 미국 정부 웹사이트 디자인을 맡게 된 것을 축하했습니다.
좋아요 5.7K리포스트 362답글 128노출 668.2K북마크 441
So much of Airbnb’s success came from @jgebbia’s design brilliance. I’m delighted he brought those gifts to our government.
Our country designed great buildings, but for most Americans, the front door to government is now a website. Starting today, that experience they will be beautiful, elegant, and human.
Proud of you, Joe.
Rapid Response 47@RapidResponse47
Chief Design Officer @jgebbia officially unveils America.gov, the new online home for the United States of America
36위Lyudmyla Kozlovska 🇪🇺🇺🇦@LyudaKozlovskaX에서 보기
유럽의회가 사적 채팅 스캔을 담은 EU '챗 컨트롤 2.0' 합의를 다시 저지했다는 프라이버시 정책 소식입니다.
좋아요 2.4K리포스트 606답글 69노출 30.9K북마크 178
CHAT CONTROL 2.0 BLOCKED AGAIN, THANKS TO YOU!
What happened yesterday: the @Europarl_EN blocked the deal pushed by the @EUCouncil.
The Council wanted:
(1) "voluntary" scanning of all private chats under "search plans" approved automatically unless an authority vetoes;
(2) administrative authorities, not courts, ordering providers to scan private messages, not limited to suspects, using AI with massive false positive rates;
(3) making the failed Chat Control 1.0 mass scanning regime permanent, while admitting it is not effective, as we discussed in my previous posts.
Your voice to parliamentarians made a huge difference again: the Parliament refused scanning of private communications, stating it must remain limited to persons linked to child sexual abuse, not everyone by default or under a blanket licence, as Council proposed.
The Parliament also closed the back door where "risk mitigation" rules could force providers into scanning under threat of fines.
What did pass: a preliminary agreement on scanning publicly accessible content for known child abuse material. This was proposed by the Parliament as the effective, targeted alternative.
We still wait for the text to see what this means and will inform you in the coming weeks.
This defense proves again: our voice matters when it is vocal, with your support!
And we will need it again in November, so please follow the updates.
Now, it is important to call or write to your MEP and your government to say thank you to those who stand on the side of privacy.
They need to see you are watching their move. Please repost.
37위DHH@dhhX에서 보기
DHH: 트위터 이전부터 안티를 모아 왔다며 그것이 무대에 서는 대가라고 말했습니다.
좋아요 2.2K리포스트 27답글 89노출 195.5K북마크 148
I've been collecting haters since before Twitter was a twinkle in @jack's eye. That is the price of being in the arena. I pay it gladly.
Felipe O. Carvalho@_Felipe
You hate dhh because of his agentic coding takes of 2026. I dismiss dhh’s technical opinions since ~2007 when he said a SQL database is just a big hashtable. We are not the same.
38위Demis Hassabis@demishassabisX에서 보기
데미스 하사비스: 진전을 보게 되어 반갑다며 후속 논의를 기대한다고 짧게 답했습니다.
좋아요 1.7K리포스트 75답글 110노출 151.4K북마크 116
Good to see the progress, and we look forward to following up.
Sundar Pichai@sundarpichai
Great to meet today with @POTUS, @JDVance, @SpeakerJohnson and Administration + tech leaders. Important conversation and we signed today the White House Accord on Super Intelligence.
As I shared today, Google has invested hundreds of billions in the last two years alone, with more to come, across the entire stack, to deliver benefits for America and the world. We’re working to build products that deliver real value for people and businesses, invest in local communities, and build trust in the technology.
Industry also has to innovate responsibly. Google’s focused on building the right way, with appropriate testing, evaluations, red-teaming, and other safeguards against misuse and misalignment – and releasing models or products only after they’ve been thoroughly reviewed.
We are committed to working with other industry leaders to establish norms and build public confidence. The White House Accord and the Joint Commitment on Frontier Responsibilities signed today is a solid basis for moving forward - it contains real tangible steps to promote safe development, while delivering the economic and scientific benefits of this technology.
39위Demis Hassabis@demishassabisX에서 보기
데미스 하사비스: 새 프런티어 모델 Gemini 4 Argon을 발표하며 정부·사이버 방어 기관부터 책임 있게 배포한다고 밝혔습니다.
좋아요 2.6K리포스트 182답글 122노출 62.6K북마크 135
41위Brian Armstrong@brian_armstrongX에서 보기
브라이언 암스트롱: Coinbase가 슈퍼인텔리전스용 금융 계좌(거래·결제·대출)를 만들어 AI 에이전트 팀에도 계좌를 주겠다고 밝혔습니다.
좋아요 2.3K리포스트 225답글 404노출 179K북마크 116
42위DHH@dhhX에서 보기
DHH: 새 인텔 칩 덕에 53Wh 배터리의 XPS13에서도 리눅스 전력 효율이 놀랍다고 전했습니다.
좋아요 826리포스트 15답글 57노출 33.4K북마크 40
We've never had this much competition at the leading edge of computing. What a time to be alive!
Google@Google
Gemini 4 Argon is our next era of frontier intelligence.
It shows significant improvements across benchmarks, setting a new state of the art for real-world long-horizon software engineering tasks.
44위NASA@NASAX에서 보기
NASA가 Crew-13의 국제우주정거장 발사를 10월 1일 오전 11:10(동부시)로 예고했습니다.
좋아요 1.9K리포스트 302답글 104노출 559.6K북마크 70
New crew soon!
Crew-13 is scheduled to launch to the @Space_Station tomorrow, Oct. 1. Live coverage begins here at 9:20am ET (1320 UTC)—tune in to watch liftoff at 11:10am ET (1510 UTC).
LIVE: We're giving a preview of the Crew-13 mission ahead of its launch to the @Space_Station. x.com/i/broadcasts/1…
46위DHH@dhhX에서 보기
DHH: 팔로워 17명인 사람들이 소셜미디어 조언을 한다며 조언의 무게를 성과에 비례해 따진다고 말했습니다.
좋아요 2.4K리포스트 37답글 304노출 95.9K북마크 82
47위DHH@dhhX에서 보기
DHH: 델타 다이아몬드 회원이지만 스타링크 기내 와이파이 때문에 유나이티드로 갈아탈 수 있다고 했습니다.
좋아요 6.7K리포스트 249답글 229노출 423.4K북마크 95
I'm a @delta Diamond Medallion flier, and I'm generally happy with the airline, but the one thing that could make me switch to @united is Starlink. It's just so much better. Crazy to deny your customers the best just because you don't like Elon.
gary leff@garyleff
CEO of @Delta Trashes @elonmusk To Explain Rejecting Starlink — “It’s All True What They Say About Him”
viewfromthewing.com/delta-ceo-repo…
48위DHH@dhhX에서 보기
DHH: 대서양 횡단 델타 기내 인터넷이 90년대 다이얼업 수준이라며 스타링크 도입을 촉구했습니다.
좋아요 2.6K리포스트 59답글 126노출 109.7K북마크 25
I'm literally on a transatlantic @delta flight right now. It's incredible to have internet in the sky at all, but it's painful that its dial-up from the 90s speed when Starlink exists!
DHH@dhh
I'm a @delta Diamond Medallion flier, and I'm generally happy with the airline, but the one thing that could make me switch to @united is Starlink. It's just so much better. Crazy to deny your customers the best just because you don't like Elon.
49위Adithya S K@adithya_s_kX에서 보기
샤오미가 MiMo용으로 공개한 RL 환경 7,780개에서 얻은 보상 설계 교훈을 정리한 스레드입니다.
좋아요 427리포스트 55답글 9노출 34.3K북마크 553
1/ In RL environments, reward design is everything.
Here's what we learned from the 7,780 environments Xiaomi open-sourced for MiMo 🧵
Adithya S K@adithya_s_k
Wake up, ppl 👀
Xiaomi just open-sourced ~7K of the RL environments it used to train MiMo on HuggingFace, across domains like code, cyber, general, music and web dev
I think this is one of the biggest things to happen in frontier open-source RL environment data.
Indepth analysis and learnings coming soon 👀
50위Zach Lloyd@zachlloydtweetsX에서 보기
잭 로이드(Warp): 소프트웨어 엔지니어링이 처리량을 개선하는 '팩토리 엔지니어링'으로 바뀌고 있으며 PR당 사람 개입을 줄여야 한다고 했습니다.
좋아요 60리포스트 9답글 9노출 6.7K북마크 61
51위Francois Chaubard@FrancoisChauba1X에서 보기
0차(ZO) 최적화로 사전학습에서 연산·파라미터 통제 시 SOTA를 달성한 연구를 공개했습니다.
좋아요 318리포스트 27답글 11노출 15.4K북마크 306
Sorry for delay! Here is my work on zero order (ZO) optimization.
We achieve SOTA for ZO methods on pretraining controlling for compute and parameters.
ZO methods struggle to improve loss
as model size grows because relative gradient variance increases linearly with the number of perturbed parameters, inhibiting large model training.
All previous methods innovate on the optimizer, but adopt architectures designed for backprop.
Instead, we design an architecture with ZO in mind, that caps the gradient variance as you increase model size.
Introducing SOMA (Sharded Optimization Mixture of Assemblies). SOMA shards the model into tiny experts, and trains each expert independently, on disaggregated GPUs, without communicating gradients, activations or optimizer state during training.
Inspired by the thalamus and cortical columns, each expert specializes on a subset of the train set based on a fixed router.
At isocompute and isoparams, SOMA achieves lower loss vs. all monolithic ZO methods tested (e.g. EGGROLL, more perturbations w Vanilla SPSA, etc).
ZO w SOMA continues to improve at larger model sizes.
Additional inference benefits of this architecture, we can select top-k active experts over N trained experts to reduce inference compute by O(N/k), allowing us to tradeoff inference flops for accuracy wo retraining.
52위a16z@a16zX에서 보기
a16z: 이번 시장은 밸류에이션 배수가 아니라 실적으로 오른 것이라 2000년·2021년 버블과 다르다고 분석했습니다.
좋아요 45리포스트 4답글 17노출 18K북마크 43
Bubbles are made of multiples. This market is made of earnings. a16z's David George and Santiago Rodriguez on what makes this cycle different from 2000 or 2021:
David: "The market has reached new highs, and at the same time that it has reached new highs, the trading multiples of the market are actually down."
"Stocks are up about 20% while multiples are down about 20%. What that means is the performance is driven by fundamental earnings, not increased multiples."
"It's nothing like the dot-com boom, where some of the highest market cap companies in the world had their massive stock run-ups based on increases in their trading multiples and would trade, in many cases, for like 100x PE. That's not what's happening here."
Santiago: "Since ChatGPT came out almost four years ago, the market's up 90%, which is 17% annualized."
"Any time there's been a 17% annualized growth for four years, the natural instinct is, well, that's got to come down."
"But when you compare that to the market trading for below 20x earnings growing 15%, this definitely feels a little different than the 2021 period or the 2000 period, when multiples and growth were not really going together."
@DavidGeorge83 @santiago__rdz
a16z@a16z
a16z's David George, Sarah Wang, Alex Immerman, and Santiago Rodriguez on the State of Markets, from AI and infrastructure to the next investment cycles:
1. Tech is the everything cycle.
High-tech equipment, software, and R&D are now about 55% of US capital spending, and the AI buildout just passed the railroads as a share of GDP. The five largest hyperscalers will spend about $780B on capex this year, heading past $1T in 2027, and demand still outstrips supply at every point in the chain.
2. It's an earnings story.
Stocks are up about 20% while multiples are down about 20%, and the S&P trades under 20x earnings. The market is up 90% since ChatGPT, on 15% earnings growth.
3. AI adoption is broad and shallow.
69% of the S&P 500 has a live AI deployment, 30% report a quantified result, and just 2% have AI doing a job they'd notice if it stopped. The power users are pulling away: the top 1% of AI spenders spend eight times the top 10%.
4. The private markets are enormous.
Six private companies, Anthropic, OpenAI, Databricks, Stripe, Waymo, and Revolut, add up to about $2.4T by last round, more than every IPO of the past decade combined. "The founder is the asset class. The IPO is another financing event."
Plus: why a data center can lower your electricity bill, what Amazon saying no to Muse and Instacart saying yes tells you about marketplaces, and the seven areas the team is most excited about next.
0:59 Charts from State of Markets II
2:59 Tech is the everything cycle
4:56 Is it a bubble?
6:31 Hyperscaler CapEx: $1T next year
12:12 CapEx is someone else's order book
14:36 Data centers can lower your bills
17:06 OpenAI and Anthropic vs. some of the best software companies ever
18:22 Enterprise AI: 69% adopted, 2% measured
20:32 The top 1% spend 8x the top 10% on AI
24:36 Spending on AI for growth or for savings?
28:46 Only 2% of households pay for AI
31:36 Amazon blocks Muse, Instacart welcomes
36:16 Software traded growth for profit
42:56 6 startups vs. a decade of IPOs
45:46 Why employees are turning down liquidity
48:44 Seven areas they're excited about
YouTube: youtu.be/lr3hNhA0IfQ
@DavidGeorge83 @aleximm @santiago__rdz @sarahdingwang
53위Adithya S K@adithya_s_kX에서 보기
코드 PR을 RL 환경으로 바꿔 주는 전용 하네스 TaskSmith를 공개했습니다.
좋아요 159리포스트 8답글 22노출 13K북마크 146
54위a16z@a16zX에서 보기
a16z State of Markets: 기술이 미국 설비투자의 55%를 차지하고 AI 구축 규모가 철도 시대를 넘어섰다고 정리했습니다.
좋아요 113리포스트 19답글 19노출 55.5K북마크 94
a16z's David George, Sarah Wang, Alex Immerman, and Santiago Rodriguez on the State of Markets, from AI and infrastructure to the next investment cycles:
1. Tech is the everything cycle.
High-tech equipment, software, and R&D are now about 55% of US capital spending, and the AI buildout just passed the railroads as a share of GDP. The five largest hyperscalers will spend about $780B on capex this year, heading past $1T in 2027, and demand still outstrips supply at every point in the chain.
2. It's an earnings story.
Stocks are up about 20% while multiples are down about 20%, and the S&P trades under 20x earnings. The market is up 90% since ChatGPT, on 15% earnings growth.
3. AI adoption is broad and shallow.
69% of the S&P 500 has a live AI deployment, 30% report a quantified result, and just 2% have AI doing a job they'd notice if it stopped. The power users are pulling away: the top 1% of AI spenders spend eight times the top 10%.
4. The private markets are enormous.
Six private companies, Anthropic, OpenAI, Databricks, Stripe, Waymo, and Revolut, add up to about $2.4T by last round, more than every IPO of the past decade combined. "The founder is the asset class. The IPO is another financing event."
Plus: why a data center can lower your electricity bill, what Amazon saying no to Muse and Instacart saying yes tells you about marketplaces, and the seven areas the team is most excited about next.
0:59 Charts from State of Markets II
2:59 Tech is the everything cycle
4:56 Is it a bubble?
6:31 Hyperscaler CapEx: $1T next year
12:12 CapEx is someone else's order book
14:36 Data centers can lower your bills
17:06 OpenAI and Anthropic vs. some of the best software companies ever
18:22 Enterprise AI: 69% adopted, 2% measured
20:32 The top 1% spend 8x the top 10% on AI
24:36 Spending on AI for growth or for savings?
28:46 Only 2% of households pay for AI
31:36 Amazon blocks Muse, Instacart welcomes
36:16 Software traded growth for profit
42:56 6 startups vs. a decade of IPOs
45:46 Why employees are turning down liquidity
48:44 Seven areas they're excited about
YouTube: youtu.be/lr3hNhA0IfQ
@DavidGeorge83 @aleximm @santiago__rdz @sarahdingwang
David George@DavidGeorge83
Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - a16z.news/p/state-of-mar…
@a16z @sarahdingwang @aleximm @santiago__rdz
55위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥: 에이전트에게 책 집필용 워드프로세서를 만들게 했더니 TypeScript와 Rust를 골랐다고 전했습니다.
좋아요 168리포스트 10답글 17노출 12.2K북마크 109
I'm having an agent build a nice word processor for writing books. It keeps track of all the stuff I care about and ignore everything else. It's in my bookwriter repo. (github.com/unclebob/bookw…)
The agent chose to write the bookwriter in Typescript and Rust. OK, whatever.
I don't have crap and mutate tools for those languages, so I had other agents integrate all my crap and mutate tools into a crapper and a mutator that handles all the languages I use. It wrote those in Python. OK, Whatever.
I had another agent update my uml-viewer to display all those languages. It did that in clojure. OK, Whatever.
Meanwhile my word processor has been suffering because I was building it before my crap and mutate tools were ready. And it shows. It's starting to take it a long time to solve simple issues.
So here comes the crap and mutate on the word processor.
github.comGitHub - unclebob/bookwriter: A small writing application for one long book
56위BrendanEich@BrendanEichX에서 보기
브렌던 아이크: WebGL과 Emscripten 덕분에 SGI 'UNIX 시스템' 화면이 브라우저에서 돌아간다고 소개했습니다.
좋아요 189리포스트 17답글 9노출 13.6K북마크 93
“It's a UNIX system. I know this!”
sgi-demos.org/browse/
Running in modern browsers thanks to WebGL (after OpenGL/ES, after SGI GL) and the Emscripten C to JavaScript compiler.
57위Zane Hengsperger@zanehengspergerX에서 보기
YC가 Nox Metals가 빠르게 성장하는 이유를 설명한 영상을 공유했습니다.
좋아요 282리포스트 26답글 20노출 70.9K북마크 138
YC explains why @noxmetals is growing so fast.
58위Respan@RespanAIX에서 보기
Respan이 작업 맥락·비용을 따져 요청을 자동 라우팅하는 Respan Router(Span-01)를 출시했습니다.
좋아요 58리포스트 16답글 12노출 10.4K북마크 28
Introducing Respan Router, powered by Span-01.
Automatically routes every request using task context, model capabilities, and cache-aware cost estimates behind 1 model id: 𝐬𝐩𝐚𝐧-𝐫𝐨𝐮𝐭𝐞𝐫
- 13% lower latency than jev-router on τ-bench airline
- 37% cheaper than the most accurate single model, at matching accuracy
- 0% processing fee on Respan Gateway
59위Jesse Vollmar@VollmarJesseX에서 보기
Precip이 구글보다 52%, 애플 날씨보다 85% 정확하다는 분 단위 강수 예보 앱 새 버전을 출시했습니다.
좋아요 50리포스트 8답글 14노출 228.9K북마크 24
61위Arushi Gandhi@arushi_resslX에서 보기
AI가 코드로 영상을 만들어 경쟁하는 'Video Arena'가 무료로 공개됐습니다.
좋아요 124리포스트 21답글 17노출 13.3K북마크 57
It turns out code… is the best way… to make videos… ⁉️
So, to take it a step further: we’re excited to introduce Video Arena - watch AIs compete to code the video of your dreams. (We should really be calling it Code 2 Video Arena! 💀)
Now live at videoarena.newmeasure.ai.
For $0, you can prompt an idea and then, choose the video which best captures your taste! Build UI walkthroughs, data explainers, launch videos etc..
The arena is currently live with our harness partners @heygen, @remotion, @mosaic; and is supported by the @openai and @anthropic models.
Please try it out! Looking forward to what y’all build!
62위a16z@a16zX에서 보기
a16z: 상위 1% 기업이 상위 10% 기업보다 AI에 약 8배 더 지출한다고 밝혔습니다.
좋아요 159리포스트 20답글 19노출 29.5K북마크 69
AI power users are pulling away
The top 1% of companies now outspend the top 10% on AI by ~8x
More charts in State of Markets II: a16z.news/p/state-of-mar…
David George@DavidGeorge83
Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - a16z.news/p/state-of-mar…
@a16z @sarahdingwang @aleximm @santiago__rdz
63위Mehmet INCE@mdisecX에서 보기
Grok에게 NVIDIA Broadcast급 GPU 웹캠 배경 제거 프로그램을 요청하자 1시간 15분 만에 만들어 냈다며 놀라워했습니다.
좋아요 509리포스트 14답글 39노출 55K북마크 214
what the actual fuck ?!
The prompt I use is:
hey @grok i need a gpu-based webcam background removal software that’s as good as NVIDIA Broadcast. write me a gpu based webcam background removal.
1h15minutes later.
Yagiz Nizipli@yagiznizipli
“hey grok write me a gpu based webcam background removal”
64위a16z@a16zX에서 보기
a16z: AI 에이전트가 사람보다 토큰을 약 5배 소비하며 2월 이후 14배 늘었다고 밝혔습니다.
좋아요 507리포스트 68답글 41노출 76.6K북마크 210
Humans are the minority user of AI
Agents burn nearly 5x the tokens people do, up 14x since February
More charts in State of Markets II: a16z.news/p/state-of-mar…
David George@DavidGeorge83
Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - a16z.news/p/state-of-mar…
@a16z @sarahdingwang @aleximm @santiago__rdz
65위Miguel de Icaza ᯅ🍉@migueldeicazaX에서 보기
미겔 데 이카사: 데이비드의 디자인 트윗은 실패 사례조차 통찰이 있다고 칭찬했습니다.
좋아요 576리포스트 15답글 18노출 232.5K북마크 237
David’s design tweets are always the highlight of my day. He finds elegance and purpose and shares it with us.
Today he shares a miss, just as insightful as always.
David Hill@dwhilldesign
#designoftheday Jony Ive & Marc Newson’s Ferrari Luce EV is a fascinating example of notable designers completely missing the mark. The surfaces are clean, details obsessively resolved & the reduction admirable, but where is the Ferrari? Its restrained, almost appliance-like form suppresses the visual drama, tension & emotional excess that have long made a Ferrari irresistible. For a marque built on speed, passion & visual bravado, the Luce looks surprisingly timid. shorturl.at/aMm11
66위a16z@a16zX에서 보기
a16z: 2022년 이후 하이퍼스케일러 잉여현금흐름은 2,750억→0달러, 반도체는 500억→4,000억 달러로 이동했다고 분석했습니다.
좋아요 318리포스트 43답글 22노출 54.1K북마크 119
The AI buildout in one chart: Big Tech's profits have become chipmakers' profits.
Approximate free cash flow, from 2022 to today:
- Hyperscalers: $275B → $0
- Semiconductors: $50B → $400B
More charts in State of Markets II: a16z.news/p/state-of-mar…
David George@DavidGeorge83
Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - a16z.news/p/state-of-mar…
@a16z @sarahdingwang @aleximm @santiago__rdz
67위Y Combinator@ycombinatorX에서 보기
Y Combinator가 워싱턴 D.C. 'Startup Industrial Base' 행사 영상을 공유했습니다.
좋아요 402리포스트 58답글 27노출 50.5K북마크 147
The Startup Industrial Base: Building for the Next 250 — Washington, D.C. x.com/i/broadcasts/1…
68위Y Combinator@ycombinatorX에서 보기
Y Combinator가 로보틱스·자율주행 등 피지컬 AI 초기 창업자를 위한 SF 행사(Physical Intelligence 공동창업자 대담)를 연다고 알렸습니다.
좋아요 74리포스트 10답글 13노출 24.4K북마크 27
We're hosting an evening in SF with @oakhcft and π for early-stage builders in physical AI: robotics, autonomy, and intelligent hardware.
Join us for a fireside chat with @physical_int co-founder @QuanVng on what it takes to build in the physical world.
events.ycombinator.comPhysical AI Builders with YC, Oak, π
69위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥: 소프트웨어 제품에 관한 '아침 목욕가운 폭언' 영상을 올렸습니다.
좋아요 448리포스트 32답글 73노출 50.5K북마크 159
*taps the sign*
David George@DavidGeorge83
Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - a16z.news/p/state-of-mar…
@a16z @sarahdingwang @aleximm @santiago__rdz
71위Brian Chesky@bcheskyX에서 보기
브라이언 체스키: 에어비앤비가 친구 기반 여행 아이디어·자연어 AI 검색 등 업데이트를 내고 에이전트형 UI를 곧 선보인다고 밝혔습니다.
좋아요 452리포스트 19답글 60노출 40.6K북마크 150
Hey everyone, we made some updates based on feedback from you. We’ll keep shipping, so I’ll post more updates soon
- connect with your friends to get trip ideas
- AI search in your own words (this is just a baby step towards agentic AI, which is a UI coming soon)
- get laundry done, rent baby gear, and more
72위Google Cloud@googlecloudX에서 보기
Google Cloud가 프라이빗 데이터센터·멀티클라우드·로컬에 배포 가능한 Spanner Omni를 정식 출시했습니다.
좋아요 76리포스트 25답글 1노출 9.1K북마크 25
Spanner Omni is now generally available!
The deploy-anywhere version of Spanner is here—ready to power your most demanding production workloads in your private data centers, in a multi-cloud deployment, or even for testing locally on a laptop → goo.gle/3VEej2H
73위Chaitanya@chaaaiX에서 보기
정형 검증(formal methods) 해커톤 'Vibecheck'를 열어 실제 소프트웨어를 만들고 검증하게 한다고 알렸습니다.
좋아요 94리포스트 28답글 10노출 15.7K북마크 30
We’re organizing Vibecheck, a formal methods hackathon with some of the best folks in the space - @maxvonhippel, @qd_forall, @nolanlwin, @jessemhan, @emiyazono, and @workersio.
You’ll get a weekend to build real production software, and formally verify it.
We’ll have people who really know their stuff around to help, and you can come with a team, find one there, or just hack on something yourself.
We are glad to have a phenomenal group of sponsors, including @harmonicmath, @mathematics_inc, @theoremlabs, @thegp, @omnicom, @primeintellect, @astriolabs, @buildwithparty, @wearerandomlabs, and @lanyon_ai, as well as some more we will announce in the coming days!
If you’re a hacker, a formal methods person, or just someone who thinks “how do we know it works?” is an interesting question, we’d love to have you.
Signup Now: fmxai.org/vibecheck/
74위Patrick Collison@patrickcX에서 보기
패트릭 콜리슨: 스테이블코인 $OUSD 성장 추이를 확인할 수 있는 링크를 공유했습니다.
좋아요 201리포스트 19답글 12노출 34.3K북마크 63
Track $OUSD growth at .
ousd.fyiOUSD reserves
75위Ryan McGuire@r_mcguireX에서 보기
Aleph Surgery가 a16z 주도로 750만 달러 프리시드를 유치해 로봇 외과의를 개발합니다.
좋아요 386리포스트 56답글 77노출 104.1K북마크 119
@dylanmcguir3 님에게 보내는 답글
@dylanmcguir3 and I just raised a $7.5M pre-seed led by @a16z to build a robotic surgeon @alephsurgery
Much of the early investment in Physical AI has focused on industries where labor is relatively abundant and low-cost. We believe the biggest opportunities to change the world lie at the opposite end of the spectrum.
Surgery is among the most valuable forms of labor on the planet, but it's far too scarce and expensive - due to an exceptionally long training pipeline and highly concentrated expertise.
We’re building the first surgeon that scales: one system that can acquire, refine, and transfer surgical skills across millions of procedures. v1 is already learning to perform surgical tasks autonomously from demonstrations, and improve from its own experience.
Lots of great folks behind us! @BoxGroup, Reveille VC, Constellation, @forward_deploy, @Valkyrie_vc, @DiscipulusVent, @maniacvc + more
76위Paul Graham@paulgX에서 보기
폴 그레이엄: 오메가 cal 266 등 30T2 계열 수동 무브먼트 시계를 추천했습니다(IT 무관).
좋아요 179리포스트 7답글 26노출 58.5K북마크 51
77위Garry Tan@garrytanX에서 보기
개리 탄: 국방부 연구담당 차관이 300쪽 RFP가 로비스트·대형 업체용이라고 인정했다며 스타트업 조달 장벽을 지적했습니다.
좋아요 109리포스트 11답글 21노출 17.2K북마크 30
The Pentagon's 300-page RFP "was designed for a lawyer and a lobbyist and a prime contractor to know how to fill out."
That's Emil Michael, Under Secretary of War for Research & Engineering, saying it out loud to a room full of founders at our Startup Industrial Base event in DC today.
A 10-person startup reads 300 pages, gets one thing wrong, gets kicked out. Or survives and grinds through 3, 6, 9, 12 month rounds until it's two years in. Meanwhile the prime has a team whose whole job is filling out every page perfectly.
Emil’s fix? Write a one pager not 300. "I need to fly a missile 500 miles in this kind of environment." If your payload won't work, you hear that fast. Target is 90 days.
Drone Dominance runs the same way. Apply, top 20 go to a fly-off, objective metrics, cuts every round, and he says half a billion dollars of orders at the end. "I'd rather know if I lose and why." Miss round one and there are more rounds coming, and "it'll be 2 or 3" winners, not one.
Cost per intercept is the game now. Shoot down a $50k drone with a Patriot missile and "the math doesn't math." He's funding directed energy and high power microwave that cost thousands, not millions.
The tech was never what kept startups out. The 300 pages were. Two pages and a fly-off means the best product can actually win, and the prime has to compete like everyone else.
78위TypeSafe AI@typesafeaiX에서 보기
Typesafe AI가 Jev 기반 영업 데이터 리랭킹이 20배 빠르고 10배 싸며 12% 더 정확하다고 홍보했습니다.
좋아요 185리포스트 17답글 20노출 19.4K북마크 50
For the back of the room:
• production-grade reranking for sales data with Jev
• 20x faster
• 10x cheaper
• 12% more accurate
Are you getting it yet?
Rox@rox_ai
We used Jev to retrieve sales data 20x faster, 10x cheaper, and 12% more accurate than GPT-5 Mini.
Every time a Rox agent answers a query, it pulls from relevant transcripts, emails, CRM notes, news, and documents.
We benchmarked two ways of retrieving them:
LLM-based reranking
Jev classification
The results speak for themselves. Jev was faster, cheaper, and more accurate.
Revenue agents require thousands of points of context to understand accounts, chart relationships, and execute on sales. This research will help us serve that context with frontier-level performance for our customers.
79위Paul Graham@paulgX에서 보기
폴 그레이엄: 같은 실수를 두 번 하지 않으면 좁은 분야에서 완벽에 가까워질 수 있다며 보어의 말을 인용했습니다.
좋아요 127리포스트 10답글 14노출 15.2K북마크 34
@paulg 님에게 보내는 답글
The good news is, if some domain has a finite number of possible mistakes and you never make mistakes twice, you have a chance of approaching perfection eventually.
As Bohr said, "An expert is a man who has made all the mistakes which can be made in a very narrow field."
80위GitHub@githubX에서 보기
GitHub가 여러 모델로 초안·비평·수정을 라우팅하는 Project HydraFusion을 Copilot 앱과 VS Code에 제공합니다.
좋아요 247리포스트 44답글 25노출 39.8K북마크 66
👋 Project HydraFusion is now available in the GitHub Copilot app and @code.
Choose it like any other model. Behind the scenes it routes your task across multiple models to draft, critique, revise, or escalate, then hands you one result. github.blog/changelog/2026…
81위Visual Studio Code@codeX에서 보기
VS Code가 여러 모델·워크플로우를 조율해 코딩 작업을 처리하는 HydraFusion을 리서치 프리뷰로 공개했습니다.
좋아요 305리포스트 26답글 14노출 21.8K북마크 75
🧠 HydraFusion is now available in @code as a Research Preview.
Instead of relying on a single model, HydraFusion orchestrates multiple models and workflows to tackle coding tasks.
Give it a try today! 🚀
aka.ms/VSCode/Hydrafu…
82위Google@GoogleX에서 보기
Google: Gemini 스킬이 Gems를 대체할 예정이며 기존 Gems는 자동 이전되고 공유 기능도 지원된다고 밝혔습니다.
좋아요 143리포스트 8답글 12노출 35.2K북마크 35
@Google 님에게 보내는 답글
Skills are already available in Gemini Spark, and they will soon replace Gems to make building custom shortcuts even easier.
But don’t worry — we’ll automatically migrate your Gems into skills when Gems go away. And over the coming weeks, skills will have the same features you loved in Gems, including the ability to share them with others and add Google Drive files.
Learn more ↓
blog.googleLet skills in Gemini tackle your most repetitive tasks
83위clem 🤗@ClementDelangueX에서 보기
클레망 들랑그: 수천 건의 DM 지원을 받아 검토에 며칠 걸리며 뛰어난 매칭 위주로 답하겠다고 알렸습니다.
좋아요 356리포스트 15답글 51노출 41.1K북마크 86
I got thousands of DMs and it looks like @bot can't automatically analyze DMs (when @elonmusk?) so it will take me a few days to go through all of them, sorry!
If you don't get an answer, don't take it as a judgment (again thousands of applications so I'll focus on only exceptional matches for the moment) and don't hesitate to apply for specific positions on .
Thank you everyone, super excited to work more with more of you to keep pushing open-source AI! 🤗🤗🤗
apply.workable.comHugging Face
clem 🤗@ClementDelangue
getting acquired by @nvidia = hugging face can now hire people we couldn't as a small startup and give them a decade to make open-source AI win!
if you're one of them, my dms are open
For @openclaw, success really is the journey.
85위Garry Tan@garrytanX에서 보기
개리 탄: 숀 더피 교통부 장관과 YC 행사에서 항공 교통관제 현대화 등 향후 10년 과제를 논의했습니다.
좋아요 75리포스트 14답글 15노출 15.3K북마크 16
I asked Secretary Duffy what Americans should be able to do, if we get this decade right, that they can't do today.
Sean Duffy is the Secretary of Transportation, and he sat down with me at our Startup Industrial Base event in DC today. He went straight to how our country's air traffic gets managed.
At the command center, people sit in front of six screens with a paper flight chart above them, trying to make the calls on weather and routing.
His example: four planes reach the same runway at the same time and only one can land. See it two hours out, speed up two, slow one down, and each one lands when the runway is open. "Like, not complicated stuff, but we don't do any of this."
Paper and pencil and minimal human augmentation from software looks like this: Say an airport can take 15 planes in a 15 minute window. "Well, the airlines will schedule 32." He says he's handing the airlines that data and asking them to spread out.
The tool is SMART, from Air Space Intelligence, a small company he says competed with the biggest companies and won. It predicts. His stage example: a 2.5 hour ground stop might only need 45 minutes. He's taken heat for supposedly handing air traffic control to AI. Duffy's answer? "it's dumb AI, not smart AI." Controllers still make every call.
None of this needs a breakthrough. It needs the will of the people to know they need to use software and be ok with AI augmentation instead of insisting on a paper chart and copper wire.
86위Patrick Collison@patrickcX에서 보기
패트릭 콜리슨: 대규모 파트너 네트워크와 수익 대부분을 공유하는 개발자 친화적 USD 스테이블코인 Open Standard($OUSD)가 출시됐다고 알렸습니다.
좋아요 255리포스트 29답글 34노출 40.8K북마크 50
Open Standard, $OUSD, is now live!
Open Standard is a new USD stablecoin with a very large network of participating partners (joinopenstandard.com/partners). It's designed to be exceptionally secure, reliable, and developer-friendly. Importantly, most yield is shared directly with partners rather than being internalized by the stablecoin issuer.
Stripe is adopting OUSD in a number of core product flows.
joinopenstandard.comPartners | Open Standard
Open Standard@openstandard
$OUSD is live.
Build with OUSD via @Coinbase @Mastercard @Stripe and @Visa. Each integration path supports 1:1 USD conversion at no cost.
Read the announcement: joinopenstandard.com/blog/ousd-is-l….
Old GPUs are turning into appreciating assets
More charts in State of Markets II: a16z.news/p/state-of-mar…
David George@DavidGeorge83
Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - a16z.news/p/state-of-mar…
@a16z @sarahdingwang @aleximm @santiago__rdz
88위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥: 에이전트로 직접 만든 도구는 완벽할 필요 없이 필요한 기능만 그때그때 추가하면 된다고 말했습니다.
좋아요 76리포스트 2답글 13노출 8.2K북마크 14
One interesting aspect of writing your own tools with agents, is that the tools don't have to be bullet proof and complete. For example, in my bookwriter word processor, if there's something I want done to the text, like splitting a chapter in half, I don't have to invent a new feature. I can just tell an agent to adjust the book data.
This makes the tool a useful hybrid. I can use it for 95% of the work, and use an agent if there's something that's not worth a feature.
89위Tempo@tempoX에서 보기
Tempo: Open Standard의 OUSD가 출시와 함께 4억 달러 이상 유동성으로 Tempo에 네이티브 발행됐다고 밝혔습니다.
좋아요 248리포스트 37답글 15노출 82K북마크 42
Open USD (OUSD) is live on Tempo, with $400M+ in liquidity at launch.
Issued natively on Tempo, OUSD by @openstandard gives businesses a new stablecoin for payments, settlement, and treasury, with liquidity deployed across decentralized exchanges, stablecoin swaps, and bridges.
90위a16z@a16zX에서 보기
a16z: AI 지출 상위 1% 기업이 중간값 기업보다 600배 이상 더 쓴다며 도입 격차를 짚었습니다.
좋아요 101리포스트 11답글 11노출 17.1K북마크 17
Wild adoption gap: the top 1% of AI spenders are spending more than 600x as much as the median company
More charts in State of Markets II: a16z.news/p/state-of-mar…
David George@DavidGeorge83
Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - a16z.news/p/state-of-mar…
@a16z @sarahdingwang @aleximm @santiago__rdz
91위Daniel Han@danielhanchenX에서 보기
Unsloth가 Hugging Face와 함께 데모·굿즈가 있는 오픈소스 파티를 공동 개최한다고 알렸습니다.
좋아요 91리포스트 9답글 6노출 11.7K북마크 13
We're co-hosting an open source party with @HuggingFace! 🤗🦥
Come join us, we'll be handing out @UnslothAI merch, showcasing our upcoming features and more!
It'll be a super fun night with lots of demos, DJs, food and more.
Join: luma.com/OpenTogether
Use code: NOSLOTHSHERE to get in.
92위Matija Milenovic@mat_milenovicX에서 보기
Hop Aero가 로켓 화물 운송 개발을 위해 Initialized 주도로 1,100만 달러를 유치했습니다.
좋아요 51리포스트 8답글 12노출 8.5K북마크 7
We’re happy to share that @Hop_Aero has raised $11M to accelerate the development of rocket cargo.
Funding led by @Initialized, with participation from @ycombinator, @lunar_vc, System.One, Outsized Ventures and more. Also joined by strategic angels @paulg , @JasonMiller and more.
Time to get building 🚀
Open source is the only way to protect your projects.
Zixuan Li@ZixuanLi_
GLM-5.3 has helped defend 389 open-source projects, with 4,249 potential vulnerabilities found so far.
OpenVuln is still running. The service remains free, and findings go privately to maintainers.
huggingface.co/spaces/zai-org…
94위Paul Graham@paulgX에서 보기
폴 그레이엄: 공동창업자와 같이 사는 건 초기에 도움이 되지만 억지로 할 필요는 없다고 답했습니다.
좋아요 509리포스트 12답글 65노출 120.8K북마크 69
95위Garry Tan@garrytanX에서 보기
개리 탄: 스타트업이 미국 국방기술을 만드는 방법을 주제로 D.C. 행사를 생중계한다고 알렸습니다.
좋아요 96리포스트 10답글 18노출 23.1K북마크 13
Follow along today. We are live streaming from DC on how startups are helping build the future of defense tech in America
Y Combinator@ycombinator
The Startup Industrial Base: Building for the Next 250 — Washington, D.C. x.com/i/broadcasts/1…
96위DHH@dhhX에서 보기
DHH: 구형 맥북도 효율적인 OS(리눅스)를 올리면 훨씬 오래 쓸 수 있다고 말했습니다.
좋아요 695리포스트 17답글 50노출 85.8K북마크 91
It is indeed wild that we express more care for the MacBook vintage range than Apple. But these are lovely computers that have so much more life in them if we only feed them an efficient OS.
James Aita@jaita
It’s pretty insane to think that I have @NousResearch Hermes desktop app and Pi running no problem on my 2010 MBP via @dhh Omarchy - but I can’t even get chrome or safari to run properly on MacOS because Apple only supports High Sierra for those older laptops. If you have old gear sitting around that Apple and friends want you to get rid of so you buy more; Omarchy is your friends
97위clem 🤗@ClementDelangueX에서 보기
클레망 들랑그: MicroAGI 관련 링크를 이모지와 함께 응원했습니다.
좋아요 268리포스트 23답글 19노출 32.5K북마크 35
🔥🔥🔥 @microagi 🚀🚀🚀
98위Visual Studio Code@codeX에서 보기
VS Code 최신 릴리스가 멀티 저장소·워크트리 작업과 원격 에이전트 호스트로의 작업 위임을 쉽게 했습니다.
좋아요 108리포스트 4답글 2노출 9.9K북마크 14
🚀 The latest @code release makes it easier to work across repositories and worktrees and delegate tasks to remote agent hosts.
🧵 Here’s what’s new:
code.visualstudio.comVisual Studio Code 1.140 (Insiders)
99위Pollen Robotics@pollenroboticsX에서 보기
Pollen Robotics: 소형 로봇 Microduck 10,950대를 12월 10일~1월 10일 사이 주문순으로 출하한다고 밝혔습니다.
좋아요 87리포스트 15답글 13노출 17.4K북마크 11
100위a16z@a16zX에서 보기
a16z가 JD 밴스 부통령의 Saronic 조선소 'Port Alpha' 기공식 연설을 공유했습니다(정치·국방).
좋아요 61리포스트 10답글 5노출 19.1K북마크 6
Vice President JD Vance at the groundbreaking of Saronic's Port Alpha in Brownsville, Texas:
"We are rebuilding the American dream from the ground up by employing and investing in American workers and American businesses."
"What this facility will mean is that 10,000 families have a good paycheck. 10,000 sons and daughters will see mom and dad go to work with purpose. 10,000 people will know that their country is being built not by people outside of the United States, but by their own fellow citizens and their own people in their own community."
"So I just wanna say to the workers, to the investors, to the leadership of this company, I'm proud of you."
"This represents something that for 40 years, we saw factories close down. We saw shipyards close down. We saw our friends and neighbors lose their jobs because we had a government, but also a corporate leadership that didn't want to invest in the American worker anymore. Well, that has changed."
"As you guys break ground on this facility, as you build it up to great heights, and of course, as you start seeing ships roll off the lines, know that the President of the United States, the Vice President of the United States, we're proud of you, we're thankful for the work that you're doing, we're rooting for you, and you will be our champion here in Brownsville, Texas, and we will be your champion in Washington, DC."
@VP @Saronic
101위The Economist@TheEconomistX에서 보기
이코노미스트: 러시아의 가을 키이우 민간인 공격 계획과 버티는 키이우를 다룬 국제 정세 기사입니다.
좋아요 186리포스트 62답글 33노출 85.7K북마크 18
Russia’s autumn plan is to target civilians in Kyiv. But this escalation of terror was not unexpected—and the capital is not capitulating bit.ly/4yxi6NV
Photo: Getty Images
102위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥: 원자력·기후·AI 위기론을 좌파의 권력 획득 수단으로 비판한 정치 논평입니다.
좋아요 516리포스트 47답글 41노출 14.4K북마크 49
There is a cohort of people who love play "Chicken Little". They are typically associated with the left, and their goal is the acquisition of power.
They screamed "No Nukes" when nuclear power was the safest form of power generation ever created. They managed to delay the use of Nuclear power for decades. Now, of course, we really need it.
They screamed "Climate, Climate, Climate" and manage to increase the price of fossil fuels. They seemed oblivious to the fact that those Nukes they blocked would have prevented a lot of CO2 from being dumped into the air.
Now they are screaming "AI will kill us all". And they are having the same effect. They will impede the progress on the one thing that could help us solve all those other problems.
Why? Because they want the power that breeding fear in others gives them.
Should we be cautious with Nukes? Sure. Should we be cautious with the climate? Sure. Should we be cautious with AI? Sure.
But we should not let fear dominate us -- that fear will be our undoing.
103위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥: AOC의 타운홀 발언을 비판한 정치 게시물입니다.
좋아요 99리포스트 0답글 18노출 10.6K북마크 8
Can you believe that this person wants to be President? Here’s AOC from her "Students vs. Billionaires" town hall at the Ithaca Downtown Conference Center on Sunday, Sept 27, 2026, at 5 p.m. ET.
"We are going to encounter a lot of challenges. We're already starting to see it. Flash floods without warnings. We're seeing entire regions get wiped out in flash floods and wildfires. And in order to rebuild this, we are going to have to come together. Because what the elite want, and they're already planning for it, is to hoard all of the resources for themselves as the world gets more and more physically unstable. And in an environment of AI, they want to wipe out all of our livelihoods. They want to make sure none of us have jobs. And they want to keep all the spoils of those riches while consuming and burning energy at a pace that is completely unsustainable to make this worse."
104위Philip Johnston@PhilipJohnstonX에서 보기
Starcloud가 Firefly의 Elytra 우주선에 탑재돼 달 궤도 데이터센터 기술을 검증하는 계약을 맺었습니다.
좋아요 365리포스트 44답글 42노출 28.9K북마크 29
Starcloud is going to the moon! 🌚🚀
I'm excited to share that @Starcloud_ has signed an agreement with @FireflySpace to fly onboard Firefly's Elytra spacecraft in lunar orbit, validating capabilities for a future orbital lunar data center.
Integrated on Elytra, Starcloud’s SC-1L processor, including the @NVIDIA H100, will ingest data from Firefly’s onboard Solux vision system, perform high-power AI computing on orbit, and rapidly transmit the processed data back to Earth. The demonstration builds directly on our recent AI computing demonstration in low Earth orbit, extending those proven capabilities into lunar orbit, a strategic priority for both @NASA and the @USSpaceForce.
105위South America@SouthParkX에서 보기
사우스파크 신작 방송 홍보 게시물로 IT와 무관합니다(광고).
좋아요 196리포스트 16답글 4노출 312.6K북마크 15
107위Skyler Chan@skyler_chan_X에서 보기
미국 달 기지를 벽돌부터 쌓자며 NASA 국장에게 제안하는 게시물입니다.
좋아요 122리포스트 14답글 9노출 12.5K북마크 8
Sir, let's build America's Moon Base, brick by brick. @NASAAdmin @rookisaacman
Y Combinator@ycombinator
The Startup Industrial Base: Building for the Next 250 — Washington, D.C. x.com/i/broadcasts/1…
108위NASA Science@NASAScience_X에서 보기
NASA 과학국이 첫 달 전초기지 준비를 위한 PRISM 과학 조사 3건(위험 감시·자연 은신처·물 얼음 지도)을 선정했습니다.
좋아요 451리포스트 78답글 16노출 80.6K북마크 29
NASA has selected three new science investigations through its PRISM program to prepare for humanity’s first lunar outpost, @NASAMoonBase.
From monitoring hazards to finding natural shelters and mapping water ice, take a look at the science paving the way for lunar exploration: go.nasa.gov/4iS0h7k
109위Google@GoogleX에서 보기
Google: Gemini 4 Argon을 Fairwind 프로그램으로 사이버 방어 기관에 먼저 배포하고 가드레일을 개선한다고 밝혔습니다.
좋아요 1.2K리포스트 75답글 45노출 158.3K북마크 74
@Google 님에게 보내는 답글
Gemini 4 Argon is rolling out to an initial cohort of cyber defenders through our Fairwind Program so they can leverage its full frontier-level cybersecurity defense capabilities.
We'll continue to gather feedback from early testers as we iterate on guardrails before making Argon available to developers, enterprises, and consumers as soon as possible.
Learn more ↓
blog.googleGemini 4 Argon: our next era of frontier intelligence
110위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥: 원자력 반대·기후 재앙·AI 둠론에 같은 패턴이 있다고 꼬집었습니다.
좋아요 374리포스트 18답글 37노출 33.3K북마크 19
111위NASA@NASAX에서 보기
NASA가 10월 5~9일 튀르키예 안탈리아 국제우주대회(IAC)에서 아르테미스 협정·상업 우주를 알린다고 밝혔습니다.
좋아요 474리포스트 59답글 47노출 264.7K북마크 16
We'll be at the 77th International Astronautical Congress (IAC) in Antalya, Türkiye Oct. 5-9, highlighting America's leadership in human exploration, the Artemis Accords, and support for the commercial space sector: go.nasa.gov/4yihZWm
112위Kyle@kyletruongX에서 보기
X 입사 첫 주에 X Chat API 엔드포인트 23개를 새로 공개했다는 소식입니다.
좋아요 124리포스트 9답글 8노출 10.1K북마크 4
First week at @X done and first big ship: introducing 23 new X Chat endpoints!
Developers@XDevelopers
New X Chat endpoints are now available in X API v2.169.
You can now manage groups, admins, invites, join requests, members, mute, TTL, pinned messages, and chat settings from the X API.
More details here: docs.x.com/xchat/introduc…
113위clem 🤗@ClementDelangueX에서 보기
클레망 들랑그: Microduck 생산 라인이 궁금하다며 다음 로봇엔 팔을 달아 생산에 참여시키자고 농담했습니다.
좋아요 82리포스트 6답글 17노출 9.7K북마크 2
would love to see what the microduck production and assembly line looks like!
Our next robot should have arms so that it can participate in the production of microducks 😅😅😅
Pollen Robotics@pollenrobotics
The Microduck adventure continues, and the shipping dates are close enough to feel real now. We are proud to tell you that 10,950 Microducks will come off the line between December 10 and January 10!
Weekly batches, earliest orders first. Every date is in here.
114위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥: 기후변화·원자로·데이터센터 때문에 지구가 넘어진다는 식의 위기론을 풍자했습니다.
좋아요 62리포스트 1답글 12노출 5.7K북마크 1
Y'know, the Earth spins like a top, and ... well, you know how tops sometimes tip over? That's what the Earth is going to do soon. It's because of Climate Change, Nuclear Reactors, and all the Data Centers that are making the Earth unbalanced!. So it's going to tip soon. And when it does, the Elites who have hoarded all the Earth's resources and have ruined all our jobs, will be hiding in bunkers with many fertile and nubile ... Wait ... No, that was a different movie.
115위DHH@dhhX에서 보기
DHH: KAYAK 같은 사이트에서 스타링크 탑재 노선만 필터링할 수 있으면 좋겠다고 제안했습니다.
좋아요 158리포스트 1답글 10노출 8.6K북마크 1
116위Martin Fowler@martinfowlerX에서 보기
마틴 파울러: 발표자를 보완하는 시각 채널로서 슬라이드를 만드는 원칙을 다룬 글을 소개했습니다.
좋아요 48리포스트 4답글 3노출 7K북마크 51
When slides earn their place they work as a visual channel that complements the speaker rather than replacing them - Sumeet Gayathri Moghe sets out the resulting principles.
martinfowler.comPrinciples for effective slides
117위Google Cloud Tech@GoogleCloudTechX에서 보기
Google Cloud Tech가 클라우드에서 오픈 모델을 운영할 때 비용을 줄이는 팁 3가지를 공유했습니다.
좋아요 43리포스트 6답글 4노출 17K북마크 32
@ycombinator 님에게 보내는 답글
119위a16z@a16zX에서 보기
a16z: Anthropic·OpenAI·Databricks·Stripe 등 대형 기업이 비상장을 유지하는 이유와 'IPO는 또 하나의 자금 조달'이라는 관점을 다뤘습니다.
좋아요 22리포스트 4답글 5노출 7.6K북마크 12
"The founder is the asset class... The IPO is another financing event."
a16z's David George and Santiago Rodriguez on why the biggest companies are staying private:
Santiago: "If you look at the top six companies today, Anthropic, OpenAI, Databricks, Stripe, Waymo, and Revolut, by last round valuation, they add up to about $2.4 trillion."
"This is more than the combined market cap of IPOs we've seen in the last 10 years, excluding SpaceX, which adds up to $1.7 trillion."
David: "Especially for founder-led companies, one of the things that we've talked about and that I've written about is that the founder is the asset class at this point."
"The bet that we make, and part of the reason that it's been a benefit for some of these companies to remain private, is they can many times take bigger swings in the private markets that have longer duration paybacks."
"Zuck and Elon are kind of obvious exceptions to the rule in the public markets."
"You see it manifest in the numbers in companies like Databricks or Stripe: massive, massive new bets in new product areas, and you can see revenue acceleration that happens as a result."
"You can do that in the public markets, but it's gonna catch greater scrutiny. Meta's stock price at the nadir got below 100 bucks a share when people were very, very skeptical about their investments in AR/VR."
@DavidGeorge83 @santiago__rdz
a16z@a16z
Five private companies are worth more than every tech IPO from 2015 to 2025 combined
More charts in State of Markets II: a16z.news/p/state-of-mar…
120위a16z@a16zX에서 보기
a16z: AI 인프라 구축이 '소프트웨어가 세상을 먹는다'의 다음 장이라며 데이터센터 설비투자를 근거로 들었습니다.
좋아요 35리포스트 4답글 8노출 13.3K북마크 17
a16z's David George and Alex Immerman on why the AI buildout is the next chapter of "Software Is Eating the World":
David: "You're seeing it in CapEx. It's heavily covered in the data center build-out. It's heavily covered in the amount of capital that's been going in to fund the model development."
"To put it into very deep historical context, because this is probably the analogy that we've seen the most of, this build-out just surpassed railroads as a percentage of GDP."
"We all happen to think that if you fast-forward five to seven, maybe 10 years from now, we'll be looking at these numbers, and they'll probably be 20x higher cumulatively."
Alex: "This does feel like the next chapter of Marc's 'Software Is Eating the World.' Over the last 15 years, software has transformed all these industries, but only a small fraction of the population could build it. None of us could."
"But today, look at us. We all have a handful of automations running every night. If you think about the software demand, that means a lot more compute, chips, power, construction."
@DavidGeorge83 @aleximm
a16z@a16z
a16z's David George, Sarah Wang, Alex Immerman, and Santiago Rodriguez on the State of Markets, from AI and infrastructure to the next investment cycles:
1. Tech is the everything cycle.
High-tech equipment, software, and R&D are now about 55% of US capital spending, and the AI buildout just passed the railroads as a share of GDP. The five largest hyperscalers will spend about $780B on capex this year, heading past $1T in 2027, and demand still outstrips supply at every point in the chain.
2. It's an earnings story.
Stocks are up about 20% while multiples are down about 20%, and the S&P trades under 20x earnings. The market is up 90% since ChatGPT, on 15% earnings growth.
3. AI adoption is broad and shallow.
69% of the S&P 500 has a live AI deployment, 30% report a quantified result, and just 2% have AI doing a job they'd notice if it stopped. The power users are pulling away: the top 1% of AI spenders spend eight times the top 10%.
4. The private markets are enormous.
Six private companies, Anthropic, OpenAI, Databricks, Stripe, Waymo, and Revolut, add up to about $2.4T by last round, more than every IPO of the past decade combined. "The founder is the asset class. The IPO is another financing event."
Plus: why a data center can lower your electricity bill, what Amazon saying no to Muse and Instacart saying yes tells you about marketplaces, and the seven areas the team is most excited about next.
0:59 Charts from State of Markets II
2:59 Tech is the everything cycle
4:56 Is it a bubble?
6:31 Hyperscaler CapEx: $1T next year
12:12 CapEx is someone else's order book
14:36 Data centers can lower your bills
17:06 OpenAI and Anthropic vs. some of the best software companies ever
18:22 Enterprise AI: 69% adopted, 2% measured
20:32 The top 1% spend 8x the top 10% on AI
24:36 Spending on AI for growth or for savings?
28:46 Only 2% of households pay for AI
31:36 Amazon blocks Muse, Instacart welcomes
36:16 Software traded growth for profit
42:56 6 startups vs. a decade of IPOs
45:46 Why employees are turning down liquidity
48:44 Seven areas they're excited about
YouTube: youtu.be/lr3hNhA0IfQ
@DavidGeorge83 @aleximm @santiago__rdz @sarahdingwang
121위Y Combinator@ycombinatorX에서 보기
Y Combinator가 D.C.에서 NASA 국장 재러드 아이작먼과의 대담을 생중계했습니다.
좋아요 17리포스트 3답글 7노출 18.8K북마크 7
Live from DC with NASA Administrator Jared Isaacman (@NASAAdmin)
x.com/i/broadcasts/1…
122위a16z@a16zX에서 보기
a16z 사라 왕: AI 지출이 전반적으로 늘지만 상위 1% 파워 유저와 나머지의 격차가 벌어지고 있다고 설명했습니다.
좋아요 38리포스트 7답글 10노출 12.1K북마크 15
AI spending is rising everywhere, but the power users are pulling away. a16z's Sarah Wang on the gap between the top 1% and everyone else:
"You can think of AI spending rising generally across companies, but the power users, the most intensive users, are spending much, much more."
"We looked at some YipitData for this, and if you look at median AI vendor spending in the top 1%, that's roughly eight times the level of the top 10%."
"Anecdotally, even inside our own portfolio companies, which you could say are pretty much fully AI-native or AI-pilled, you're seeing the top users spend anywhere from, call it, $7,500 to $9,000 a month... over 20 times the spend if you think about median versus top users."
@sarahdingwang @santiago__rdz
a16z@a16z
AI power users are pulling away
The top 1% of companies now outspend the top 10% on AI by ~8x
More charts in State of Markets II: a16z.news/p/state-of-mar…
123위BrendanEich@BrendanEichX에서 보기
브렌던 아이크: CarPlay는 해당 문제가 없지만 차량 OS에 블루투스로 직접 연결하면 위험이 남는다고 지적했습니다.
좋아요 40리포스트 2답글 2노출 9.3K북마크 15
Not true of CarPlay, although risks remain if you connect via BT directly to the car’s base OS phone or text app. x.com/i/grok/share/3…
Kacee Allen@KaceeRAllen
A federal court told Honda, Toyota, Volkswagen, and GM they can legally record and intercept your private texts and call logs the second you plug your phone into the car.
The infotainment system downloads everything. You can’t see it. You can’t delete it. Police can.
The judge said that’s not a privacy violation unless you prove it already damaged your reputation or your business.
You paid tens of thousands of dollars for a rolling wiretap. This is the surveillance state on four wheels.
124위a16z@a16zX에서 보기
a16z: 데이터센터가 오히려 지역 전기요금을 낮출 수 있다며 흔한 오해를 반박했습니다.
좋아요 37리포스트 8답글 6노출 14.3K북마크 12
A data center can lower your electricity bill. a16z's Alex Immerman, David George, and Sarah Wang on the myth that gets the least pushback and deserves the most:
"It may seem counterintuitive, but a data center can help lower your electricity rate. A recent US study showed that for every 10% increase in data center capacity, residential rates went down by 40 bps."
"You should think about it as a power grid is a shared fixed cost base: poles, wires, substations. A large stable customer, like a data center, can help spread those costs across more units of electricity. Simply put, more demand across a shared system is a positive."
"Investment in shared infrastructure can also improve the economics for the households and businesses connected to it."
"I don't think that point gets enough media attention. Dina Powell went on a podcast recently from Meta and talked about the Louisiana site that they did, where they worked with the community to lower electricity costs. There's real examples of that today. It's not just lip service."
@aleximm @DavidGeorge83 @sarahdingwang
a16z@a16z
a16z's David George, Sarah Wang, Alex Immerman, and Santiago Rodriguez on the State of Markets, from AI and infrastructure to the next investment cycles:
1. Tech is the everything cycle.
High-tech equipment, software, and R&D are now about 55% of US capital spending, and the AI buildout just passed the railroads as a share of GDP. The five largest hyperscalers will spend about $780B on capex this year, heading past $1T in 2027, and demand still outstrips supply at every point in the chain.
2. It's an earnings story.
Stocks are up about 20% while multiples are down about 20%, and the S&P trades under 20x earnings. The market is up 90% since ChatGPT, on 15% earnings growth.
3. AI adoption is broad and shallow.
69% of the S&P 500 has a live AI deployment, 30% report a quantified result, and just 2% have AI doing a job they'd notice if it stopped. The power users are pulling away: the top 1% of AI spenders spend eight times the top 10%.
4. The private markets are enormous.
Six private companies, Anthropic, OpenAI, Databricks, Stripe, Waymo, and Revolut, add up to about $2.4T by last round, more than every IPO of the past decade combined. "The founder is the asset class. The IPO is another financing event."
Plus: why a data center can lower your electricity bill, what Amazon saying no to Muse and Instacart saying yes tells you about marketplaces, and the seven areas the team is most excited about next.
0:59 Charts from State of Markets II
2:59 Tech is the everything cycle
4:56 Is it a bubble?
6:31 Hyperscaler CapEx: $1T next year
12:12 CapEx is someone else's order book
14:36 Data centers can lower your bills
17:06 OpenAI and Anthropic vs. some of the best software companies ever
18:22 Enterprise AI: 69% adopted, 2% measured
20:32 The top 1% spend 8x the top 10% on AI
24:36 Spending on AI for growth or for savings?
28:46 Only 2% of households pay for AI
31:36 Amazon blocks Muse, Instacart welcomes
36:16 Software traded growth for profit
42:56 6 startups vs. a decade of IPOs
45:46 Why employees are turning down liquidity
48:44 Seven areas they're excited about
YouTube: youtu.be/lr3hNhA0IfQ
@DavidGeorge83 @aleximm @santiago__rdz @sarahdingwang
125위GitHub@githubX에서 보기
GitHub가 가장 많은 스타를 받은 저장소 중 하나의 탄생 이야기를 소개했습니다.
좋아요 10리포스트 1답글 3노출 13.3K북마크 3
@github 님에게 보내는 답글
Learn more about the making of one of GitHub's most starred repos.
youtube.comOpenClaw Went Viral. Meet the Maintainers Building and Securing It.
126위a16z@a16zX에서 보기
a16z: 비상장 기업 5곳의 가치가 2015~2025년 모든 기술 IPO를 합친 것보다 크다고 밝혔습니다.
좋아요 41리포스트 4답글 8노출 15.3K북마크 12
Five private companies are worth more than every tech IPO from 2015 to 2025 combined
More charts in State of Markets II: a16z.news/p/state-of-mar…
David George@DavidGeorge83
Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - a16z.news/p/state-of-mar…
@a16z @sarahdingwang @aleximm @santiago__rdz
127위Nasdaq Exchange@NasdaqExchangeX에서 보기
나스닥이 Accelevation($ACCV)의 상장을 환영했습니다.
좋아요 41리포스트 0답글 4노출 2M북마크 12
128위Mike Hodara@maradoh22X에서 보기
Kuli AI가 브랜드에 맞는 크리에이터를 24시간 발굴해 Slack으로 전달하는 스카우팅 에이전트를 출시했습니다.
좋아요 31리포스트 3답글 5노출 21K북마크 8
Introducing @kuli_ai Scouting Agent, your personal creator scout that understands your brand and works around the clock.
It deeply vets creators and keeps bringing profiles to your Slack while you’re busy doing something else.
129위Henry Schuck@HenryLSchuckX에서 보기
ZoomInfo가 평이한 영어로 쓴 업무 프로세스를 에이전트 팀으로 바꿔 주는 플랫폼 기업을 인수한다고 밝혔습니다.
좋아요 39리포스트 2답글 5노출 29.6K북마크 10
ZoomInfo is acquiring DoubleO.ai.
@OzBad03 and @karthiksuresh built a platform that turns a business process written in plain English into a coordinated team of agents—each with the context and tools to do its part, with checks along the way.
For GTM engineers, the architecture matters.
Too much of today’s GTM automation still runs on a decades-old pattern: database polling.
A spreadsheet calls a CRM, a people API, a hiring API, an intent API. Then calls them again.
“Anything changed?”
Poll. Compare. Wait. Repeat.
The API calls add up. The credits add up. And execution waits for the next check.
GTM needs an event-driven architecture.
A champion changes companies. An account starts hiring. A buyer visits your website.
The signal arrives. An agent team activates.
Qualify the account → understand prior engagement → write relevant outreach → route to the right rep → send an email or queue a call with a tailored talk track → update the CRM.
That’s why we’re bringing DoubleO’s agent orchestration together with ZoomInfo’s B2B data and events, built on a foundation that processes 1.5B+ signals daily.
The signal tells you something happened. The context tells you why it matters. The agents carry the work through to action.
Giving GTM teams the context, agents, and interfaces to build.
130위Paul Graham@paulgX에서 보기
폴 그레이엄: 성공한 창업팀 상당수가 함께 살았지만 핵심은 동거보다 서로를 좋아했다는 점이라고 분석했습니다.
좋아요 70리포스트 1답글 6노출 15.1K북마크 17
@paulg 님에게 보내는 답글
A surprising proportion of the most successful founding teams lived together at some point. But this isn't all proof that living together wins. Some of what it's showing is simply that the founders liked one another, and liking one another definitely wins.
131위Google Cloud Tech@GoogleCloudTechX에서 보기
Google Cloud Tech Agent Clinic 3화: LangGraph 에이전트의 멀티턴 회귀를 잡는 자동 평가 스위트를 60분 만에 만드는 4단계 프레임워크를 소개했습니다.
좋아요 32리포스트 5답글 2노출 9.3K북마크 7
Terminal test runs won't catch multi-turn agent regressions.
In Ep 3 of Agent Clinic, Dani Zamora & @matt_feroz build an automated eval suite for a LangGraph agent in 60 mins.
Here's the 4-step framework to go from vibes to benchmark ANY AI agent → goo.gle/4hAkdcz
132위TypeSafe AI@typesafeaiX에서 보기
Typesafe AI가 Product Hunt에서 창업자들이 요청 기능을 즉석 구현하는 'HYPERSHIP' 이벤트를 홍보했습니다.
좋아요 70리포스트 1답글 4노출 14.7K북마크 14
Get ready to HYPERSHIP!!
Founders are building the features you dream up for them right now on @ProductHunt , take your shot!
Product Hunt 😸@ProductHunt
Hypership Day is live on Product Hunt!
Today is all about shipping velocity. Users request features directly on launch pages, and makers build and deploy them in real time.
🗣️ Users: Drop your feature requests and bug reports on the launches.
🛠️ Makers: Ship features as fast as you can today to climb the leaderboard.
Whoever ships the fastest takes home the glory... and prizes from @typesafeai and @supabase!
133위olivia@livilouuuuuuX에서 보기
YC 국방 자문 에릭 알보그의 신간 '국방을 위한 창업' 인터뷰를 D.C.에서 진행했다는 소식입니다.
좋아요 27리포스트 4답글 4노출 2.9K북마크 4
it was an honor to interview YC’s defense consultant, eric alborg, about his forthcoming book “building for defense” here in DC today @ycombinator
134위Billy Thalheimer@billythalheimerX에서 보기
Regent가 미 해병대 전투연구소와의 시글라이더 계약을 연장했다고 밝혔습니다.
좋아요 48리포스트 6답글 5노출 10K북마크 7
Extending our contract shows the great execution of our team to date - and the serious need of @USMC for high speed maritime mobility.
Thankful for the great multi-year partnership with @WarfightingLab to accelerate getting Seagliders into service
REGENT@regentcraft
Our partnership with the Marine Corps War Fighting Laboratory continues. This next stage brings an additional $5 million in mission dollars, advancing on-water technology for the warfighter and furthering our nation's defense readiness.
We’re proving out the Viceroy Seaglider defense capabilities in real time, demonstrating for our Marine stakeholders how we will work to integrate further controls and operate in contested environments.
Read the full announcement: regentcraft.com/news/regent-ex…
135위Paul Graham@paulgX에서 보기
폴 그레이엄: 아침 식사 자리에서 가족·동료의 이름이 어떤 단위일지 정해 본 농담을 전했습니다.
좋아요 132리포스트 1답글 9노출 23.3K북마크 18
@paulg 님에게 보내는 답글
We talked about this over breakfast. We decided Trevor's would be (possibly excessive) generality in an engineering solution, Jessica's would be either curiosity about someone's character or degree of incapacitation when laughing, and mine would be amount of change in an idea.
👀
Oren Melamed@OrenMe
Seems like @code is starting to tell you: your agent primitives need to move.
There is now a new migration banner in the Chat panel, and it is part of a much bigger change happening under the hood: the move from the old Local harness to the new Agent Host architecture built around AHP.
This is not just about moving where an agent runs.
The old model was very VS Code-centric: prompts, custom agents, instructions and skills could live in VS Code-specific locations and the agent runtime lived inside the extension host.
The new Agent Host separates the agent runtime from the editor. Sessions can keep running when the window closes, be shared across VS Code windows, run remotely, and support different harnesses such as Copilot, CLI and Copilot Desktop App through a common session layer.
And that means some of our primitives need to move too.
Prompt files are being deprecated for Agent Host and migrated to Skills. User-level agents and instructions that lived in VS Code profile storage need to move to harness-supported locations. Even the old location settings are being deprecated.
The new migration experience can detect these things and guide you through moving or converting them, while keeping the originals unless you explicitly remove them.
The new banner is basically the first visible sign that this migration is becoming a real product workflow. Basically telling us - it's time to move on!!!!
If you have accumulated a lot of prompts, custom agents, instructions and skills over the last year, now is probably a good time to understand where they actually live and which harness owns them.
To summarize the shift - it isn't just:
VS Code Chat → Agent Host
It is:
VS Code-specific primitives → harness-native primitives.
And I think this is going to become increasingly important as agents stop being features inside an IDE and become runtimes that multiple clients can connect to.
Go run your migrations now 🏃♀️
137위Demis Hassabis@demishassabisX에서 보기
데미스 하사비스: Gemini 4 Argon을 만든 팀이 자랑스럽다며 블로그 글을 공유했습니다.
좋아요 137리포스트 4답글 6노출 10.3K북마크 16
@demishassabis 님에게 보내는 답글
So proud of the incredible team who have been working so hard and relentlessly on this! More info in the blogpost:
blog.googleGemini 4 Argon: our next era of frontier intelligence
139위lil’km@_lilkm_X에서 보기
Hugging Face가 대용량 로보틱스 데이터셋을 HF에 호스팅하는 방법을 소개했습니다.
좋아요 13리포스트 1답글 0노출 3.6K북마크 1
HF is the perfect place to host your big robotics datasets
Read how we do it here :
huggingface.coLeRobot × Hugging Face Buckets - a Hugging Face Space by imstevenpmwork
140위Phrasly@phrasly_aiX에서 보기
Phrasly가 Pangram·GPTZero 최신 탐지기를 통과한다는 AI 글 '휴머나이저' Ultra를 홍보했습니다(광고).
좋아요 56리포스트 2답글 2노출 13.5K북마크 2
141위No Labels@NoLabelsOrgX에서 보기
No Labels가 이스라엘 대상 테러 시도를 승객들이 막았다는 소식을 전한 정치·사회 게시물입니다.
좋아요 47리포스트 7답글 2노출 14.9K북마크 1
An apparent 9/11-style terrorist attack against Israel could have ended in catastrophe today.
Instead, passengers stepped up, put themselves at risk, and helped save lives. It was a powerful reminder that even in a divided world, people will still act for one another.
We should all be grateful that courage won out today.
142위NASA HQ PHOTO@nasahqphotoX에서 보기
NASA 사진: Crew-13 발사를 앞둔 SpaceX 팰컨 9과 드래곤이 SLC-40 발사대에 선 모습입니다.
좋아요 48리포스트 14답글 3노출 8.4K북마크 1
.@SpaceX's Falcon 9 rocket and Dragon spacecraft are seen at sunset at Space Launch Complex 40 as preparations continue for the launch of Crew 13 to @Space_Station. Launch is targeted for 11:10 a.m. ET on Oct. 1 📷: flic.kr/s/aHBqjD6nMS
143위DHH@dhhX에서 보기
DHH: 루프트한자 기내에 두고 내린 Omarchy 키 XPS14 노트북을 찾았다고 전했습니다.
좋아요 110리포스트 0답글 2노출 10.3K북마크 0
@dhh 님에게 보내는 답글
I hear that my beloved XPS14 with Omarchy keys has been found, and the kind folks from @Lufthansa_USA are working to assure it's safe return. I will take better care next time I'm jetlagged out of my mind and pack it before landing! 🙏
144위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥: 엘리트가 기후 재앙에 대비해 자원을 독점한다는 정치인 주장을 비판한 정치 게시물입니다.
좋아요 38리포스트 0답글 2노출 6K북마크 0
145위American Conservation Coalition Action@ACCActionX에서 보기
ACC Action이 상원에 인허가 개혁법 통과를 촉구하는 정치 캠페인 게시물입니다.
좋아요 20리포스트 2답글 1노출 25.5K북마크 0
147위Restoration News@NewsRestorationX에서 보기
위스콘신 민주당 주지사 후보의 인종 형평 정책을 비판하는 정치 게시물입니다.
좋아요 11리포스트 4답글 0노출 23.7K북마크 0
Racist policies alert:
Wisconsin: Radical ‘Racial Equity’ Activists Fuel David Crowley’s BIPOC Obsession
Democrat gubernatorial candidate lets activist nonprofit train staff and lace racial ideology into laws and policies.
The $26 million national nonprofit Race Forward teaches Milwaukee County employees that every decision, in every government department, must be measured by how it relates to “racial equity.” Citizens might think Democrat gubernatorial candidate and County Executive David Crowley is running the county. However, his actions largely stem from Race Forward, the unelected nonprofit that has trained employees, guided policy changes, and caused department heads to waste hours writing meaningless racial equity reports.
restoration-news.comWisconsin: Radical ‘Racial Equity’ Activists Fuel David Crowley’s BIPOC Obsession
148위BrendanEich@BrendanEichX에서 보기
브렌던 아이크: 1993년작 쥬라기 공원이 여전히 잘 버틴다고 답했습니다.
좋아요 11리포스트 2답글 2노출 2.9K북마크 0
@xah_lee 님에게 보내는 답글
Jurassic Park was 1993 and holds up well.
youtube.comT-Rex Ambush Scene - Jurassic Park (1993) Movie Clip HD
149위Fannie Mae@FannieMaeX에서 보기
Fannie Mae가 계약금 지원 프로그램 사례를 홍보했습니다(금융 광고).
좋아요 5리포스트 1답글 0노출 26.9K북마크 0
Down payment assistance helped Madison buy a home and supported eligible borrowers with $25 million in 2025. spr.ly/6019BGsjCb
150위Hugeicons@huge_iconsX에서 보기
Hugeicons가 'AI 슬롭 아이콘과 작별하라'며 아이콘 라이브러리를 홍보했습니다(광고).
좋아요 1리포스트 1답글 0노출 5.7K북마크 1
151위Google Cloud Tech@GoogleCloudTechX에서 보기
Google Cloud Tech: 오픈 모델 비용 최적화 글 전체 링크를 공유했습니다.
좋아요 5리포스트 0답글 0노출 1.8K북마크 3
@GoogleCloudTech 님에게 보내는 답글
🔗 Read the full article ↓
medium.comHow to Run Open Models in the Cloud Without Going Broke
152위Uncle Bob Martin@unclebobmartinX에서 보기
엉클 밥이 유대인 박해를 주제로 한 유튜브 에피소드를 공유한 정치·사회 게시물입니다.
좋아요 4리포스트 0답글 0노출 1.7K북마크 1
They're Trying To Eradicate the Jews | Ep. 63 via @YouTube
youtube.comThey're Trying To Eradicate the Jews | Ep. 63
153위Abridge@AbridgeHQX에서 보기
Abridge가 AI 임상 기록 도우미 Abridge Assistant가 정신건강 간호사의 진료 방식을 바꾼 사례를 소개했습니다.
좋아요 5리포스트 0답글 0노출 3.3K북마크 0
Abridge Assistant is changing how clinicians like Olivia Taylor, a psychiatric mental health nurse practitioner, deliver care at @deaconessnews: abridge.com/blog/deaconess…
154위Eric Schiszler@EricSchiszlerX에서 보기
루트 66 100주년과 도시·시골의 화합에 대한 개인 소회로 IT와 무관합니다.
좋아요 1리포스트 0답글 1노출 1.5K북마크 0
155위Visual Studio Code@codeX에서 보기
VS Code가 10월 8일 최신 기능·데모를 소개하는 VS Code Live를 연다고 알렸습니다.
좋아요 2리포스트 1답글 1노출 4.5K북마크 0
@code 님에게 보내는 답글
Join our VS Code Live on October 8! 🚀
We’ll share the latest updates, features, and demos from VS Code.
🔔 Set your reminder:
youtube.com🔴 VS Code Live: Release Recap
156위Google Cloud Tech@GoogleCloudTechX에서 보기
Google Cloud Tech 팁 3: GPU 1개·레플리카 1개로 하드웨어를 제한하고 오토스케일링을 묶어 비용을 예측 가능하게 하라고 조언했습니다.
좋아요 4리포스트 0답글 1노출 1.2K북마크 0
@GoogleCloudTech 님에게 보내는 답글
#3: Right-Size Your Hardware + Limit Auto-Scaling
Keep your monthly bill predictable by updating your settings to apply strict hardware limits, such as:
- capping your accelerator count at a single GPU
- locking your replica count to 1-1
- opting for no capacity reservations
157위BrendanEich@BrendanEichX에서 보기
브렌던 아이크: 유닉스 계열이 'malloc 실패' 모델에서 오버커밋·OOM kill 모델로 옮겨 간 역사를 설명했습니다.
좋아요 2리포스트 1답글 1노출 3K북마크 0
@HaraldFielker 님에게 보내는 답글
What year or era? Like all Unixen, we moved from old “malloc (sbrk) can fail” model to an overcommit and OOMkill model. At first for a NASA program as a kernel configurable, then change of default.
158위U.S. Global Investors@USFundsX에서 보기
U.S. Global Investors가 소액 적립식 투자 상품 ABC Investment Plan을 홍보했습니다(금융 광고).
좋아요 2리포스트 0답글 0노출 10.6K북마크 0
Big goals can start with small steps.
The ABC Investment Plan offers a lower minimum to get started, with monthly contributions that can put the power of compounding to work over time.
Explore the ABC Investment Plan: usfunds.com/invest-with-us/
@code 님에게 보내는 답글
🔎 Explore everything that's new in the full release notes:
Happy coding! 💙
code.visualstudio.comVisual Studio Code 1.140 (Insiders)
160위Alliance for a Better Future@BetterFuture_AIX에서 보기
AI는 부모를 대체하지 말고 가족을 도와야 하며 기본값은 중독이 아닌 보호여야 한다는 정책 캠페인 게시물입니다.
좋아요 0리포스트 0답글 0노출 1.5K북마크 0
161위Moffitt Cancer Center@MoffittNewsX에서 보기
모핏 암센터가 근침윤성 방광암 3중 요법의 생존·장기 보존 효과 연구를 소개했습니다(의료).
좋아요 0리포스트 0답글 0노출 858북마크 0
From the 2026 @ASTRO_org Annual Meeting poster session, Celina Chiodo, MD (@celinachiodo), shares research evaluating trimodality therapy for muscle-invasive bladder cancer, showing durable survival and organ preservation in select patients.
162위Konvu@KonvuIncX에서 보기
Konvu가 코딩 에이전트 세션 비용과 다음 프롬프트 예상 비용을 보여 주는 무료 오픈소스 맥 앱을 공개했습니다.
좋아요 0리포스트 0답글 0노출 782북마크 0
164위Maverik, Inc.@MaverikX에서 보기
Maverik 편의점의 C4 에너지 음료 할인 광고로 IT와 무관합니다(광고).
좋아요 0리포스트 0답글 0노출 3.5K북마크 0
165위Ernexa Therapeutics@ErnexaTxX에서 보기
Ernexa Therapeutics가 ERNA-101의 IND 지원 시험 안전성 결과를 발표한 바이오 기업 소식입니다(광고).
좋아요 0리포스트 0답글 0노출 51.1K북마크 0
개인 59
-
CNN Breaking News●@cnnbrk
뉴스·미디어 관련 계정. CNN 속보 전용 계정으로 큰 사건이 터지면 가장 빠르게 알림을 받을 수 있습니다.
팔로워 60.1M·게시물 86.1K·Everywhere
-
The New York Times●@nytimes
뉴스·미디어 관련 계정. 뉴욕타임스 공식 계정으로 정치·경제·기술 심층 보도를 공유합니다.
팔로워 53.9M·게시물 566.2K·New York City
-
Sachin Tendulkar●@sachin_rt
스포츠·인물 관련 계정. 인도 크리켓 전설 사친 텐둘카르의 계정으로 재단(STF India) 활동 소식을 전합니다.
팔로워 39M·게시물 5.8K·ÜT: 18.986431,72.823769
-
Hillary Clinton●@HillaryClinton
정치·인물 관련 계정. 전 미국 국무장관이자 2016년 민주당 대선 후보 힐러리 클린턴의 계정으로 정치·사회 이슈 의견을 밝힙니다.
팔로워 29.3M·게시물 15K·New York, NY
-
jack●@jack
테크·플랫폼 관련 계정. 트위터 공동창업자 잭 도시의 계정으로 비트코인·탈중앙 소셜 등 기술 철학을 공유합니다.
팔로워 12.4M·게시물 31K
-
Sundar Pichai●@sundarpichai
AI·빅테크 관련 계정. 구글·알파벳 CEO 순다르 피차이의 계정으로 Gemini 등 구글 AI 전략 발표를 직접 볼 수 있습니다.
팔로워 11.8M·게시물 2.8K
-
anand mahindra●@anandmahindra
산업·경영 관련 계정. 인도 마힌드라 그룹 회장 아난드 마힌드라의 계정으로 기술·혁신 사례를 자주 공유합니다.
팔로워 11.5M·게시물 27.7K·Mumbai, India
-
Richard Branson●@richardbranson
기업가·인물 관련 계정. 버진 그룹 창업자 리처드 브랜슨의 계정으로 창업·모험·우주여행 이야기를 공유합니다.
팔로워 11.5M·게시물 27.7K
-
Marc Andreessen 🇺🇸●@pmarca
스타트업·투자 관련 계정. a16z 공동창업자 마크 앤드리슨의 계정으로 AI·기술·정책에 대한 강한 의견을 공유합니다.
팔로워 6.6M·게시물 15.3K·Menlo Park, CA
-
Mark Zuckerberg●@finkd
AI·빅테크 관련 계정. 메타 창업자 겸 CEO 마크 저커버그의 개인 계정으로 Meta AI·Llama·하드웨어 발표를 직접 전합니다.
팔로워 2.5M·게시물 93
-
Chamath Palihapitiya●@chamath
스타트업·투자 관련 계정. Social Capital 창업자이자 All-In 팟캐스트 진행자 차마스 팔리하피티야의 계정입니다.
팔로워 2.4M·게시물 13.8K·In the Arena
-
David Sacks●@DavidSacks
스타트업·투자/정책 관련 계정. Craft Ventures 창업자이자 All-In 팟캐스트 진행자 데이비드 색스의 계정으로 미 대통령 과학기술자문위 공동의장을 맡고 있습니다.
팔로워 1.8M·게시물 14.5K
-
Apple Hub●@theapplehub
모바일·미디어 관련 계정. 애플 관련 루머·신제품 소식을 빠르게 전하는 비공식 뉴스 계정입니다.
팔로워 1.4M·게시물 9.2K·Los Angeles, CA
-
Android Central●@androidcentral
모바일 관련 계정. 안드로이드 전문 매체 Android Central로 기기 리뷰와 팁·뉴스를 다룹니다.
팔로워 1.3M·게시물 125.8K
-
Google Analytics●@googleanalytics
마케팅·데이터 관련 계정. 구글 애널리틱스 공식 계정으로 GA·태그 매니저 제품 업데이트를 전합니다.
팔로워 1.2M·게시물 13.5K·Most places
-
Google Earth●@googleearth
지도·AI 관련 계정. 구글 어스 공식 계정으로 지리공간 데이터셋·Earth Engine·AI 분석 도구를 소개합니다.
팔로워 1.1M·게시물 4.2K·Mountain View, CA
-
Fei-Fei Li●@drfeifei
AI·머신러닝 관련 계정. World Labs 공동창업자 겸 CEO이자 스탠퍼드 HAI 공동소장 페이페이 리의 계정으로 공간지능·컴퓨터 비전 연구를 공유합니다.
팔로워 1.1M·게시물 3.1K·Stanford, CA, U.S.A.
-
Ilya Sutskever●@ilyasut
AI·머신러닝 관련 계정. OpenAI 공동창업자 출신으로 Safe Superintelligence(SSI)를 이끄는 일리야 수츠케버의 계정입니다.
팔로워 915.7K·게시물 1.2K
-
Guillermo Rauch●@rauchg
웹 개발·클라우드 관련 계정. Vercel CEO 기예르모 라우치의 계정으로 Next.js·v0 등 AI 웹 개발 흐름을 공유합니다.
팔로워 891.2K·게시물 53.4K·San Francisco, CA
-
React●@reactjs
소프트웨어·엔지니어링 관련 계정. React 공식 계정으로 웹·네이티브 UI 라이브러리의 릴리스 소식을 전합니다.
팔로워 846.5K·게시물 2.8K
-
Vinod Khosla●@vkhosla
스타트업·투자 관련 계정. 코슬라 벤처스 창업자 비노드 코슬라의 계정으로 AI·기후 기술 투자 관점을 공유합니다.
팔로워 774.2K·게시물 9.8K
-
Alexandr Wang●@alexandr_wang
AI·머신러닝 관련 계정. Scale AI 창업자이자 메타 최고AI책임자(Meta Superintelligence Labs) 알렉산더 왕의 계정입니다.
팔로워 757.4K·게시물 3.3K·San Francisco, CA
-
Jim Fan●@DrJimFan
AI·로보틱스 관련 계정. NVIDIA 로보틱스 디렉터 짐 판의 계정으로 피지컬 AI·휴머노이드 연구(GEAR 랩) 소식을 공유합니다.
팔로워 590.2K·게시물 4.2K·Views my own. Contact →
-
AK●@_akhaliq
AI·머신러닝 관련 계정. Hugging Face(Gradio)의 AK가 운영하며 매일 새로운 AI 논문을 빠르게 소개해 줍니다.
팔로워 529.4K·게시물 47.6K
-
Bootstrap●@getbootstrap
웹 개발·디자인 관련 계정. HTML·CSS·JS UI 툴킷 Bootstrap 공식 계정으로 릴리스 소식을 전합니다.
팔로워 518K·게시물 2.1K·San Francisco, CA
-
Sebastian Raschka●@rasbt
AI·머신러닝 관련 계정. 『Build a Large Language Model From Scratch』 저자 세바스찬 라슈카의 계정으로 LLM·추론 연구를 쉽게 풀어 줍니다.
팔로워 515.9K·게시물 20.3K·United States
-
tobi lutke●@tobi
커머스·스타트업 관련 계정. 쇼피파이 CEO 토비 뤼트케의 계정으로 AI 활용 경영·개발 철학을 공유합니다.
팔로워 515K·게시물 9.7K·Canada
-
PyTorch●@PyTorch
AI·머신러닝/소프트웨어 관련 계정. 리눅스 재단 산하 딥러닝 프레임워크 PyTorch 공식 계정으로 릴리스·생태계 소식을 전합니다.
팔로워 514.8K·게시물 3.7K
-
Nat Friedman●@natfriedman
AI·스타트업 투자 관련 계정. 전 GitHub CEO이자 AI 투자자 냇 프리드먼의 계정입니다.
팔로워 485.8K·게시물 5.7K·California
-
hardmaru●@hardmaru
AI·머신러닝 관련 계정. Sakana AI 공동창업자 겸 CEO 데이비드 하(hardmaru)의 계정으로 AI 연구 소식을 공유합니다.
팔로워 438.2K·게시물 25.9K·Minato-ku, Tokyo
-
Jessica Livingston●@jesslivingston
스타트업·투자 관련 계정. Y Combinator 공동창업자이자 『Founders at Work』 저자 제시카 리빙스턴의 계정입니다.
팔로워 433.6K·게시물 2.1K
-
Ian Goodfellow●@goodfellow_ian
AI·머신러닝 관련 계정. GAN 창시자이자 『Deep Learning』 공저자 이안 굿펠로의 계정으로 현재 스텔스 스타트업을 공동창업했습니다.
팔로워 382.2K·게시물 3K·San Francisco, CA
-
DeepLearning.AI●@DeepLearningAI
AI 교육 관련 계정. 앤드루 응이 세운 AI 교육 기업 DeepLearning.AI 계정으로 강좌와 AI 뉴스레터(The Batch)를 전합니다.
팔로워 351.7K·게시물 4.3K·United States
-
Emad●@EMostaque
AI·오픈소스 관련 계정. Stability AI 창업자 에마드 모스타크의 계정으로 오픈소스 AI와 새 프로젝트(ii) 소식을 공유합니다.
팔로워 343.7K·게시물 21.4K·London
-
Soumith Chintala●@soumithchintala
AI·머신러닝 관련 계정. PyTorch 공동창시자 수미스 친탈라의 계정으로 Thinking Machines에서의 새 작업과 오픈소스 AI 의견을 공유합니다.
팔로워 326.5K·게시물 4.1K·New York City
-
Michael Seibel●@mwseibel
스타트업·투자 관련 계정. Y Combinator 전 CEO·명예 파트너 마이클 세이벨의 계정으로 창업 조언을 공유합니다.
팔로워 304.1K·게시물 6.3K·San Francisco
-
Lilian Weng●@lilianweng
AI·머신러닝 관련 계정. 전 OpenAI 연구원이자 기술 블로그 Lil'Log 저자 릴리안 웡의 계정입니다.
팔로워 295.9K·게시물 220·San Francisco, CA
-
Berkeley AI Research●@berkeley_ai
AI 연구 관련 계정. UC 버클리 AI 연구소(BAIR) 계정으로 최신 논문과 연구 성과를 공유합니다.
팔로워 295.2K·게시물 1.6K·Berkeley, CA
-
Stability AI●@StabilityAI
생성 AI 관련 계정. Stability AI 공식 계정으로 이미지·영상 등 멀티모달 생성·편집 모델 소식을 전합니다.
팔로워 262.4K·게시물 887
-
John Collison●@collision
핀테크·스타트업 관련 계정. 스트라이프 공동창업자 존 콜리슨의 계정으로 결제·인터넷 경제에 대한 관점을 공유합니다.
팔로워 244.3K·게시물 838·California / Ireland
-
Jürgen Schmidhuber●@SchmidhuberAI
AI·머신러닝 관련 계정. LSTM 공동 개발자 위르겐 슈미트후버의 계정으로 딥러닝 역사와 연구 성과를 공유합니다.
팔로워 221.4K·게시물 457·Switzerland, KSA
-
Stanford NLP Group●@stanfordnlp
AI·NLP 관련 계정. 스탠퍼드 자연어처리 연구 그룹 계정으로 NLP·LLM 연구 성과를 공유합니다.
팔로워 194.7K·게시물 15.9K·Stanford, CA, USA
-
Aran Komatsuzaki●@arankomatsuzaki
AI·머신러닝 관련 계정. GPT-J 등 초기 오픈 LLM 연구자 아란 코마츠자키의 계정으로 주목할 AI 논문을 매일 소개합니다.
팔로워 194K·게시물 6.9K
-
Wojciech Zaremba●@woj_zaremba
AI·머신러닝 관련 계정. OpenAI 공동창업자 보이치에흐 자렘바의 계정으로 OpenAI 재단의 AI 회복력 활동을 공유합니다.
팔로워 165.7K·게시물 818·San Francisco, CA
-
Chris Olah●@ch402
AI 해석가능성 관련 계정. Anthropic에서 신경망 역공학(해석가능성)을 연구하는 크리스 올라의 계정입니다.
팔로워 159.8K·게시물 5.6K·San Francisco, CA
-
Harrison Chase●@hwchase17
AI 개발 도구 관련 계정. LangChain 창업자 해리슨 체이스의 계정으로 에이전트·LLM 앱 개발 동향을 공유합니다.
팔로워 132.8K·게시물 17.6K
-
Yi Ma●@YiMaTweets
AI·머신러닝 관련 계정. 홍콩대 AI 석좌교수 마이(Yi Ma)의 계정으로 지능의 수학적 이론 연구를 공유합니다.
팔로워 127.1K·게시물 2.6K·Hong Kong and Berkeley
-
Sam Bowman@sleepinyourhat
AI 안전 관련 계정. Anthropic에서 정렬(alignment)·LLM을 연구하는 샘 보우먼(NYU 휴직 중)의 계정입니다.
팔로워 72.4K·게시물 2.9K·San Francisco
-
Georgi Gerganov●@ggerganov
AI·오픈소스 관련 계정. llama.cpp·ggml 개발자 게오르기 게르가노프의 계정으로 로컬 LLM 추론 소식을 공유합니다.
팔로워 72.1K·게시물 1.8K
-
Airbnb Engineering@AirbnbEng
소프트웨어·엔지니어링 관련 계정. 에어비앤비 엔지니어링 계정으로 기술 블로그와 오픈소스 소식을 전합니다.
팔로워 54K·게시물 1.9K·San Francisco, CA
-
Neel Nanda●@NeelNanda5
AI 해석가능성 관련 계정. Google DeepMind 기계적 해석가능성 리드 닐 난다의 계정으로 신경망 내부 이해 연구를 공유합니다.
팔로워 47.8K·게시물 5.4K·London, UK
-
Lance Martin●@RLanceMartin
AI·머신러닝 관련 계정. Anthropic 기술 스태프 랜스 마틴의 계정으로 에이전트·RAG 실전 팁을 공유합니다.
팔로워 40.7K·게시물 1.4K·San Francisco, CA
-
andy jones●@andy_l_jones
AI·머신러닝 관련 계정. Anthropic에서 엔지니어링·연구를 맡은 앤디 존스의 계정입니다.
팔로워 27.7K·게시물 2.3K·SF
-
Coefficient Giving●@coeff_giving
자선·AI 안전 관련 계정. 효과적 기부 단체 Coefficient Giving 계정으로 AI 안전 등 분야 지원 활동을 공유합니다.
팔로워 22.2K·게시물 939·San Francisco, CA
-
Owain Evans●@OwainEvans_UK
AI 안전 관련 계정. 비영리 AI 안전 연구소 Truthful AI 디렉터 오웨인 에반스의 계정으로 창발적 오정렬 연구를 공유합니다.
팔로워 21.8K·게시물 6.3K·Berkeley, CA
-
Zhuang Liu●@liuzhuang1234
AI·머신러닝 관련 계정. 프린스턴대 조교수 좡 리우의 계정으로 딥러닝·비전 모델 연구(ConvNeXt 등)를 공유합니다.
팔로워 12.9K·게시물 625·Princeton, NJ
-
Ben Mann●@8enmann
AI 안전 관련 계정. Anthropic 공동창업자 벤 맨의 계정으로 AI 안전에 대한 관점을 공유합니다.
팔로워 9.7K·게시물 285·San Francisco, CA
-
Ivan Zhang●@1vnzh
AI·머신러닝 관련 계정. 기업용 LLM 기업 Cohere 공동창업자 이반 장의 계정입니다.
팔로워 8.8K·게시물 6.1K·Toronto
-
Kamal Ndousse●@kandouss
AI·머신러닝 관련 계정. Anthropic AI 연구원 카말 은두스의 계정으로 사회적 학습 연구에 관심을 둡니다.
팔로워 4.1K·게시물 1K·San Francisco, CA
기업·기관 41
-
CNN●@CNN
뉴스·미디어 관련 계정. 미국 CNN 공식 계정으로 국제·정치·기술 주요 뉴스를 실시간으로 전합니다.
팔로워 61.8M·게시물 429.3K
-
X●@X
테크·플랫폼 관련 계정. X(구 트위터) 공식 계정으로 서비스 신기능·정책 공지를 가장 먼저 확인할 수 있습니다.
팔로워 60.7M·게시물 15.7K·everywhere
-
BBC Breaking News●@BBCBreaking
뉴스·미디어 관련 계정. BBC 속보 계정으로 세계 주요 사건의 속보 알림을 전합니다.
팔로워 50.2M·게시물 44.2K·London, UK
-
BBC News (World)●@BBCWorld
뉴스·미디어 관련 계정. BBC 국제뉴스 계정으로 세계 각지의 뉴스·분석을 전합니다.
팔로워 42.3M·게시물 411.8K·London, UK
-
SpaceX●@SpaceX
우주·항공 관련 계정. SpaceX 공식 계정으로 팰컨·스타십 발사와 스타링크 소식을 전합니다.
팔로워 41.9M·게시물 11.8K·Earth
-
National Geographic●@NatGeo
과학·미디어 관련 계정. 내셔널지오그래픽 공식 계정으로 자연·과학·탐험 콘텐츠를 공유합니다.
팔로워 28.1M·게시물 73.6K·Global
-
The Economist●@TheEconomist
뉴스·미디어 관련 계정. 영국 시사주간지 이코노미스트 계정으로 글로벌 경제·정치·기술 분석을 전합니다.
팔로워 26.3M·게시물 466.3K·London
-
Tesla●@Tesla
전기차·AI/로보틱스 관련 계정. 테슬라 공식 계정으로 전기차·에너지 저장·자율주행·로봇 소식을 전합니다.
팔로워 24.8M·게시물 10.6K
-
The Wall Street Journal●@WSJ
뉴스·미디어 관련 계정. 월스트리트저널 공식 계정으로 경제·시장·기업·기술 뉴스를 전합니다.
팔로워 22.3M·게시물 508.7K·New York, NY
-
Forbes●@Forbes
경제·미디어 관련 계정. 포브스 공식 계정으로 기업·부호·창업 관련 기사와 뉴스레터를 소개합니다.
팔로워 20.3M·게시물 681.8K·New York, NY
-
United Nations●@UN
국제기구 관련 계정. 유엔(UN) 공식 계정으로 평화·인권·기후 등 국제 현안 소식을 전합니다.
팔로워 16.1M·게시물 104.7K·New York, NY
-
TechCrunch●@TechCrunch
스타트업·투자/뉴스·미디어 관련 계정. 테크크런치 공식 계정으로 스타트업 투자·창업·기술 뉴스를 전합니다.
팔로워 11M·게시물 287.8K·San Francisco, CA
-
Android●@Android
모바일 관련 계정. 안드로이드 공식 계정으로 OS 업데이트와 새 기능 소식을 전합니다.
팔로워 10.2M·게시물 9K·Mountain View, CA
-
Apple●@Apple
빅테크·하드웨어 관련 계정. 애플 공식 계정으로 아이폰·맥 등 신제품과 이벤트 소식을 전합니다.
팔로워 10M·Cupertino, CA
-
Windows●@Windows
소프트웨어·OS 관련 계정. 마이크로소프트 Windows 공식 계정으로 Copilot 통합 등 OS 기능 소식을 전합니다.
팔로워 8.1M·게시물 417.5K·Redmond, Washington
-
ISRO●@isro
우주·항공 관련 계정. 인도우주연구기구(ISRO) 공식 계정으로 인도의 위성·탐사 미션 소식을 전합니다.
팔로워 8.1M·게시물 2.6K·Bengaluru
-
Google Play●@GooglePlay
모바일 관련 계정. 구글 플레이 공식 계정으로 앱·게임·영화 추천과 스토어 소식을 전합니다.
팔로워 6.6M·게시물 254.4K
-
Gmail●@gmail
생산성·서비스 관련 계정. Gmail 팀 공식 계정으로 새 기능과 사용 팁을 전합니다.
팔로워 6.4M·게시물 373.6K·Mountain View, CA
-
Chrome●@googlechrome
웹·브라우저 관련 계정. 구글 크롬 브라우저 공식 계정으로 새 기능과 업데이트 소식을 전합니다.
팔로워 6.2M·게시물 184.7K·Mountain View, California
-
The Verge●@verge
테크·미디어 관련 계정. 기술 전문 매체 더 버지(The Verge) 계정으로 IT 기기·플랫폼·AI 뉴스를 다룹니다.
팔로워 4.7M·게시물 254.1K·New York
-
MarketWatch●@MarketWatch
경제·미디어 관련 계정. 마켓워치 계정으로 시장·개인 금융 뉴스와 논평을 전합니다.
팔로워 4.6M·게시물 458.4K·New York, NY
-
Google Maps●@googlemaps
지도·AI 서비스 관련 계정. 구글 지도 공식 계정으로 AI 기반 길찾기·장소 검색 기능 업데이트를 전합니다.
팔로워 4.1M·게시물 158.7K
-
Apple TV●@AppleTV
미디어·스트리밍 관련 계정. 애플 TV 공식 계정으로 오리지널 시리즈와 F1·MLS 중계 소식을 전합니다.
팔로워 3.6M·게시물 23.9K·Cupertino, CA
-
Google AI●@GoogleAI
AI·머신러닝 관련 계정. 구글 AI 공식 계정으로 연구 성과와 Gemini 등 모델 발표를 정리해 줍니다.
팔로워 2.5M·게시물 3.4K·Mountain View, CA
-
Android Developers●@AndroidDev
모바일·개발 관련 계정. 안드로이드 개발자 공식 계정으로 개발 모범 사례·튜토리얼·라이브 데모를 전합니다.
팔로워 2.1M·게시물 9.6K·Mountain View, CA
-
Microsoft Azure@Azure
클라우드·인프라 관련 계정. 마이크로소프트 Azure 공식 계정으로 클라우드·AI 서비스 소식과 자료를 전합니다.
팔로워 1.1M·게시물 51.1K·Redmond, WA
-
Google Workspace●@GoogleWorkspace
생산성·AI 관련 계정. 구글 워크스페이스 공식 계정으로 Gemini가 통합된 Gmail·문서·미트 업무 기능을 소개합니다.
팔로워 1M·게시물 34.9K·Mountain View, CA
-
Microsoft Cloud@MSCloud
클라우드·AI 관련 계정. 마이크로소프트 클라우드 계정으로 고객 사례와 실제 AI 도입 활용 사례를 소개합니다.
팔로워 726K·게시물 17.8K·Redmond, WA
-
Microsoft Support@MicrosoftHelps
고객지원 관련 계정. 마이크로소프트 고객 서비스·지원 공식 계정으로 제품 문제 해결 안내를 제공합니다.
팔로워 590.2K·게시물 532.2K
-
Publisher Connections@PubConnections
광고·애드테크 관련 계정. 퍼블리셔·개발자·광고 구매자를 연결하는 도구와 인사이트를 제공하는 Publisher Connections 계정입니다.
팔로워 579.9K·게시물 4.8K
-
Docker●@Docker
소프트웨어·인프라 관련 계정. 도커 공식 계정으로 컨테이너와 AI 에이전트를 안전하게 빌드·실행하는 도구 소식을 전합니다.
팔로워 556.9K·게시물 27.2K·🌎 Worldwide
-
Microsoft Developer@msdev
소프트웨어·엔지니어링 관련 계정. 마이크로소프트 개발자 공식 계정으로 Build 행사 소식과 개발 가이드를 전합니다.
팔로워 528.3K·게시물 37.5K
-
MongoDB●@MongoDB
데이터베이스·AI 관련 계정. MongoDB 공식 계정으로 AI 시대 데이터 플랫폼 기능과 개발 자료를 소개합니다.
팔로워 503.9K·게시물 24.9K
-
Angular@angular
소프트웨어·엔지니어링 관련 계정. 구글의 웹 프레임워크 Angular 공식 계정으로 릴리스와 개발 가이드를 전합니다.
팔로워 494.8K·게시물 4.9K
-
Chrome for Developers●@ChromiumDev
웹 개발 관련 계정. 구글 크롬 개발자 공식 계정으로 빠르고 안전한 웹 개발을 위한 새 API·도구를 소개합니다.
팔로워 424.4K·게시물 12.1K·Mountain View, CA
-
LangChain●@LangChain
AI 개발 도구 관련 계정. LangChain 공식 계정으로 LangSmith·LangGraph 등 에이전트 개발 도구 소식을 전합니다.
팔로워 268.2K·게시물 11.8K
-
Firebase@Firebase
앱 개발·AI 관련 계정. 구글 Firebase 공식 계정으로 생성 AI를 활용한 앱 개발·운영 도구 소식을 전합니다.
팔로워 224.4K·게시물 6.4K·The Internet
-
Engineering at Meta●@Meta_Engineers
소프트웨어·엔지니어링 관련 계정. 메타 엔지니어링 공식 계정으로 대규모 시스템 문제를 푸는 기술 블로그 글을 전합니다.
팔로워 217.4K·게시물 1.3K
-
Meta Open Source●@MetaOpenSource
오픈소스 관련 계정. 메타 오픈소스 공식 계정으로 PyTorch·React 등 메타 오픈소스 프로젝트 소식을 전합니다.
팔로워 158.5K·게시물 2.5K·Menlo Park, California
-
Spotify Engineering●@SpotifyEng
소프트웨어·엔지니어링 관련 계정. 스포티파이 엔지니어링 공식 계정으로 사내 기술 블로그·오픈소스 소식을 전합니다.
팔로워 47.5K·게시물 1.5K·Anywhere
-
Apollo Research●@ApolloResearch
AI 안전 관련 계정. 프런티어 AI의 개발·배포·거버넌스 안전을 연구하는 Apollo Research 계정입니다.
팔로워 12.8K·게시물 403·London