2026년 11월 4일

홈 타임라인 게시물 91개, 인기순

1위DHH@dhhX에서 보기

Omarchy 4.5는 M1부터 M3, Neo까지 완전히 호환되며 듀얼 부팅 설치로 Apple 하드웨어를 쉽게 쓸 수 있게 된다고 발표한다.

좋아요 6.9K리포스트 343답글 534노출 242.6K북마크 882

DHH@dhh
Omarchy 4.5 will ship with full compatibility for M1, M2, M3, and Neo! Incredible achievement from the Omarchy M team. The installer makes this so smooth, the dual boot makes it a no brainer. Apple hardware liberated. M4-6 on the way. WE CAN FIX EVERYTHING!

2위Paul Graham@paulgX에서 보기

paulg는 마이크 존슨 하원의장이 하원 운영을 막을 수 없게 되었으니 이제 엡스타인 파일의 내용이 공개되어야 한다고 주장한다.

좋아요 14.3K리포스트 1.2K답글 1.4K노출 567.2K북마크 461

Paul Graham@paulg
Now we'll finally learn what's in the Epstein files. Mike Johnson is no longer in a position to shut down the House in order to prevent votes on measures to force their release. It's time. It's past time.

3위Anvisha@anvishaX에서 보기

Voyager는 Higgsfield보다 절반 가격에 AI 영상 모델을 제공하고 API 호출을 줄여 생성 비용을 크게 낮췄다고 홍보한다.

좋아요 823리포스트 74답글 128노출 324.8K북마크 818

Anvisha@anvisha
We just made a 5x cheaper Higgsfield. Voyager gives you AI video models like Seedance 2.5 at half Higgsfield’s price. More importantly, it cuts wasted generations. Users get final outputs in 2–3x fewer API calls. Half the price per generation. A fraction of the generations.

4위Ricursive Intelligence@RicursiveAIX에서 보기

RicursiveAI는 처음부터 사전학습한 650억 파라미터 밀집 LLM인 Ripple-65B Preview를 공개했으며, 에이전트 코딩 과제에서 미국 오픈 웨이트 모델 중 최고 성능을 보였다고 주장한다.

좋아요 2.6K리포스트 321답글 264노출 310.3K북마크 1.4K

5위Node.js@nodejsX에서 보기

Node.js는 2026년 11월 11일 이후 Node.js 22, 24, 26의 보안 취약점 수정 릴리스를 낼 예정이며, 가장 높은 심각도는 HIGH라고 사전 공지했다.

좋아요 4.5K리포스트 736답글 138노출 657.2K북마크 1.7K

6위Google@GoogleX에서 보기

Gemini 3.8 Flash TTS의 음성 리믹싱 기능이 Google AI Studio와 Gemini API에 출시되었으며, 2,000개 이상의 음성에 억양, 속도, 음높이 등을 프롬프트로 조정할 수 있다고 소개한다.

좋아요 620리포스트 65답글 104노출 130.6K북마크 186

Google@Google
Voice Remixing for Gemini 3.8 Flash TTS is now live in @GoogleAIStudio and the Gemini API. Start with a preset from our library of 2,000+ voices or a custom-designed voice, and use prompts to fine-tune timbre, pitch, pace, and accent, like "add a subtle Southern U.S. accent" or "make it sound like a fire-breathing dragon." 🐉

7위OpenAI@OpenAIX에서 보기

OpenAI 보안 연구팀은 자율 취약점 발견·공격·패치 파이프라인을 만들어 한 번의 실행으로 리눅스 커널의 로컬 권한 상승 취약점 948개를 검증하고 패치를 생성했으며, 총 API 비용은 120만 달러였다고 밝혔다.

좋아요 2.1K리포스트 216답글 177노출 165.3K북마크 535

OpenAI@OpenAI
Our security researchers built an autonomous cyber vulnerability discovery, exploitation, and patching pipeline. In a single run with an upcoming cyber model, it validated and generated patches for 948 local privilege escalation vulnerabilities in the Linux kernel at a total API cost of $1.2 million.

8위Anthropic@AnthropicAIX에서 보기

Anthropic의 생물학 연구팀은 Claude와 함께 박테리아 유래 vault 입자의 첫 3D 구조를 밝혔고, 수개월 걸릴 작업을 며칠로 단축했지만 AI가 낸 오류도 있었다고 설명한다.

좋아요 2.6K리포스트 271답글 266노출 218.4K북마크 615

Anthropic@AnthropicAI
We’re sharing how our biology research team worked with Claude to map the first 3D structure of a vault particle from bacterial life. Vault particles are a longstanding mystery. They are barrel-shaped objects, among the largest protein structures found in our cells. They were first discovered by Leonard Rome and Nancy Kedersha in 1986, but their function remains unknown. With Claude, we’ve found results that might bring us closer to solving that mystery. Until now, vaults were only known in the cells of complex things like animals and plants. We have shown that similar proteins from bacteria (and the viruses that infect them) can also assemble into vaults. Finding the structures in these simpler forms of life may make it easier to investigate their function in our own cells. For our life sciences team, the project was also a demonstration of the potential (and the limitations) of using AI to accelerate science. In our full piece, we discuss how Claude was able to dramatically speed up some stages—manually determining the vault particle’s structure would have taken weeks to months of painstaking work, but here it took days. We also discuss how Claude made some simple mistakes along the way, and how we collaborated with our machine learning researchers to address them. Read more:
anthropic.comInvestigating vault particles with Claude

10위OpenAI@OpenAIX에서 보기

OpenAI는 ChatGPT 이미지를 2.6 버전으로 업그레이드해 GIF와 참조 이미지를 지원하고, API에 이미지 생성이 50% 빠른 두 가지 새 모델을 추가한다고 발표했다.

좋아요 4.4K리포스트 403답글 254노출 201.6K북마크 819

11위Patrick Collison@patrickcX에서 보기

패트릭 콜리슨은 지난 10년간 미국 민주주의가 유권자의 선호에 잘 반응해 왔다고 평가하며, 토크빌의 미국의 민주주의가 지금도 현재와 잘 맞아떨어진다고 말한다.

좋아요 2.8K리포스트 152답글 109노출 208K북마크 510

Patrick Collison@patrickc
It seems to me that democracy has on the whole worked very well in America over the last decade, and perhaps better than in any other major country in the West. National outcomes have repeatedly been responsive to underlying voter preferences on meaningful questions, and significant changes in policy have ensued. I’ve been reading Democracy in America this past week, and it’s striking how much it still rhymes nearly two centuries later.

12위Greg Brockman@gdbX에서 보기

OpenAI가 Pierre Computer 회사를 환영하며, 이 회사의 오픈소스 개발 도구가 이미 Codex 일부에 쓰이고 있어 함께 AI 기반 소프트웨어 협업의 미래를 만들겠다고 밝힌다.

좋아요 800리포스트 27답글 96노출 56K북마크 129

13위Uncle Bob Martin@unclebobmartinX에서 보기

코드를 이해하면 언어를 몰라도 전체 구조와 이름으로 대부분의 코드를 점검할 수 있으며, 선장은 배의 작동 원리를 알면 되고 볼트 하나하나의 세부 사항까지 알 필요는 없다고 비유한다.

좋아요 1.7K리포스트 90답글 84노출 74.5K북마크 240

Uncle Bob Martin@unclebobmartin
If you understand code, you can spot-check most languages — even if you don’t know them. Mid and high level structure are pretty obvious in virtually all languages. Names are self-evident. The captain knows how the ship works at a high level. He does not know the torque rating of every bolt.

14위Kimi.ai@Kimi_MoonshotX에서 보기

Moonshot AI가 Kimi K3.1 Pro의 모델 가중치를 공개하며 오픈 인텔리전스를 계속 발전시키겠다고 밝혔다.

좋아요 5.9K리포스트 627답글 277노출 361.1K북마크 748

15위Charlie Holtz@charlieholtzX에서 보기

Conductor 팀이 Meta에 합류해 MSL의 Nat Friedman, Alexandr Wang과 함께 다음 세대 인터페이스를 만들 계획이라고 발표하며 초기 지지자와 사용자에게 감사를 전한다.

좋아요 3.4K리포스트 114답글 610노출 347.8K북마크 414

Charlie Holtz@charlieholtz
I am very excited to share that Conductor is joining Meta! We have some big, crazy plans for the next interface. If you liked Conductor, you’re gonna love what’s coming. We’re super excited to work with @natfriedman and @alexandr_wang at MSL to build it. They’ve been on an absolute tear, and I’m convinced MSL has the highest talent density of all the labs. A huge thank you to our earliest believers @aaron_epstein, @ilyasu, and @nabeel. And to everyone here who tried Conductor, sent us feedback, and prompt requests - you made this so much fun. Thank you ❤️🚂 Choo choo!

16위Mitchell Hashimoto@mitchellhX에서 보기

Mitchell Hashimoto는 OpenAI가 Pierre를 인수한 것을 축하하며, 그곳에 친구들이 많이 있고 자신도 소액 투자자라고 말한다.

좋아요 1.8K리포스트 10답글 42노출 74.7K북마크 141

Mitchell Hashimoto@mitchellh
Congrats to Pierre! I have many great friends that work there and I’m so happy for them. I’m also a minor investor so I’m happy for myself. LOL.
OpenAI Newsroom@OpenAINewsroom
Today we’re announcing that OpenAI will acquire @pierrecomputer to accelerate our vision for making Codex a place where teams can build software together with AI, from first idea to the version they ship. Pierre’s open-source developer tools already power parts of Codex today. Together, we’ll build on that foundation for the agentic era. openai.com/index/openai-t…

18위Steve Rattner@SteveRattnerX에서 보기

Steve Rattner는 일론 머스크가 중간선거에 거의 5억 달러를 썼지만 특정 후보에게 쓴 돈의 98%가 낙선한 후보에게 돌아갔다고 지적했다.

좋아요 2.5K리포스트 726답글 106노출 85.2K북마크 138

19위NASA@NASAX에서 보기

NASA는 11월 5일 미국 동부 시간 오전 8시 30분에 우주비행사들이 우주정거장에 두 번째 새 태양전지판을 설치하는 우주유영을 중계한다고 안내한다.

좋아요 1.2K리포스트 142답글 75노출 268.5K북마크 57

21위DHH@dhhX에서 보기

Omarchy가 많은 사람의 호응을 얻고 있으며, 미래를 적극적으로 받아들이고 반대 의견을 무시한 채 계속 고쳐 나갈 때 좋은 일이 일어난다고 말한다.

좋아요 846리포스트 27답글 44노출 30.3K북마크 26

22위Linum@LinumIncX에서 보기

두 명으로 이루어진 연구소가 Pyramid-JiT 구조로 약 2만 달러를 들여 학습한 20억 파라미터 텍스트-이미지 모델을 Apache 2.0 라이선스로 공개했다고 알린다.

좋아요 89리포스트 5답글 4노출 16.9K북마크 94

Linum@LinumInc
We’re a 2-person lab researching more efficient architectures for generative image and video. linum.ai/field-notes/pj… Last month, we published Pyramid-JiT, a pixel-space architecture that converged 11x faster than its latent-diffusion counterpart. Today, we're releasing a 2B Pyramid-JiT text-to-image model as a milestone towards more affordable video models. Trained from scratch for only ~$20K. See blog for details & weights (Apache 2.0).

23위shawn@shawn_panaX에서 보기

Phone Harness는 에이전트가 개인 휴대폰이나 클라우드의 iPhone/Android를 사용해 실제 앱을 테스트하고, 사람에게 메시지와 전화를 보내며, 2단계 인증 코드를 가져올 수 있게 해 주는 오픈소스 도구이다.

좋아요 591리포스트 40답글 67노출 62.6K북마크 619

shawn@shawn_pana
Agents can now control phones. Introducing Phone Harness (YC F26) Let your agent use your personal phone, or its own iPhone/Android in the cloud > Test and use real mobile apps > Message and call people > Grab verification codes for 2FA Open source. Connect through MCP, CLI, or API. Get started at phone-harness.com

24위a16z@a16zX에서 보기

a16z는 2007~2016년 빈티지 상위 5% 미국 펀드 기준으로 벤처 펀드의 TVPI가 8.5배로 바이아웃 펀드의 3.8배보다 높다는 분석을 공유한다.

좋아요 43리포스트 5답글 13노출 15.6K북마크 32

a16z@a16z
The higher up you go, the more venture pulls ahead of buyout Among the top 5% of US funds from 2007-16 vintages: - Venture funds hit 8.5x TVPI, with 5.3x returned in cash - Buyout funds hit 3.8x TVPI, with 3.5x returned in cash "In venture, winning tends to beget winning: the best firms typically get the next call from the best founders." Full piece from a16z's Jen Kha: a16z.news/p/is-private-e… Not an offer or solicitation. None of the above should be taken as investment advice or an advertisement for investment services. Please see a16z.com/disclosures for more information.
Jen Kha@jkhamehl
x.com/i/article/2117…

25위a16z@a16zX에서 보기

a16z의 Jen Kha는 기술 벤처가 기술 바이아웃보다 수익성이 높았다는 점과 바이아웃의 부담 증가를 근거로, 새로운 기술 물결이 시작되고 있다고 주장한다.

좋아요 109리포스트 11답글 14노출 18.7K북마크 79

a16z@a16z
Top-quartile venture beat top-quartile buyout through the internet buildout and again through mobile and cloud, and a16z's Jen Kha argues we're at the start of the next wave. Tech buyouts have long been the "safe" bet in private markets. Now the stress is showing: - In the first half of 2026, venture funds marked their tech holdings up 16%, while buyout funds marked theirs down 4% - 70% of the debt foreclosed in 2025 and 2026 came from deals done in 2021 or 2022 - A typical buyout now needs 10 to 12% EBITDA growth a year to make 2.5x in five years, up from about 5% in 2015 - Top 5% venture funds from 2007-16 returned 5.3x their capital in cash, versus 3.5x for top 5% buyout funds "Tech buyout is a late-cycle asset class. It's most effective once a technology wave has settled and when business models are predictable enough to borrow against. When a new technology wave drops, all bets are off. By contrast, that's when venture shines." Full piece from a16z's Jen Kha: a16z.news/p/is-private-e… Not an offer or solicitation. None of the above should be taken as investment advice or an advertisement for investment services. Please see a16z.com/disclosures for more information.
Jen Kha@jkhamehl
x.com/i/article/2117…

26위a16z@a16zX에서 보기

a16z가 Parsec 창업팀이 만든 원격 데스크톱 솔루션 Phaze에 투자했으며, 이 제품은 사람과 AI 에이전트 모두에게 원격 워크스테이션에서 로컬 같은 경험을 제공하도록 설계되었다.

좋아요 53리포스트 5답글 10노출 15K북마크 38

a16z@a16z
We’re thrilled to invest in Phaze. @Phaze_in is a high-performance remote desktop solution for giving humans and agents a local-like experience on remote workstations. It’s the best solution for interactive, latency-sensitive workloads, and it was designed from the ground up for working with agents, which makes compatibility, management and administration seamless. It’s not a surprise Phaze is as good as it is given the team behind it. @boxerbk and Chris Dickson, the founders of Phaze, previously built Parsec, which at the time was the leading remote desktop solution for graphics-intensive applications. We’re very excited to be working with this team again to bring desktops and applications to agents. And of course to humans too. By @martin_casado
Benjy Boxer@boxerbk
Today we’re announcing an $8.75M seed round for Phaze. Companies aren’t ready for how many computers people and AI agents are about to need. And it’s happening just as computer prices are up 3x. Chris and I spent years building Parsec to make a computer thousands of miles away feel like it’s sitting right in front of you. We sold that company, and we’re starting over because people aren’t the only users of computers anymore. Now, agents need access to real computers to run applications, test software, and get work done. They need to work alongside people, hand off control when necessary, and operate within the security policies companies already have in place. And companies need to start managing computers like infrastructure. That’s why we’re building @Phaze_in: secure, high-performance access to computers, wherever those machines happen to be. An engineer can connect to a CAD workstation from any device and forget about being physically stuck at that workstation. A video editor can ask an AI agent to connect to a render machine to monitor and fix broken renders. An IT support person can fix multiple machines at the same time with an AI agent connecting to employees’ computers. If the agent needs help, it can hand control back to the technician in the same session. We loved working with the firms who backed Parsec, so we’re running it back. Our Seed round is led by @martin_casado at @a16z, with @bryce at @indievc, @davidbeisel at @NextViewVC, @BenjLerer at @LererHippeau, and @Ryann_Lai at @makersfundvc.

27위yuj@yyjhaoX에서 보기

Floot은 AI에 요청만 하면 구독과 일회성 결제, 배송 라벨과 추적 기능을 API 키나 웹훅 설정 없이 앱에 추가할 수 있게 해준다.

좋아요 78리포스트 5답글 43노출 18.7K북마크 50

yuj@yyjhao
Build your idea, and get paid. Introducing Floot payments + shipping. Your Floot app can now take payments and ship orders. Ask your AI for “a $9/month plan”, “a checkout for this course” or “shipping labels for my orders”, and it’s done. No need to wrangle API keys or webhooks. Set up subscriptions and one-time payments, paid out straight to your bank. Selling something physical? Get live rates, printable labels and tracking right in your app. Start making money today at Floot P.S. We are also live today at product hunt with a 30% off coupon. Please support us if you like what we do: producthunt.com/products/floot

28위a16z@a16zX에서 보기

Eve는 AI로 의료 기록 검토 등 시간이 많이 드는 소송 업무를 줄여 소형 법률 사무소의 사건 처리 능력을 키워주며, 350,000건 이상의 사건에서 35억 달러 이상을 회수했다고 밝히고 채용 중이라고 알린다.

좋아요 61리포스트 4답글 13노출 16.4K북마크 38

a16z@a16z
Plaintiff lawyers only get paid if they win, so every legal firm hits a limit on how many cases it can take, and every case it turns down leaves someone without a lawyer. Eve builds AI that handles the time-intensive casework, like reviewing medical records, drafting key documents, and opening insurance claims, cutting work that took days to under an hour. One firm went from about 100 clients per attorney to 500. That gives small firms a real shot against insurers with far more resources. Co-founder and CEO Jay Madheswaran: "Think about giving David a slingshot to take down Goliath. That's what we're doing." Three years in, 1,700+ of the roughly 29,000 plaintiff firms in the US use Eve. It's been used in 350,000+ cases and helped recover $3.5 billion for clients. Eve is hiring across engineering, machine learning, sales, and customer success in SF, the Bay Area, and remotely across the US: a16zjobs.substack.com/p/with-164m-in… @jayanthmadhesw1 @evelegalai
Jayanth Madheswaran@jayanthmadhesw1
Huge milestone: Eve has helped 350,000 people take on a system built to wear them down. Take Mike Dickerson. A four-car pileup, two hip surgeries, then an insurer betting he’d take a lowball offer. He didn’t. His lawyers @GeigerGroup fought for years and, with Eve’s help, got him 15x the first offer. Firms on Eve have recovered $3.5B+ for their clients, and we’re just getting started. Next milestone: 1 million people helped.

29위a16z@a16zX에서 보기

a16z는 AI 어시스턴트 벤치마크를 만든 David Pawlan과 함께, 사용자들이 가장 두려워하는 신뢰 문제와 에이전트 비교 테스트, 에이전트의 미래에 대해 이야기하는 대담을 소개한다.

좋아요 50리포스트 7답글 35노출 21.3K북마크 28

a16z@a16z
David Pawlan, creator of Assistant Benchmark, has tracked 147 AI assistants so far, and in his group chats of AI power users, the number one fear is trust. It comes up in 22% of roughly 3,600 messages. He puts agents through head-to-head tests, like ordering same-day flowers as cheaply as possible or planning a weekend in LA that's both luxurious and cost-efficient, to see which ones can hold competing preferences. He joins a16z's Anish Acharya and Josh Elman to discuss why the scariest agent failures are the embarrassing ones, why agents are where e-commerce was in 1997, why scale is still a moat, and the "dark agent" that could speak to the parts of ourselves we don't reveal. 0:52 Inside a meetup for 100 agent nerds 2:52 Why we'll all have an army of agents 5:36 Trust is power users' top fear 6:28 Building fences around your agent 10:14 Privacy loses to free pizza 12:41 Can agents balance price and speed? 14:19 How agents define luxury differently 19:44 Migrating from one agent to another 21:06 What your photo roll tells an agent 25:48 Will agents make us more rational? 27:38 Why paternalism kills products 31:43 Gossipy agents as social glue 33:41 Scale is still a software moat 36:55 From OpenClaw to agents hanging out 39:23 Agents that introduce you to people 46:15 Predictions and the dark agent idea YouTube: youtu.be/dSiTwPl0t1U @DavidPawlan @illscience @joshelman

30위Y Combinator@ycombinatorX에서 보기

YC의 Startup School 에피소드에서 Francois가 연구자가 훌륭한 창업가가 되기 위해 필요한 10가지 사고방식의 전환을 설명한다.

좋아요 198리포스트 31답글 40노출 22.1K북마크 97

Y Combinator@ycombinator
Google, Anthropic, and Databricks were all started by founders with PhDs. But they’re exceptions to the rule. Many of the instincts that help researchers succeed in academia can often work against them when building a startup. YC Visiting Partner @FrancoisChauba1 knows both worlds. He’s currently a PhD student, previously did research in Fei-Fei Li’s lab, and founded and ran Focal Systems (W16). In this episode of Startup School, Francois breaks down 10 mindset shifts researchers need to make to become formidable founders, from learning to thrive without grades or deadlines to distinguishing interesting problems from important ones, moving faster with imperfect information, and starting with what customers actually want. 00:53 - Vanity metrics vs. what people want 01:45 - Interesting vs. important 02:27 - Novelty vs. plagiarism 03:54 - There is no finish line 04:32 - Get on a plane and talk to people 05:53 - Hyperbole is okay 06:28 - The 70% rule 07:11 - You can't delegate the hard stuff 07:52 - Become a capitalist 08:16 - Nobody cares how smart you are 09:04 - Get out of your office

31위antirez@antirezX에서 보기

antirez는 LLM의 은닉 상태는 다음 토큰에 필요한 정보만 담고 있어 전체 표현이 아니며 인코더도 아니라고 주장하면서, JEPA와 LLM을 비교할 때 이 점을 기억하라고 말했다.

좋아요 417리포스트 26답글 34노출 32.6K북마크 180

antirez@antirez
Remember one thing when you read JEPA vs LLMs in the context of the platonic representation hypothesis. Hidden state of an LLM does not represent the full representation of the LLM, it carries only what is needed for the next token. It's not an encoder.

32위Llama@Llama_appX에서 보기

Llama 앱이 Hugging Face Hub의 수많은 LLM을 몇 번의 클릭으로 내 컴퓨터에서 빠르게 실행할 수 있게 해주며, 비공개이고 로컬에서 동작하며 오픈소스라고 소개한다.

좋아요 255리포스트 30답글 16노출 18.9K북마크 108

33위Google for Developers@googledevsX에서 보기

Google은 Gemini 3.8 Flash TTS에 음성 리믹스 기능을 출시해, 사용자가 프리셋이나 디자인된 음성의 음색, 음높이, 속도, 억양을 프롬프트로 세밀하게 조정할 수 있게 했다.

좋아요 349리포스트 38답글 10노출 26.2K북마크 147

Google for Developers@googledevs
Voice Remixing is now live in Gemini 3.8 Flash TTS. Fine-tune voices without starting from scratch. Pick from over 2K preset library voices or custom designed voices in @GoogleAIStudio and the Gemini API to dial in timbre, pitch, pace, and accent using simple prompts. How it works: ✅ Start from any preset or designed voice ✅ Prompt adjustments for pitch, pace, accent, or timbre (like "soften the delivery") ✅ Audition three generated takes and save your favorite Try "Remix this voice" in @GoogleAIStudio today.

35위Verne@VerneRoboticsX에서 보기

Verne가 산업용 로봇을 위한 오픈 웨이트 세계 행동 모델 V0.5를 공개했으며, LIBERO와 RoboTwin 2.0 벤치마크에서 미국산 오픈 웨이트 모델 중 선두라고 밝힌다.

좋아요 251리포스트 65답글 30노출 17K북마크 100

Verne@VerneRobotics
Introducing V0.5, Verne’s open-weight world action model for industrial robots. We believe democratizing robotics starts with making frontier foundation models open to everyone. Bringing robots to every industry will take many teams, building on shared progress across many workflows. V0.5 is the leading US-made open-weight model on LIBERO and RoboTwin 2.0 benchmarks. We hope you’ll join us in putting V0.5 to work.

36위a16z@a16zX에서 보기

a16z는 PE 투자 소프트웨어의 성장률이 공개 소프트웨어보다 낮고 벤처 투자 기업은 훨씬 빠르게 성장하며, 최고 기업들이 비공개로 더 오래 머물면서 성장이 상장 전에 일어나고 있다고 분석한다.

좋아요 62리포스트 4답글 11노출 8.7K북마크 22

a16z@a16z
PE-backed software now grows slower than public software, while venture-backed companies grow several times faster Median revenue growth (latest available): - PE-backed software grows 3 to 7% - Public software grows 10 to 16% - Venture-backed private software grows 18 to 30% "Because the best companies are staying private longer than ever, more of that growth is happening before they ever hit the public markets." Full piece from a16z's Jen Kha: a16z.news/p/is-private-e… Not an offer or solicitation. None of the above should be taken as investment advice or an advertisement for investment services. Please see a16z.com/disclosures for more information.
Jen Kha@jkhamehl
x.com/i/article/2117…

37위Kaustubh Bansal@bansal_kaustubhX에서 보기

YC F26 기업 Bayworks는 기술자들의 차량 수리 음성과 영상 데이터를 수집해 딜러십 운영과 수리 표준화를 돕는 시스템을 만들고 있으며, 110일 만에 80곳 이상의 딜러십으로 확장했다.

좋아요 160리포스트 16답글 30노출 37.3K북마크 49

Kaustubh Bansal@bansal_kaustubh
Self-driving companies record every mile a car drives. Nobody records how it gets fixed. Bayworks (YC F26) does that, building the intelligence that keeps America’s vehicles running. In 110 days, Bayworks has gone from 0 to 80+ dealerships covering 4,000 repairs a day, including a top-50 US dealer group ($3B+ annual revenue). We capture voice and video data as technicians repair cars and verify the work, turning this ground truth into a system that runs the dealership’s operations and standardizes every repair.

38위Isa@isausmanovX에서 보기

에이전트 중심으로 설계된 스마트폰 Raven을 소개하며, 해커톤 우승을 계기로 얻은 관심과 함께 사전 주문이 시작되었음을 알린다.

좋아요 156리포스트 21답글 49노출 10K북마크 46

Isa@isausmanov
Winning OpenAI hackathon got me into YCombinator. 1,000+ likeminded people have reached out since. Introducing Raven, a new phone rebuilt around agents. You say what you want, your agents do it, and the result comes back to your phone. More done, less screen time. Pre-orders open today.

39위Andrew Ng@AndrewYNgX에서 보기

Andrew Ng는 11월 30일부터 12월 1일까지 뉴욕에서 열리는 AI Dev 행사에 참석을 권하며, AI 엔지니어 직무의 빠른 성장과 에이전트, 메모리, 평가 등 실습 세션을 소개한다.

좋아요 96리포스트 5답글 42노출 20K북마크 27

Andrew Ng@AndrewYNg
Join me and 2000 other AI engineers and builders at AI Dev in New York City, Nov 30-Dec 1! A LinkedIn study found that over July-September, the job title AI Engineer has been one of the 10 fastest-growing occupations in seven major economies. AI engineering has become a core skill for all software work, and AI Dev is designed to help you develop it. Our sessions give you hands-on practice with: - Coding agents and harness design: Harness tuning, repository context, autonomous coding - Agent memory and knowledge graphs: GraphRAG, vector search, persistent memory, hybrid retrieval - Agents in production: Evals, observability, simulated testing, browser and voice agents @ylecun and I will also talk about what’s next in world models and physical AI. I hope to see you there! Tickets:
ai-dev.deeplearning.aiAI Dev NYC 2026 — DeepLearning.AI

40위Slash@slashappX에서 보기

slashapp이 YC 지원서를 제출한 뒤 6년 만에 연 매출 5억 4천만 달러 규모에 이르렀다고 자축하는 글이다.

좋아요 112리포스트 3답글 11노출 10K북마크 31

41위Capy@capydotaiX에서 보기

Capy 등은 11월 15일 YC 사무실에서 소비자 모바일 앱 해커톤을 열고, 우승팀에게 YC 면접 기회와 상금을 주며 11월 13일까지 지원을 받는다고 알린다.

좋아요 128리포스트 17답글 15노출 16.1K북마크 35

42위Hugging Face@huggingfaceX에서 보기

Hugging Face는 100대 이상의 실제 로봇과 140곳 이상의 실제 장면으로 로봇 모델을 평가하는 오픈 이니셔티브 RoboColosseum의 창립 파트너이며, 모든 실행 기록을 공개한다고 밝혔다.

좋아요 181리포스트 36답글 66노출 18.5K북마크 44

Hugging Face@huggingface
Evaluating robot models in the real world is hard, but solvable. That's why we're a founding partner of RoboColosseum 🤗 100+ real robots, 140+ in-the-wild scenes, and every rollout, failures included, open on the Hub.
RoboColosseum@RoboColosseum
Introducing RoboColosseum An open robot eval initiative by @huggingface @BitRobotNetwork @eaic_sg @DJiafei Access: - 100+ robots (@UnitreeRobotics G1s, @LeRobotHF SO-101s, YAMs, Franka and more) - 140+ in-the-wild scenes - Open rollout data - 10,000 FREE eval hrs Thread ↓

44위Rishi@rishimehta30X에서 보기

Ascenix는 요양원 운영을 위한 맞춤형 소프트웨어 플랫폼으로, 보고서 작업과 환급 누락, 규정 준수 문제를 하나의 시스템으로 해결하며 데모를 안내한다.

좋아요 71리포스트 21답글 30노출 7.5K북마크 16

Rishi@rishimehta30
We’re launching Ascenix (@ycombinator F26): custom software for running senior homes. Our first customer was the retirement home we lived in for months. Now we’re growing 15% WoW across Canada and the US. Senior homes face bottlenecks like time-consuming reporting, missed reimbursement revenue, and compliance gaps that risk fines and reputational damage. Solving each with separate tools can mean dozens of vendors, hundreds of thousands of dollars in spending, multiple logins, and siloed data. We replace that stack with one platform tailored to each home. Teams describe what they need, and Ascenix turns their data and instructions into dashboards, reports, or automated workflows. Our reusable building blocks let us add new workflows as new bottlenecks arise, without another vendor or product. Visit ascenix.co for a demo.

45위Robert Xu@robbyuchenX에서 보기

Fern은 텔레옵 에피소드 3개만으로 물리와 시각이 정확한 로봇 시뮬레이션을 만들고, API로 수많은 롤아웃을 병렬 실행하게 해주는 플랫폼이다.

좋아요 78리포스트 14답글 21노출 5.5K북마크 16

Robert Xu@robbyuchen
Today we're excited to launch Fern by @ishikilabs, backed by @ycombinator: high-fidelity simulation platform for robotics! Robotics teams are bottlenecked by hardware evals. Running evals in the real world is slow and expensive, and building a simulation that's a) high fidelity, b) fast, and c) scales to millions of rollouts is difficult. Using only 3 teleop episodes, Fern creates a high fidelity simulation that's grounded in both physics and visual fidelity. Then the platform allows you to run thousands of rollouts in parallel through our API, with no infrastructure, sims, or solvers to manage. One customer has run 900k+ sims on Fern. Hear how their experience has been 👇 If running high fidelity, scalable evals is a need, schedule a meeting with us at fern.bot !

46위Y Combinator@ycombinatorX에서 보기

Weave는 AI 코딩 도구가 엔지니어링 팀의 생산성을 실제로 높이는지 측정하며, 최근 1,350만 달러 시리즈 A 투자를 유치하고 Weave Router를 출시했다는 YC 패널 대담을 소개한다.

좋아요 86리포스트 15답글 27노출 15.7K북마크 16

Y Combinator@ycombinator
Weave (W25) helps engineering teams understand whether their AI coding tools are actually making them more productive. Its models measure the work humans and agents produce and compare that output with AI spending. The company recently raised a $13.5M Series A led by Standard Capital and launched Weave Router, which selects models for different coding tasks. In this fireside, @WorkWeave co-founders @adambcohen93 and @AndrewChurchiII join YC's @snowmaker to share how the company evolved alongside the rapid rise of AI coding, from measuring the impact of tools like Cursor to helping teams build software factories that can work autonomously. They explain how the best engineering teams are learning to measure the real output of their agents, spread the habits of their strongest AI users, and optimize for a future where much more software is written without a human at the keyboard. 00:00 — Optimizing AI Coding With Weave 04:02 — Getting More From Your AI Budget 08:39 — Measuring Work Beyond Lines of Code 12:30 — From AI Adoption to Optimization 16:47 — Measuring AI ROI and Coaching Engineers 20:16 — Building Software Factories 25:38 — The Future of Model Routing

47위Anthropic@AnthropicAIX에서 보기

Anthropic이 vault 입자 구조 연구의 사전 인쇄본 전문 링크를 공유한다.

좋아요 271리포스트 25답글 12노출 44.1K북마크 50

48위Juan Pablo Linares@jplinares99X에서 보기

Blinka가 글로벌 기업을 위한 USD 뱅킹 플랫폼을 공개했으며, 출시 두 달 만에 고객 400곳과 1,800만 달러 이상의 결제 실적을 올렸다고 알린다.

좋아요 96리포스트 16답글 24노출 8.1K북마크 17

Juan Pablo Linares@jplinares99
Today we are announcing @GetBlinka to the world. A USD banking platform for global businesses. You can open a USD business account from anywhere and receive investor and customer payments into FDIC-insured accounts*, send money internationally without the FX markup, earn yield on idle USD, and issue corporate cards with 0 international fees and 1.5% cash back. We are backed by @ycombinator and in just two months reached 400 customers and done over $18M in payments.

49위Greg Brockman@gdbX에서 보기

gdb는 자사 취약점 파이프라인으로 리눅스 로컬 권한 상승 취약점 약 1,000개를 발견·검증하고 패치를 준비했으며, 취약점당 비용은 약 1,000달러라고 밝혔다.

좋아요 409리포스트 28답글 39노출 23.4K북마크 65

Greg Brockman@gdb
we've found, validated, and prepared patches for almost 1,000 linux local privilege escalations using our cyber vulnerability pipeline, at the cost about $1,000 per vulnerability. an emerging scalable approach to significantly hardening both open source and proprietary projects
OpenAI@OpenAI
Our security researchers built an autonomous cyber vulnerability discovery, exploitation, and patching pipeline. In a single run with an upcoming cyber model, it validated and generated patches for 948 local privilege escalation vulnerabilities in the Linux kernel at a total API cost of $1.2 million.

50위Visual Studio Code@codeX에서 보기

VS Code 1.145 업데이트로 별도 기기의 에이전트 세션 개선, cron 기반 자동화 일정 설정, 채팅 알림 줄이기, Blobby 기능 등 이번 주 새 기능을 알린다.

좋아요 125리포스트 15답글 7노출 11.2K북마크 19

51위OpenAI@OpenAIX에서 보기

OpenAI는 Images 2.6을 ChatGPT와 Codex에서 모든 요금제에 제공하기 시작했고, GPT-Image-2.6 Flare와 Sunburst는 API에서도 이용할 수 있다고 알렸다.

좋아요 284리포스트 17답글 6노출 32.4K북마크 43

53위OpenAI@OpenAIX에서 보기

OpenAI는 개발자용 신규 이미지 모델 GPT-Image-2.6 Flare와 Sunburst를 공개했으며, Flare는 빠른 생성을, Sunburst는 고품질과 편집 일관성을 강점으로 내세운다고 밝혔다.

좋아요 275리포스트 25답글 4노출 32.5K북마크 35

OpenAI@OpenAI

@OpenAI 님에게 보내는 답글

Two new image models for developers: GPT-Image-2.6 Flare and Sunburst. Flare generates images in as little as six seconds. Sunburst brings higher-quality generation and greater consistency across edits. Image markup and SVG exports give you more ways to refine and use what you create.

54위Yann LeCun@ylecunX에서 보기

얀 르쿤은 1996년부터 2002년까지 Vapnik의 학과장을 맡아 연구 여건을 지원했으며, ML 접근법에 대해 견해가 달랐지만 그의 연구를 적극 알렸다고 밝혔다.

좋아요 844리포스트 24답글 14노출 53.8K북마크 96

Yann LeCun@ylecun

@Arjunjain 님에게 보내는 답글

Plot twist: I was Vapnik's Department Head between 1996 and 2002, and made sure he had the space, resources, and collaborators to do his best work. To people who think I'm dogmatic: I famously disagreed with Vapnik about the best way forward in ML, but did what I could to promote his work, because research is about exploration and science is about results.

55위RunAnywhere@RunAnywhereAIX에서 보기

RunAnywhere는 Trainium에서 최신 오픈 모델을 지원하기 위해 TrnRT라는 추론 엔진을 만들어 오픈소스로 공개했고, 한 대의 Trainium2 장비에서 GLM 모델이 분당 357만 토큰을 처리한다고 밝혔다.

좋아요 27리포스트 12답글 11노출 3.8K북마크 3

RunAnywhere@RunAnywhereAI

@AnthropicAI 님에게 보내는 답글

@AnthropicAI is betting up to 5 GW on @awscloud Trainium. We wanted the latest open models flying on it too but there is ZERO support for those. So we wrote megakernels and open-sourced TrnRT: Our Tranium inference engine Running @Zai_org's 320B GLM-5.3-Flash on one Trainium2 box, producing 3.57 M tokens per minute, measured on real coding agent traffic via @SemiAnalysis_'s AgentX.
Andy Jassy@ajassy
Trying a new AI chip used to be a lot of work. You often had to rewrite your code, revalidate it, and sink many weeks in, betting this new chip is the right one for your workload. That cost time and money, and wore on the team. Your best people got buried in plumbing instead of building the thing they came to build. TorchNeuron is going to largely take that off the table, enabling your existing PyTorch code to run natively on Trainium with one configuration change and no rewrite. With Trainium’s very appealing price-performance characteristics, TorchNeuron is going to make it even easier for more companies to use AI more cost-effectively. aboutamazon.com/news/innovatio…

56위Engineering@EngineeringX에서 보기

X 웹을 하나의 구현으로 통합해 재구축한 결과, 느린 연결에서도 프로필의 첫 게시물이 22초 이상에서 2.7초로 줄었고 국가별로 순차 출시 중이라고 소개한다.

좋아요 162리포스트 14답글 6노출 14.2K북마크 16

Engineering@Engineering
Three implementations became one, and a profile's first post now loads in 2.7 seconds on a slow connection. Read how we rebuilt X web for speed and efficiency:
Yagiz Nizipli@yagiznizipli
The web rewrite is working. On a slow connection, the old X web took more than 22 seconds to show the first post on a profile. The new X web does it in 2.7 seconds. Here's how we did it, and how we're rolling it out one country at a time. blog.x.com/en/engineering…

57위Aaron Epstein@aaron_epsteinX에서 보기

Conductor 팀이 Meta에 합류해 MSL의 나트 프리드먼, 알렉산더 왕과 함께 차세대 인터페이스를 만든다는 소식을 전하며 팀의 성장을 축하한다.

좋아요 115리포스트 6답글 7노출 6.4K북마크 9

Aaron Epstein@aaron_epstein
Conductor is joining Meta! It's been so cool to see their evolution from pivot hell during the S24 batch, to building a product that they themselves wanted to use (back in the stone age of 2025 when people still coded by hand!), to building a community and movement around their product. All of their success is so well deserved, couldn't happen to better people. Congrats @charlieholtz @jacksondecampos and the entire @conductor_build team!
Charlie Holtz@charlieholtz
I am very excited to share that Conductor is joining Meta! We have some big, crazy plans for the next interface. If you liked Conductor, you’re gonna love what’s coming. We’re super excited to work with @natfriedman and @alexandr_wang at MSL to build it. They’ve been on an absolute tear, and I’m convinced MSL has the highest talent density of all the labs. A huge thank you to our earliest believers @aaron_epstein, @ilyasu, and @nabeel. And to everyone here who tried Conductor, sent us feedback, and prompt requests - you made this so much fun. Thank you ❤️🚂 Choo choo!

58위Anaïs Howland, CFA@AnaisHowland18X에서 보기

Oasive는 지방채 분석용 AI 서비스를 출시해, 차입자별 신용 관련 뉴스와 이사회 회의 요약, 출처가 표시된 재무 데이터, 채권 구조를 한 페이지에 보여주며 MCP 서버와 API도 계획하고 있다.

좋아요 41리포스트 5답글 14노출 3.8K북마크 3

Anaïs Howland, CFA@AnaisHowland18
Today I'm launching @OasiveAI for municipal bonds, backed by @ycombinator! Figuring out if a city or hospital can pay back its bonds means digging through PDFs, board meeting videos and local news. Oasive for Munis puts it on one page per borrower: → Credit-relevant news with rating changes and distress flagged → Board meetings and hearings, summarized and linked to the recording → Financials and sector metrics, every number sourced → Each bond's security structure → Ask Oasive anything, get a sourced answer in seconds Next we're planning an MCP server and API for sub-sector data. What would you want in it? Early access: DM me or oasive.ai

59위NASA@NASAX에서 보기

NASA의 Anil Menon 우주비행사가 국제우주정거장에서 MarkRober와 대화하며 학생들의 우주 생활 질문에 답한 소식을 전한다.

좋아요 517리포스트 70답글 53노출 335.5K북마크 30

60위Jeff Dean@JeffDeanX에서 보기

Jeff Dean이 어제 투표율이 높았던 것을 기쁘게 여기며 투표에 참여한 모든 사람에게 감사를 전한다.

좋아요 662리포스트 12답글 8노출 24.8K북마크 10

Jeff Dean@JeffDean
Awesome to see high voter turnout yesterday. Thanks to everyone who took the time to vote!
Jeff Dean@JeffDean
Election Tuesday is just two days away. For those eligible to vote in the U.S., please make time to go to the polls to vote (or mail in your ballot if you haven't already). Voting is a great privilege of living in a democracy and how we all make our collective voices heard. 🗳️

61위Kushal Byatnal@kushalbyatnalX에서 보기

Extend UI가 월 npm 다운로드 약 200만 건에 이르렀다며, 문서 뷰어와 편집기, PDF 도구 등 다양한 오픈소스 기능을 소개한다.

좋아요 35리포스트 4답글 3노출 5.2K북마크 30

Kushal Byatnal@kushalbyatnal
Extend UI is now approaching 2M monthly npm downloads 📈 - fully open source (MIT licensed) - 1.6k GitHub stars - officially supported by our team many of the largest AI startups/enterprises already use it in production if you're building anything where users interact with documents, there's probably something in here you need: - DOCX editor & viewer: rich text formatting, comments, tracked changes, tables, images, import/export - PDF viewer: text selection, search, page thumbnails, zoom, rotation, and bounding box overlays - PDF editor: highlights, annotations, comment threads, redaction, form filling, and page reordering - XLSX viewer + editor: sheet tabs, formulas, cell formatting, frozen panes, merged cells, and export - PPT viewer: in-browser slide rendering, thumbnail navigation, and zoom - CSV/TSV viewer: virtualized grids for large datasets, search, and cell navigation - source citations + extraction review: connect extracted fields to bounding boxes in the original document, edit values, and compare JSON outputs - schema builder: visual editor for extraction schemas, nested objects, arrays, enums, and a synced JSON view - document splitting: drag pages between groups and reorder them with page previews - e-signatures: signature and initials placement, date fields, required-field checks, and signed PDF export - file browser: Finder-style folders with icon, list, column, and gallery views - file uploads: drag-and-drop, batch uploads, validation, progress, and retry states - file thumbnails: preview tiles with loading and fallback states - OCR + layout inspection: block, line, and word overlays with confidence coloring and navigation back to the source page
Extend@ExtendHQ
Here are a few cool public projects built with our open-source UI kit! 1) @kortix CompanyOS - AI agents that work across your company’s tools, tasks, and shared knowledge 2) OpenMeld by @felixlu1018 - a shared workspace where people and AI agents collaborate and pick up where others left off 3) Jevbox, built by our own @andrewlu0 - a document library that organizes your files and answers questions with page-level citations 4) Neutree Agent Platform - a self-hosted platform for building and sharing AI agents, with Extend UI powering document previews

62위Adam Cohen@adambcohen93X에서 보기

Weave Checks가 오픈소스로 공개되었으며, 에이전트가 모든 PR을 검토해 마크다운으로 정의한 코드 품질 기준을 자동으로 강제한다.

좋아요 39리포스트 5답글 3노출 5.2K북마크 28

Adam Cohen@adambcohen93
Today we're open sourcing Weave Checks, to help you enforce the code principles you care about (without a human reading every line). By running an agent against every PR, Weave Checks ensures it meets criteria you define in Markdown files. Each Markdown file is a single code quality standard you want to uphold. The agent checks that the standard is upheld and leaves a review comment if it's not. It self-resolves once its own comments have been addressed. Top Weave Checks seen we've setup: 1. No tautological tests 2. Over abstraction 3. Defensive programming 4. Dead-code paths 5. Intermediate variable cleanup Let us know what you think! weaveos.com/checks

63위Aaron Epstein@aaron_epsteinX에서 보기

Vyra는 원본 영상과 프롬프트만 올리면 숏폼 영상을 편집해 주는 개인 영상 편집 에이전트라고 소개하며 적극 추천하는 글이다.

좋아요 70리포스트 2답글 4노출 8.5K북마크 43

Aaron Epstein@aaron_epstein
The simplest way to build an audience is to just share what you're working on, but most founders don't do it because creating and editing video content is too painful. Vyra just makes it so easy now, it's like having a dedicated video editor on your team. Def check it out!
sulan@sulansart
Introducing Vyra! We're building the personal video editing agent for shortform content. Just upload your raw footage, prompt, and Vyra edits it for you. Try now at usevyra.com

64위Steve Endow@steveendowX에서 보기

스티브 엔도우는 컴퓨터의 탄생 역사와 첫 컴퓨터에 얽힌 이야기를 담은 책에 대해 저자 unclebobmartin에게 감사를 표하며, 그 책이 자신을 프로그래머의 길로 이끌었다고 말한다.

좋아요 22리포스트 2답글 2노출 3.3K북마크 12

65위a16z@a16zX에서 보기

a16z의 마틴 카사도는 투자자가 가진 재무적, 생산적, 프런티어, 투기적이라는 네 가지 마인드셋을 설명하며, 자신의 신념을 알아야 개인 편향을 정당화하지 않게 된다고 말한다.

좋아요 23리포스트 4답글 7노출 9K북마크 10

a16z@a16z
a16z's Martin Casado on the four mindsets people bring to investing, and why knowing yours stops you from justifying your own bias: 1. Financial "The correct proxy for the health of a business is the market it's going into and its finances. That's what it is. What are the SaaS metrics? What is the size of the market? That is a proxy for the entire business." 2. Productive "You believe you're investing in a productive asset, and the productive asset is what's creating value. If you believe in that, as it creates value, it will create market... The type of diligence you do if you're a productive investor is more like, what is the technical approach? What are the adjacent markets?" 3. Frontier "I think this is the romantic, idealized version of VC from the 60s and 70s, which is kind of YOLO. It's like, founders know better than VCs, and they've seen something that I don't see, and they're gonna go out there and do it. And, like, who am I to second guess them?" 4. Speculative "It is always so surprising how prevalent it is... They don't even care about the asset. They just want to time a market, right?... They're like, 'Well, this looks cheap. It's clearly gonna go up in value' or, 'This is the top of the market, like, I'm gonna get out.'" "...You can justify any deal or not justify any deal. So if you don't know what your beliefs are, you'll always basically just kind of justify whatever personal bias you have." @martin_casado w/ @patrick_oshag on @InvestLikeBest
martin_casado@martin_casado
x.com/i/article/2115…

66위Google Cloud@googlecloudX에서 보기

Google Cloud는 AI 에이전트를 신뢰할 수 있는 비즈니스 맥락에 기반하게 하고, 멀티클라우드와 SaaS, 온프레미스 데이터를 활용해 자율 에이전트 운영을 확장하는 방법을 다루는 방송을 안내했다.

좋아요 27리포스트 4답글 4노출 4.1K북마크 11

67위a16z@a16zX에서 보기

a16z의 아니시 아차리야와 데이비드 파울란은 사람들이 말하는 프라이버시 우려와 실제 행동 사이의 괴리를 다루며, 사람들이 작은 혜택에도 데이터를 쉽게 내준다고 지적한다.

좋아요 21리포스트 1답글 9노출 9K북마크 8

a16z@a16z
a16z's Anish Acharya and David Pawlan, creator of Assistant Benchmark, on the gap between people's stated and revealed preferences on privacy, and what that means for AI agent adoption: "I think in revealed preferences, consumers will share any amount of data for even the most trivial amount of upside." "People have been connecting their cards, their credit cards, and their bank accounts via Plaid since the Venmo era. But when you ask them, they're like, 'Oh, yes, I'm very concerned about trust and safety.'" "I think an example though of where consumers are very concerned about sort of trust and assumptions is around being embarrassed and social signaling... The social mechanics actually feel so much higher stakes than the actual, like, real-world implications." "[One company] offered a million dollars worth of pizza to anyone that wanted to use their product... In order to get that $30 of pizza, you were granting access to your entire Google Workspace, all of your emails, all of your data for 30 bucks." "It just proves it right there, that people are willing to just give up all that data for some free pizza." @illscience @DavidPawlan
a16z@a16z
David Pawlan, creator of Assistant Benchmark, has tracked 147 AI assistants so far, and in his group chats of AI power users, the number one fear is trust. It comes up in 22% of roughly 3,600 messages. He puts agents through head-to-head tests, like ordering same-day flowers as cheaply as possible or planning a weekend in LA that's both luxurious and cost-efficient, to see which ones can hold competing preferences. He joins a16z's Anish Acharya and Josh Elman to discuss why the scariest agent failures are the embarrassing ones, why agents are where e-commerce was in 1997, why scale is still a moat, and the "dark agent" that could speak to the parts of ourselves we don't reveal. 0:52 Inside a meetup for 100 agent nerds 2:52 Why we'll all have an army of agents 5:36 Trust is power users' top fear 6:28 Building fences around your agent 10:14 Privacy loses to free pizza 12:41 Can agents balance price and speed? 14:19 How agents define luxury differently 19:44 Migrating from one agent to another 21:06 What your photo roll tells an agent 25:48 Will agents make us more rational? 27:38 Why paternalism kills products 31:43 Gossipy agents as social glue 33:41 Scale is still a software moat 36:55 From OpenClaw to agents hanging out 39:23 Agents that introduce you to people 46:15 Predictions and the dark agent idea YouTube: youtu.be/dSiTwPl0t1U @DavidPawlan @illscience @joshelman

68위Vivian Midha Shen@vivianmshenX에서 보기

Anthropic에서 막 퇴사한 엔지니어 trq212가 SOTAwives 에피소드에 출연하며, 전체 방송은 내일 오전 10시 PT에 공개된다고 알린다.

좋아요 38리포스트 2답글 2노출 6.4K북마크 13

Vivian Midha Shen@vivianmshen
a hot new bombshell has entered the villa @trq212 joins @MilksandMatcha and me on @SOTAwives right after his last day at Anthropic full episode drops tomorrow 10am pt. set a reminder!
SOTAwives@SOTAwives
We adopted an ex-Anthropic engineer as our third wife. This week, @trq212 joins @vivianmshen and @MilksandMatcha right after his last day at Anthropic. From returning what was left of his Series A, to cross-lab friendships and why tech Twitter parties are overrated (ours excluded, of course). Dropping tomorrow across all platforms. Episode 2 sneak peek below:

69위TypeSafe AI@typesafeaiX에서 보기

이번 금요일 Product Hunt에서 HYPERSHIP 행사가 열리며, 참가자들이 요청을 받아 자정 전까지 제품을 출시하는 내용을 알린다.

좋아요 44리포스트 3답글 3노출 6.9K북마크 15

70위a16z@a16zX에서 보기

a16z의 분석에 따르면 최근 미국 바이아웃 펀드는 레버리지를 많이 쓰고도 러셀 3000 지수 수익률을 대체로 밑돌았으며, 벤처 펀드는 같은 기간 지수를 앞질렀다.

좋아요 19리포스트 5답글 8노출 6.7K북마크 6

a16z@a16z
Recent US buyout funds would have done better just buying the stock market US buyout returns minus the Russell 3000, by vintage: - About 9.7 points a year ahead for 2001-05 vintages - About 2.0 points a year ahead for 2006-20 vintages - Behind for every vintage since 2019 "Buyout has trailed public stocks while carrying far more debt. Over the last 10 years, US Buyout returned 14.5% a year, underperforming the Russell 3000. In contrast, US venture as a whole returned 17.6% a year over the same period (pooled across all funds), 2.3 points ahead of the Russell 3000. Unlevered." Full piece from a16z's Jen Kha: a16z.news/p/is-private-e… Not an offer or solicitation. None of the above should be taken as investment advice or an advertisement for investment services. Please see a16z.com/disclosures for more information.
Jen Kha@jkhamehl
x.com/i/article/2117…

71위David Lieb@dfliebX에서 보기

dflieb는 유전적 병력 때문에 자녀 갖기를 망설이던 가족들에게 AI와 유전체학이 도움이 될 수 있으며 이 분야가 크게 성장할 것이라고 전망했다.

좋아요 10리포스트 0답글 0노출 3K북마크 3

David Lieb@dflieb
This could help families afraid to have kids due to genetic history finally feel able to. AI+genetics is going to be huge.
Reticular (YC F24)@ReticularAI
Couples with a family history of severe autism have been reaching out, hoping we could find a genetic cause to screen for. But the genetic signal was always sparse. Now our new work on rare variants & GPN-star has increased performance by over 50%. Another step towards ending genetic disease.

72위Google Cloud@googlecloudX에서 보기

AlloyDB Express는 개발자와 코딩 에이전트를 위해 2초 안에 만들어지고 유휴 비용이 없으며 원클릭으로 프로덕션에 전환할 수 있는 PostgreSQL 서비스다.

좋아요 34리포스트 0답글 2노출 3K북마크 10

73위a16z@a16zX에서 보기

기술 분야 PE 인수 거래 비중이 2023년 34%에서 2026년 1분기 11%로 떨어졌고, 2010년대식 PE 호황은 금리 인하만으로 되살아나기 어렵다고 분석한다.

좋아요 44리포스트 4답글 9노출 11.1K북마크 11

a16z@a16z
PE's appetite for tech peaked just as the AI wave arrived Tech's share of global buyout deal value: - 23% in 2018 - 34% in 2023, the year after ChatGPT launched - 29% in 2025 - 11% in the first quarter of 2026 "SaaS, of course, was also the manna from heaven for private equity: capital-light, consistent recurring revenue that was easy to lend against, lots of small markets with clear leaders, and a small but critical line in the customer's budget, so prices could go up." "It is safe to say that the 2010s aren't coming back anytime soon: a supply of fixer-uppers, a decade of falling rates, and fast-growing funds that kept buying from each other at higher prices. A few rate cuts help the debt; they don't re-create that decade." Full piece from a16z's Jen Kha: a16z.news/p/is-private-e… Not an offer or solicitation. None of the above should be taken as investment advice or an advertisement for investment services. Please see a16z.com/disclosures for more information.
Jen Kha@jkhamehl
x.com/i/article/2117…

74위Uncle Bob Martin@unclebobmartinX에서 보기

Orbit Slinger 게임의 Level 35 Contraflow에서 37,617점으로 개인 최고 기록을 세웠다는 게임 홍보 글이다.

좋아요 9리포스트 0답글 2노출 4.9K북마크 2

75위Julia Muiruri@juliamuiruri4X에서 보기

GitHub Universe의 해킹 배지로 집 네트워크에서 Octolamp를 제어하는 프로젝트를 만들었다는 개발자의 경험을 공유한다.

좋아요 14리포스트 3답글 0노출 6.1K북마크 3

76위Google@GoogleX에서 보기

AI Studio에서 'Remix this voice'를 사용하면 생성된 세 가지 음성을 나란히 비교해 골라 계속 리믹스하고 저장해 여러 스크립트에서 재사용할 수 있다고 안내한다.

좋아요 32리포스트 3답글 4노출 18.6K북마크 6

77위OpenAI@OpenAIX에서 보기

OpenAI는 방어자가 공격자보다 먼저 취약점을 찾아 고칠 수 있도록 리눅스 관리자 및 Patch the Planet 파트너와 협력해 보고서와 패치를 진행하고 있다고 밝혔다.

좋아요 145리포스트 6답글 4노출 22.9K북마크 25

78위Google Cloud Tech@GoogleCloudTechX에서 보기

구글 클라우드 라이브에서 Gabe Weiss가 비즈니스 맥락을 반영한 Gemini로 팀이 데이터를 분류하고 초기 모델을 학습해 데이터 엔지니어에게 작동하는 프로토타입을 넘기는 방법을 소개한다.

좋아요 19리포스트 2답글 0노출 4.5K북마크 3

79위Google Cloud@googlecloudX에서 보기

Google Cloud는 AlloyDB, 보더리스 레이크하우스, Knowledge Catalog 등을 포함한 Agentic Data Cloud의 주요 업데이트를 발표했다.

좋아요 20리포스트 2답글 3노출 2.8K북마크 3

81위Greg Brockman@gdbX에서 보기

gdb는 오픈소스 코드베이스의 취약점을 무상으로 찾아 패치하는 데 상당한 프런티어 컴퓨팅을 투입하고 있으며, 리눅스 커널 연구에서 일주일 만에 약 1,000개의 로컬 권한 상승 취약점을 찾았다고 설명했다.

좋아요 34리포스트 2답글 2노출 5.8K북마크 4

Greg Brockman@gdb

@gdb 님에게 보내는 답글

additional context on the project:
Eric Wallace@Eric_Wallace_
We are now allocating a significant amount of our frontier compute towards discovering and patching vulnerabilities in open source codebases, fully pro bono. In one case, we studied the Linux kernel, which is a particularly load-bearing piece of software at OpenAI. Before AI, security researchers might discover ~20 local privilege escalation (LPE) vulnerabilities in Linux in a typical year. In one week's time, our upcoming frontier cyber model was able to find ~1000 LPE vulnerabilities using about $1M of tokens. In the process, it was able to discover at least 10 novel classes of Linux exploitation techniques.

83위OpenAI@OpenAIX에서 보기

OpenAI는 에이전트가 전문가들이 참신하다고 판단한 리눅스 익스플로잇 기법을 10가지 이상 발견해 사용했다며, 프런티어 LLM이 기존 기법 적용을 넘어 새로운 공격과 패치 방식을 찾을 수 있음을 시사한다고 밝혔다.

좋아요 138리포스트 9답글 4노출 24K북마크 9

OpenAI@OpenAI

@OpenAI 님에게 보내는 답글

The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel. This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.

84위Jordan Harband@ljharbX에서 보기

nvm v0.41.0은 카테고리별 버전 선택과 LTS 별칭 등 새 기능을 추가했고, 59개의 버그 수정과 7개의 보안 수정을 포함하며 그중 25건은 Anthropic OSS Scanner의 도움으로 찾았다.

좋아요 22리포스트 5답글 3노출 3K북마크 1

85위Oliver Wesche@oliwesche16X에서 보기

Verno가 정식 출시되었으며, 엔지니어가 P&ID 도면 변경을 설명하면 이 도구가 도면을 수정하고 관련 문서를 갱신하며 엔지니어링 규칙을 검사한다고 소개한다.

좋아요 11리포스트 2답글 5노출 2.9K북마크 0

86위Uncle Bob Martin@unclebobmartinX에서 보기

사용자는 테슬라 사이버트럭이 과소평가되었다는 내용의 유튜브 영상을 공유한다.

좋아요 19리포스트 0답글 1노출 5.7K북마크 0

87위Visual Studio Code@codeX에서 보기

전체 릴리스 노트에서 업데이트 내용을 확인하라고 안내하며 개발자들에게 즐거운 코딩을 응원한다.

좋아요 1리포스트 0답글 0노출 3.4K북마크 2

89위Visual Studio Code@codeX에서 보기

Blobby의 새 점프 게임을 소개하며, 스페이스 키로 차징하고 화살표 키로 조준하는 조작법과 채팅에서 실행하는 방법을 안내한다.

좋아요 8리포스트 1답글 2노출 3.6K북마크 2

Visual Studio Code@code

@code 님에게 보내는 답글

🎮 Need a short break? Try Blobby's new jump game! Hold Space to charge, use the arrow keys to aim, and beat your high score. Enter /vscode-pet in chat, then choose Play Jump Game from Blobby's menu in the Agents window.

90위Google Cloud@googlecloudX에서 보기

Google Cloud는 신뢰할 수 있는 에이전트와 Agentic Data Cloud의 AI 활용 사례를 다룬 추가 콘텐츠 링크를 공유했다.

좋아요 5리포스트 1답글 0노출 896북마크 0