2026년 10월 28일

홈 타임라인 게시물 90개, 인기순

1위Tim Cook@tim_cookX에서 보기

Apple은 새로운 iPad mini, 14인치 MacBook Pro, iMac에 이어 PROLED XDR 디스플레이를 갖춘 MacBook Pro에도 터치 기능을 도입해 Mac에서 처음으로 터치를 지원한다고 발표했습니다.

좋아요 52.5K리포스트 3.5K답글 2.3K노출 4.8M북마크 5.6K

Tim Cook@tim_cook
We brought Touch to the Mac! From the all-new iPad mini to the 14-inch MacBook Pro and iMac... and now MacBook Pro with PROLED XDR Display and Touch, creators have new tools to do their best work yet.

2위Google@GoogleX에서 보기

Google이 최신 프론티어 모델 Gemini 4 Carbon을 소개했으며, 여러 분야에서 성능이 향상되어 Intelligence Index 1위와 사이버보안 분야 신기록을 달성했고, Fairwind 프로그램을 통해 사이버 방어자에게 먼저 제공된다고 밝혔다.

좋아요 8K리포스트 790답글 315노출 494.7K북마크 810

Google@Google
Introducing Gemini 4 Carbon, our latest frontier model. Carbon delivers improved performance across domains — ranking #1 on v5.0 of the @ArtificialAnlys Intelligence Index and sets a new state-of-the-art in cybersecurity and long-horizon capabilities. It’s rolling out starting this week to cyber defenders through our Fairwind Program, with expanded access coming soon 🧵

3위Google DeepMind@GoogleDeepMindX에서 보기

Gemini 4 Carbon이 차세대 프론티어 모델이며, 사이버보안과 장기 과제 수행 능력에서 최고 수준이라고 소개한다.

좋아요 10.8K리포스트 1K답글 504노출 817.2K북마크 1.1K

4위Manny Hernandez@m9rcnX에서 보기

m9rcn은 Ploid를 소개하며, 전 세계 20억 명 이상을 색인해 공개 웹을 모든 사람과 기업의 하나의 프로필로 만들고 사람과 에이전트에게 저렴하게 제공한다고 밝혔다.

좋아요 480리포스트 58답글 129노출 312K북마크 601

Manny Hernandez@m9rcn
Introducing Ploid. I think the world's data should belong to everyone, so we indexed over 2 billion people globally. Ploid turns the public web into one profile for every person and company on earth, and serves it to you and your agents for cents. We're mapping the human economy, and we're live today. ploid.com

5위NVIDIA AI@NVIDIAAIX에서 보기

NVIDIA는 약 4B 활성 파라미터를 가진 30B MoE 오픈 웨이트 모델 Nemotron 3.5 Transcribe를 공개했으며, 이 모델은 40개 언어 전사, 화자 구분, 번역, 오디오 이해를 한 체크포인트에서 처리한다.

좋아요 5.3K리포스트 405답글 188노출 471.3K북마크 4K

NVIDIA AI@NVIDIAAI
Meet NVIDIA Nemotron 3.5 Transcribe 🎙️ An open-weight, promptable 30B MoE model with ~4B active parameters that brings transcription in 40 languages, speaker attribution, translation, contextual prompting and audio understanding into one checkpoint. Watch it keep track of who said what when everyone’s talking at once. 1/ 🧵

6위Vercel@vercelX에서 보기

Vercel은 베타로 Vercel CI를 선보였으며, Turborepo로 작업을 지능적으로 건너뛰고 내부 저장소에서 2~6배 빠르며 에이전트에 최적화되어 있다고 밝힌다.

좋아요 813리포스트 49답글 75노출 98.8K북마크 427

7위peetzweg/@peetzwegX에서 보기

리눅스용 OpenDisplay의 Sender와 Receiver가 출시되었고, 초기 단계라 피드백과 기여를 받고 있다고 알린다.

좋아요 950리포스트 48답글 14노출 42.1K북마크 471

8위Suhail@SuhailX에서 보기

Suhail은 Particle 출범을 알리며, 개발자 중심 모델 시리즈의 첫 단계인 particle-1이 성능을 유지하면서 토큰을 약 45% 적게 쓴다고 소개했다.

좋아요 1.4K리포스트 82답글 187노출 204.4K북마크 618

Suhail@Suhail
1/ Excited to announce our new company Particle This is our first step in a series of models focused on developers. With particle-1, we focused on token efficiency and managed to keep parity in performance while using ~45% fewer tokens. Live now: particle.ai

9위Satya Nadella@satyanadellaX에서 보기

GitHub Universe에서 Microsoft는 Copilot Fusion, 에이전트를 구축하는 Foundry의 통합 Copilot 하네스, 여러 리포지토리에서 에이전트를 조율하는 Copilot Projects 등 에이전트 시대에 맞춘 소프트웨어 개발 수명 주기의 업데이트를 발표했습니다.

좋아요 964리포스트 114답글 105노출 117.1K북마크 334

Satya Nadella@satyanadella
We're redefining the SDLC for the agentic era. Lots of GitHub updates this morning at Universe: • GitHub Copilot Fusion delivers frontier-level quality, 40% faster and at up to 77% lower cost • Our unified Copilot harness is now in Foundry, a managed service for building long-running agents that reason, use tools, and get work done • Copilot Projects gives you an orchestration loop to build with multiple agents across repos And we’re expanding capacity and strengthening GitHub’s architecture to make the home of agentic development more resilient.

10위ChatGPT@ChatGPTX에서 보기

ChatGPT는 Pro 사용자(미국)를 대상으로 미리보기 출시된 기능을 소개하며, 작업을 맡기면 텍스트, 사진, 음성 메모로 지시받고 관련 업데이트를 먼저 문자로 보내주며 iMessage나 RCS로 안전하게 주고받을 수 있다고 설명한다.

좋아요 5.3K리포스트 266답글 441노출 661.6K북마크 1.2K

ChatGPT@ChatGPT
By popular request, you can now text your dot. Give it a task to take off your plate or something to keep an eye on, using text, photos, or even voice notes. It will text you proactively with relevant updates. Safe, secure texting via iMessage or RCS. You’re the only one who can text your dot, and your dot can only text you. Rolling out now as a preview for Pro users in the US.

11위Artificial Analysis@ArtificialAnlysX에서 보기

GitHub Copilot Fusion은 Anthropic과 OpenAI 모델을 조합해 Claude Code의 Opus 5.5보다 점수는 2점 낮지만 작업당 비용은 절반 수준이라는 벤치마크 결과이다.

좋아요 416리포스트 24답글 115노출 58.4K북마크 70

Artificial Analysis@ArtificialAnlys
GitHub has released Copilot Fusion in GitHub Copilot, which orchestrates between Anthropic and OpenAI models. It scores 2 points below Opus 5.5 in Claude Code at half the Cost per Task in the Coding Agent Index GitHub describes Copilot Fusion as a runtime orchestration system, which selects and coordinates models to optimize performance, cost, and latency for each task. Copilot Fusion chooses between a number of models from Anthropic and OpenAI, including Claude Opus 5.5, Claude Sonnet 5.5, GPT-6 Astra, and GPT-6 Sol. We benchmarked its max setting in GitHub Copilot CLI in the Artificial Analysis Coding Agent Index v2.0. Token costs are charged based on the underlying models used to complete a request. Key takeaways: ➤ Cost-efficient for near-frontier performance: Copilot Fusion in GitHub Copilot costs $5.59 per task, versus $11.42 for Claude Code with Opus 5.5 (max). Devin Fusion with Opus 5.5 (max) + GPT-6 Luna (high) scores 62 at $8.00 per task. ➤ Slight lead in Terminal-Bench 4.0: The Coding Agent Index v2.0 weights its three constituent benchmarks equally. Copilot Fusion in GitHub Copilot leads Terminal-Bench 4.0 with a score of 71. It ranks third on FrontierCode in the chart at 51, behind Claude Code with Opus 5.5 (max) at 56 and Devin Fusion with Opus 5.5 (max) + GPT-6 Luna (high) at 54. Its SWE-Atlas-QnA-AA score lags at 57, between GPT-6 Astra in Codex and Gemini 4 Argon in Antigravity CLI. ➤ Half the token use of Opus 5.5 in Claude Code: GitHub Copilot Fusion used 7.2M tokens per task in the Coding Agent Index, approximately half of the tokens used by Claude Opus 5.5 in Claude Code, and 30% less than Devin Fusion with Opus 5.5 (max) + GPT-6 Luna (high).

12위Jarred Sumner@jarredsumnerX에서 보기

jarredsumner는 VC 투자로 Bun을 만들었으며 투자자들이 요구한 것은 성장 지표뿐이었고, Bun은 여전히 MIT 라이선스로 계속 개발되고 있다고 주장했다.

좋아요 930리포스트 20답글 50노출 51.8K북마크 90

Jarred Sumner@jarredsumner
The hate for VC-funded open-source always seemed deranged They gave me millions of dollars to spend on salaries for engineers to work on open-source, and together we built Bun. In return, all the VCs asked for was “make number go up” We’re still MIT licensed & still shipping
Jarred Sumner@jarredsumner
After the Series A, seemed like we’d only have the bandwidth to add & maintain a linter + formatter after growing the team by another -5 engineers - but being more than like 15 people with $0 in revenue and no progress on commercialization seemed crazy to me.

13위Satya Nadella@satyanadellaX에서 보기

사티아 나델라가 실적 발표를 마쳤으며, Microsoft Cloud 매출이 630억 달러를 넘고 30% 성장했다고 밝힌다.

좋아요 1.2K리포스트 91답글 139노출 129.5K북마크 102

Satya Nadella@satyanadella
Just wrapped our earnings call. It was a fast start to our fiscal year. Microsoft Cloud revenue surpassed $63B, with growth accelerating to 30% YoY. Lots of innovation as we build the new tech stack for the agentic wave. A few highlights:

14위Google AI@GoogleAIX에서 보기

Google이 사이버보안과 장기 과제 수행 능력이 뛰어난 최신 모델 Gemini 4 Carbon을 발표했으며, 우선 신뢰할 수 있는 사이버 방어자들에게 Fairwind 프로그램으로 제공하고 접근을 곧 넓힐 예정이라고 밝혔다.

좋아요 3.7K리포스트 361답글 174노출 189.5K북마크 287

Google AI@GoogleAI
Announcing Gemini 4 Carbon, our strongest model with state-of-the-art cybersecurity and long horizon capabilities. Carbon represents a step up across most benchmarks, delivering strong frontier performance in autonomous coding and cyber defense. We’re first making Carbon available to a subset of trusted cyber defenders in our Fairwind program. We’ll expand access as soon as possible.

15위Mitchell Hashimoto@mitchellhX에서 보기

Mitchell Hashimoto는 Siri가 간단한 질문조차 제대로 답하지 못하고, macOS Siri 앱 디자인도 매우 조잡하다고 강하게 비판했다.

좋아요 1.6K리포스트 51답글 142노출 92.6K북마크 125

Mitchell Hashimoto@mitchellh
Siri is still bad, like really really bad. I opened an email with my Uber Eats confirmation, there was a live activity on the screen with progress, and I pulled down and asked Siri "what did I order for lunch?" And it thought for like 2 minutes and then said it doesn't know. 🤦‍♂️ Okay so the harness and models are clearly bad. Well, then next I opened up the Siri GUI app on macOS and HOLY. It's the worst looking, beginner, basic bitch app I've ever seen. Did someone spend more than 5 minutes on this? Whoever worked on this app along with the full chain of people who checked off on this needs to be terminated immediately. This is 0/10 terrible software, so embarrassed for Apple to ship this.

16위Demis Hassabis@demishassabisX에서 보기

Gemini 4 Carbon이 주요 벤치마크에서 SOTA를 기록했으며, 최첨단 역량과 심층적인 안전 및 보안 연구를 결합했다고 전한다.

좋아요 5.2K리포스트 290답글 233노출 214.6K북마크 268

17위a16z@a16zX에서 보기

a16z의 CISO 설문에 따르면 AI가 보안 우선순위를 바꿔 2027년 상위 우선순위 7개가 모두 AI 관련 항목이며, 내부 AI 에이전트 보안과 AI SOC 등이 가장 크게 늘었다.

좋아요 483리포스트 70답글 47노출 63.9K북마크 477

a16z@a16z
In a single year, AI has rewritten the security priority list for CISOs. Attackers are moving faster. Bain estimates AI compresses a typical cyberattack from about four weeks to about 18 hours, and models keep getting better at finding and exploiting weaknesses. There's also more to protect, with employees shipping apps they couldn't have built a year ago and agents getting access to source code, documents, and production systems. In a survey of 100 CISOs at companies with 1,000+ employees, all seven of the most-selected 2027 priorities involve AI. The biggest jumps since 2026: - Internal AI agent security (+28 pts) - AI SOC (+26 pts) - AI offensive security (+24 pts) - AI code security (+23 pts) - AI vulnerability remediation (+22 pts) Last year's staples: - Exposure & vulnerability management (-21 pts) - SIEM / security data (-18 pts) - Privileged access management (-18 pts) From a16z's @OwenColby and @steph_zhang on The New Security Agenda: a16z.news/p/the-new-secu…
Steph Zhang@steph_zhang
x.com/i/article/2115…

18위Microsoft Developer@msdevX에서 보기

Microsoft는 Foundry Agent Service에 Copilot 하네스를 도입해 오케스트레이션, 도구, 컨텍스트, 제어를 제공하므로 개발자가 실행 계층을 다시 만들지 않고도 에이전트를 쉽게 구축할 수 있게 했다.

좋아요 212리포스트 17답글 8노출 26.8K북마크 178

Microsoft Developer@msdev
The @github engineering team built a best-in-class, truly multi-model harness, which we've implemented across Microsoft products. Introducing Microsoft Copilot harness in Foundry Agent Service, a simpler way to build agents without rebuilding the execution layer. It handles orchestration, tools, context, and controls so you can focus on your agent’s skills and outcomes. Watch @jeffhollan's demo to see it in action.

19위Hugging Face@huggingfaceX에서 보기

Hugging Face는 AI 및 머신러닝 연구에 입문하기 좋은 NeurIPS 2026 재현 해커톤을 곧 열 예정이라고 알렸다.

좋아요 464리포스트 22답글 8노출 31.1K북마크 371

20위a16z@a16zX에서 보기

a16z의 설문 조사에 따르면 CISO의 18%가 현재 공격적 보안 테스트 대부분을 지속적으로 수행하고 있고, 52%는 1년 안에 그렇게 할 계획이며, 70%는 AI로 수정 사항이 실제로 효과가 있는지 재검증하고 싶어 한다.

좋아요 76리포스트 12답글 15노출 12.5K북마크 57

21위Seungmin Hong@shong0504X에서 보기

Simantic Bot은 AI가 시뮬레이션에서 펌웨어를 실행하고 GitHub PR 안에서 ISR, 주변장치 타이밍, 메모리 버그를 테스트하도록 돕는 도구를 출시했다.

좋아요 152리포스트 12답글 13노출 9.8K북마크 107

Seungmin Hong@shong0504
We're launching Greptile for firmware. AI code reviewers are great for catching bugs in software, but when it comes to firmware, they miss critical bugs in ISR and peripheral timings, and runtime memory. We're launching Simantic Bot to fix that. We gave AI the ability to run firmware in simulation, and test scenarios like providing external outputs from sensors and wireless communication, all inside your Github PRs. Check it out at simantic.com

22위a16z@a16zX에서 보기

Mark Cuban은 a16z와의 대담에서 Cost Plus Drugs의 성장 경험, AI가 가져올 큰 변화, 그리고 AI를 잘 활용하지 못하는 기업은 사라질 것이라는 견해를 이야기했다.

좋아요 219리포스트 34답글 22노출 46.2K북마크 145

a16z@a16z
Mark Cuban's advice for young people boils down to "there's gonna be two types of companies: those who are great at AI and those who have gone out of business." "And then there's two types of people in how they use AI: those who use AI to learn anything, and those who use AI so they don't have to learn anything." "And if you're one of those people who uses AI so that you can learn anything, and you realize what type of companies and problems aren't being solved by AI, don't be shy to just talk to your AI." "And I say this not just to kids, but to 50 60, 70-year-old entrepreneurs. The biggest challenge right now with entrepreneurs using AI is the comfort level of talking to your AI... there's still a lot of hesitancy." "Just like back in the day, those first spreadsheets, people still had their 10-key adding machine, and they would check it to make sure the spreadsheets were right." "So when you ask me, is there one thing I truly believe, it's everything. Everything will change. This is the greatest transitional technology we have ever seen, and I've been around for all of them, right? This blows them all away." @mcuban @julesyoo @eriktorenberg
a16z@a16z
Mark Cuban scaled Cost Plus Drugs from about 100 generics to roughly 4,000 at a flat 15% markup, forcing incumbents to change prices in ways decades of policy never did. He thinks AI could transform healthcare, but it won't fix opaque prices and broken incentives on its own. Mark sits down with a16z's Julie Yoo and Erik Torenberg to discuss what he's learned building Cost Plus Drugs, why trust is its core product, and how greater transparency could put pressure on healthcare costs. They also explore why drugs designed for your DNA could cost close to nothing, the football field-sized drug plant his team shrank into tractor trailer-sized pods, his plan to make generics in Dallas for less than he pays manufacturers in India, and why every company will either be great at AI or go out of business. 0:50 Cost Plus beat decades of policy 3:15 111 generics to 4,000 drugs 5:00 A drug plant shrunk to 5K sq ft 5:55 $20M drug factories in a trailer 8:10 Hospital deals with zero denials 9:50 Why health plans are rigged 14:05 1,335 pages no human can check 18:35 "Where are they f**king me over?" 22:45 Making drugs cheaper than India 26:50 Why you still need entry-level hires 31:45 Nonprofit hospitals must offer aid 35:55 DNA-matched drugs for near $0 39:30 Why AI will mean more doctors 48:00 Mark's plan: "bye bye insurers" 52:25 Why AI labs' ads are useless 56:35 Get good at AI or go out of business YouTube: youtu.be/5UWDUD6qaM4 @mcuban @julesyoo @eriktorenberg

23위Maithra Raghu@maithra_raghuX에서 보기

maithra_raghu는 금융 리서치용 모델 Samaya Sterling을 소개하며, 작은 30B 모델을 네 단계로 학습시켜 FrontierFinance에서 Claude Fable 5.1보다 높은 정확도를 내고 비용은 11배 낮췄다고 설명했다.

좋아요 81리포스트 13답글 5노출 12.7K북마크 47

Maithra Raghu@maithra_raghu
Introducing Samaya Sterling, our new model post-trained for challenging financial research that outperforms the frontier! Sterling started off as a two person research effort to see if we could increase efficiency in agentic orchestration. To our surprise we were able to post-train a much smaller model (Nemotron 3.5 Lightning, a 30B mixture-of-experts model with 3B active parameters) to greater accuracy (58.9% vs 55.9% on FrontierFinance) and 11x lower cost ($0.61 vs $6.89) than Claude Fable 5.1, the best performing frontier model. We did this via a four-stage training process, two stages of SFT where the model first learns to research with our financial tools and how to construct a good agentic research strategy. The second two stages are RL, with an offline GRPO stage where the model further hones its research strategy from eight teachers that provide diverse approaches, and an online GRPO stage where the model can make end to end attempts and learn from its mistakes. We see substantial gains from the co-adaptation of the model and harness through post-training, both through the model learning to effectively use the high-quality financial tools, and better compaction approaches that let even a small model execute on longer and deeper research. We are exploring multiple follow-up directions, from multi-agent research swarms using this small model, to training firm-specific models to encode unique judgement, to training a larger model, where early results show even greater gains from the training process. The full research is linked below!

24위a16z@a16zX에서 보기

a16z의 100명 CISO 설문에 따르면 기업 전반에 AI가 이미 도입되어 있고, 주요 우려는 AI 공격이 대응 속도를 앞지르는 것, 직원의 데이터 유출, AI 에이전트의 권한 초과 또는 침해다.

좋아요 110리포스트 9답글 26노출 15.5K북마크 61

a16z@a16z
AI is already running across the enterprise, and securing it is now one of CISOs' biggest concerns Already in production at their companies: - Employee generative AI tools (92%) - AI coding assistants or agents (83%) - MCPs or agent-to-app integrations (55%) - Autonomous agents with tool access or write permissions (34%) Their biggest concerns: - AI attacks outpacing their team (55%) - Employees leaking data into AI tools (47%) - AI agents exceeding permissions or getting compromised (46%) From a16z's @OwenColby and @steph_zhang, breaking down a survey of 100 CISOs at companies with 1,000+ employees: a16z.news/p/the-new-secu…
Steph Zhang@steph_zhang
x.com/i/article/2115…

25위E2B@e2bX에서 보기

Y Combinator가 내부 에이전트 하네스인 QM을 오픈소스로 공개했으며, E2B 샌드박스를 연결하면 팀원마다 격리된 에이전트 컴퓨터를 제공할 수 있다.

좋아요 30리포스트 7답글 13노출 3.6K북마크 15

E2B@e2b
.@ycombinator open-sourced the agent harness it uses internally, @qm__dev. It gives everyone on your team their own agent in Slack, so every agent needs a computer to work in. That computer can now be an E2B sandbox. Set "sandbox.backend": "e2b" and every person and channel gets an isolated sandbox from the qm-sandbox template, with the GitHub CLI, the AWS CLI, Claude Code and Codex preinstalled. Ask QM to run your tests and open a PR, and it does the work in your sandbox without touching anyone else’s. Run all of QM on E2B with e2b sandbox create qm, or point your existing deployment at it.
docs.e2b.devQM - E2B Docs

26위a16z@a16zX에서 보기

Martin Casado는 VC 논쟁이 서로 충돌하는 네 가지 투자 관점이 뒤섞여 혼란스러워진다며, 팀이 하나의 핵심 철학을 정해 일관되게 지키거나 단계별로 발전시켜야 한다고 조언한다.

좋아요 55리포스트 3답글 13노출 14.1K북마크 27

a16z@a16z
a16z's Martin Casado says many debates about VC turn into "a muddled, incoherent mess" because investors pull arguments from four investing mindsets that often conflict: 1. Financial: the business's finances are the asset, so investors judge deals on metrics like churn, margin, and growth, plus the size of the market. 2. Productive: the tech is the asset, and it creates new markets, so investors focus on tech, team, and market fit, and treat today's TAM as secondary. 3. Frontier: the founders are the asset, so investors back the best team in a wildly new space well before anything is working. 4. Speculative: investors have little visibility into the company, so they bet on the market's perception of the deal, with a heavy focus on price and timing. Each one is coherent on its own, but using them as a grab bag breaks down, because "it's hard to directly refute a finance argument with a product argument since they make entirely different assumptions on what the bet is for value creation." Martin's answer is to pick one core philosophy as a team and stick to it across deals, or evolve it by stage: frontier before there's a product, productive before traction, and financial once the business is repeatable. He's "strictly a productive investor" and thinks new tech tends to create its own market, like "Uber and taxis, GitHub and the repository, Cursor and the IDE." In his words, "If you truly believe a technology is revolutionary, TAM concerns should be third order at best." Full piece from Martin:
a16z.comThe Evolving Landscape of Venture Capital Mindsets | Andreessen Horowitz
martin_casado@martin_casado
x.com/i/article/2115…

27위Y Combinator@ycombinatorX에서 보기

Broccoli는 배관, 냉난방, 전기 업체의 전화를 24시간 AI 에이전트가 응대해 작업을 예약하는 서비스로, 시리즈 A와 B에서 5,200만 달러를 조달했고 1,500개 이상의 업체가 사용 중이다.

좋아요 87리포스트 7답글 16노출 13K북마크 39

Y Combinator@ycombinator
Congrats to @call_broccoli (W22) on raising $52M across their Series A and Series B! Broccoli’s AI agents answer calls for plumbing, HVAC, and electrical companies around the clock, booking jobs directly into the scheduling software contractors already use and following up on open estimates. After their EV startup stalled, @abhishekjain25 and @adityaketkar knocked on 140 contractor doors until a Sacramento plumbing owner gave them five minutes. They built the first product from her office in three weeks, and today she’s Broccoli’s COO. More than 1,500 home service businesses now run on Broccoli, with over 10 million customer conversations handled and 2.5 million appointments booked.
fortune.comMeet the Baby Boomer whose life changed when 2 Gen Z AI founders walked into her plumbing business....

28위a16z@a16zX에서 보기

a16z는 AI의 사이버 공격 능력이 약 2년 사이 크게 향상되었다고 말하며, 2024년 GPT-4o는 2단계에서 멈췄지만 최신 모델은 토큰 1억 개까지 계속 개선되어 평균 4분의 3 지점까지 진행하고, 최고 시도는 32단계를 모두 끝내 네트워크 전체를 장악했다고 설명한다.

좋아요 105리포스트 11답글 23노출 14.4K북마크 47

a16z@a16z
AI's cyber skills have taken a huge leap in about two years In a test that simulates a 32-step cyberattack on a company network, tracking how far each model gets as it uses more tokens: - GPT-4o, a 2024 model, stalled at step 2 and never improved with more tokens - Today's top models kept improving all the way to 100M tokens, getting about three-quarters of the way through on average - The best attempts finished all 32 steps and took over the entire network From a16z's @OwenColby and @steph_zhang on The New Security Agenda: a16z.news/p/the-new-secu…
Steph Zhang@steph_zhang
x.com/i/article/2115…

29위Vercel@vercelX에서 보기

Vercel은 AI Gateway에 가상 모델 기능을 도입해, 여러 에이전트에서 하나의 가상 모델 ID로 모델 설정을 중앙 관리하고 새 모델로 즉시 전환하며 Zero Data Retention을 적용할 수 있다고 발표했다.

좋아요 145리포스트 7답글 20노출 19.6K북마크 63

30위Jeff Hollan@jeffhollanX에서 보기

jeffhollan은 Microsoft Foundry의 Agent Service에 새 Microsoft Copilot 하네스가 추가되어 코드 없이 장기 실행 기업용 에이전트를 만들고 관리 및 보안 처리할 수 있다고 알린다.

좋아요 89리포스트 19답글 8노출 19.7K북마크 38

Jeff Hollan@jeffhollan
Building powerful long running agents requires a powerful agent harness. Now it's easier than ever inside of @Microsoft Foundry with the new Microsoft Copilot harness in Foundry Agent Service! 🧑‍💻 Build and govern a powerful agent powered by the same harness optimized across GitHub, Microsoft 365, and Copilot scenarios for agents that can accomplish complex and long running enterprise tasks -- no code required. All secured and governed in Foundry 🔐 Check it out: devblogs.microsoft.com/foundry/micros…

31위a16z@a16zX에서 보기

Mark Cuban이 AI 시대의 인력, 헬스케어 가격 투명성, Cost Plus Drugs 경험을 이야기한 a16z 대담을 소개한다.

좋아요 40리포스트 6답글 14노출 9.6K북마크 17

a16z@a16z
Mark Cuban's rule for the AI-era workforce: "People who know how to use AI replace the jobs of people who don't know how to use AI." "[Healthcare] looks the same as any other industry. You have to still hire entry-level people." "No matter what the model is, the minute it's released, it's behind. There's latency associated with every model." "You need people who are aware of everything that's going on, that create context and deal with the models, be the human in the loop, understand the consequences." "You need those people to grow with the organization over time so that as all these things change, they've seen the history of the implementation, and they can bring value to the organization." @mcuban @julesyoo @eriktorenberg
a16z@a16z
Mark Cuban scaled Cost Plus Drugs from about 100 generics to roughly 4,000 at a flat 15% markup, forcing incumbents to change prices in ways decades of policy never did. He thinks AI could transform healthcare, but it won't fix opaque prices and broken incentives on its own. Mark sits down with a16z's Julie Yoo and Erik Torenberg to discuss what he's learned building Cost Plus Drugs, why trust is its core product, and how greater transparency could put pressure on healthcare costs. They also explore why drugs designed for your DNA could cost close to nothing, the football field-sized drug plant his team shrank into tractor trailer-sized pods, his plan to make generics in Dallas for less than he pays manufacturers in India, and why every company will either be great at AI or go out of business. 0:50 Cost Plus beat decades of policy 3:15 111 generics to 4,000 drugs 5:00 A drug plant shrunk to 5K sq ft 5:55 $20M drug factories in a trailer 8:10 Hospital deals with zero denials 9:50 Why health plans are rigged 14:05 1,335 pages no human can check 18:35 "Where are they f**king me over?" 22:45 Making drugs cheaper than India 26:50 Why you still need entry-level hires 31:45 Nonprofit hospitals must offer aid 35:55 DNA-matched drugs for near $0 39:30 Why AI will mean more doctors 48:00 Mark's plan: "bye bye insurers" 52:25 Why AI labs' ads are useless 56:35 Get good at AI or go out of business YouTube: youtu.be/5UWDUD6qaM4 @mcuban @julesyoo @eriktorenberg

32위Harpreet Dhaliwal@harpreetd14X에서 보기

YC F26 소속 AI 네이티브 트럭킹 회사 TrueMile은 화물 계획, 탐색, 중개인 협상, 기사 배차를 수행하는 AI를 출시했으며, 이 AI가 고른 화물은 시장보다 마일당 17% 높은 수익을 냈다.

좋아요 57리포스트 6답글 21노출 13.9K북마크 23

33위Autosana (YC S25)@autosanaaiX에서 보기

Autosana는 iOS, Android, 웹 등 전체 스택에 걸친 종단간 코드 검증 도구를 무료로 제공하며, 이는 소프트웨어의 가장 큰 문제를 해결하려는 것이다.

좋아요 40리포스트 14답글 13노출 3.5K북마크 16

34위Finn Mallery@fin465X에서 보기

fin465은 Origami 회사의 지난 2년간 월 매출 그래프를 공유하며, 회사는 포기하지 않는 한 끝나지 않으니 무언가 통할 때까지 계속해야 한다고 말했다.

좋아요 124리포스트 8답글 18노출 10.9K북마크 48

Finn Mallery@fin465
Monthly revenue graph (MRR) for our company Origami over the last 2 years: Your company only dies if you decide it’s over Keep going until something works (^for reference, these 4 dips were each months long periods 😁)

35위GitHub@githubX에서 보기

github는 개발자의 역할이 바뀌었지만 의미 있는 소프트웨어 작업이라는 목표는 같다며, 배포될 내용을 통제하면서 창작을 돕는 업데이트를 출시했다고 말한다.

좋아요 102리포스트 15답글 21노출 21.8K북마크 37

36위Garry Tan@garrytanX에서 보기

Garry Tan은 개발자 중심 모델 시리즈의 첫 단계로 Particle을 발표했으며, particle-1은 성능을 유지하면서 토큰을 약 45% 적게 사용한다고 밝혔다.

좋아요 148리포스트 4답글 35노출 17.5K북마크 52

Garry Tan@garrytan
Particle will be on the frontier for smartest model at lowest dollar per unit of intelligence
Suhail@Suhail
1/ Excited to announce our new company Particle This is our first step in a series of models focused on developers. With particle-1, we focused on token efficiency and managed to keep parity in performance while using ~45% fewer tokens. Live now: particle.ai

37위Jay Parikh@jayparikhX에서 보기

GitHub는 #GitHubUniverse에서 에이전트 시대에 맞춰 GitHub를 어떻게 발전시킬지 발표했다.

좋아요 62리포스트 23답글 5노출 12.9K북마크 21

39위GitHub@githubX에서 보기

GitHub는 Copilot Fusion Max가 Claude Code Opus 5.5 Max와 2점 이내 성능에 API 비용은 약 절반이라며 11월 30일까지 30% 할인을 홍보했다.

좋아요 350리포스트 35답글 53노출 46.8K북마크 105

40위Visual Studio Code@codeX에서 보기

GitHub Copilot의 새로운 기능을 소개하며, 여러 기기에서 에이전트 실행, 모델 결과 비교, 에이전트 워크플로 맞춤 설정 등 최신 업데이트를 안내한다.

좋아요 139리포스트 15답글 6노출 13.9K북마크 40

41위Greptile@greptileX에서 보기

TREX에 컴퓨터 사용 기능이 추가되어 변경 사항에 맞는 서비스를 실행하고, 데스크톱 앱, 웹 사용자 흐름, CLI를 처음부터 끝까지 직접 테스트할 수 있게 되었습니다.

좋아요 86리포스트 7답글 17노출 11.2K북마크 22

Greptile@greptile
Introducing Computer Use in TREX. TREX now spins up the relevant services for your changes and uses computer use to test them end to end. It can test desktop apps, click through a web app in a browser to simulate user flows, or use a terminal to test your CLI.

43위Spenser Skates@spenserskatesX에서 보기

Paul과 Jessica는 Amplitude가 가장 힘들던 시기에 저와 오랜 시간을 함께했고, YC가 창업자를 오래 돕는 이유도 이들처럼 진심으로 신경 쓰는 투자자이기 때문이라고 말합니다.

좋아요 113리포스트 4답글 5노출 20.3K북마크 27

Spenser Skates@spenserskates
Both Paul and Jessica spent extended time with me a few years ago during the toughest parts of Amplitude. I met up with them multiple times in Palo Alto and then Anne and I got to visit them in the Cotswolds as well. It's partly that they were there at the beginning, but I think it's deeper in their case, they can't help but care. That's a very powerful combintation. Funnily I don't have a single picture of Paul with a straight face.
Paul Graham@paulg
One of the most unexpected things about YC is how long we continue helping the founders. Even though we're generally the first investors, we're the ones still advising them 10 years later. I think it's because we're the investing analogue of old friends.

44위Google@GoogleX에서 보기

Google은 Search와 Gemini 앱에서 개발자가 앱을 한 곳에서 만들고 관리할 수 있는 Connected Apps Platform 프리뷰를 출시하고, Adobe, Calm, Peloton 등 파트너의 새 연결 앱도 함께 공개했다고 발표했다.

좋아요 340리포스트 30답글 42노출 70.6K북마크 76

45위Arav Mathur@aravmathur23X에서 보기

aravmathur23은 앱에서 누가 자금을 옮기거나 고객 데이터를 내보내는지, 사람이든 AI든 무엇을 할 수 있는지 통제하는 GrayPass를 소개하며, 90만 건 이상의 세션을 처리하고 이번 달 매출 5만 달러를 달성했다고 밝혔다.

좋아요 110리포스트 21답글 47노출 6.9K북마크 24

46위Tianyi Evans Gu@iamtianyiguX에서 보기

Confetti는 평범한 영어로 하드웨어를 설계하고 만들 수 있게 해 주는 AI 플랫폼으로, 아이디어부터 3D 설계와 부품 배송까지 한곳에서 처리한다.

좋아요 128리포스트 8답글 53노출 5.4K북마크 24

Tianyi Evans Gu@iamtianyigu
Introducing Confetti, a platform that helps you design and build hardware with plain English. Here’s how: - explain your idea at confetti.build - that’s it. Confetti is the first AI platform for vibe-manufacturing which lets anybody go from idea to 3D design to parts at your doorstep all in one place. Creating hardware has always required skills and resources, meaning our ideas are often left as just ideas. But making things should be for everybody.

47위Google@GoogleX에서 보기

Google의 Gemini 앱에 새로운 몰입형 탐색 라이브러리가 추가되어 우주 탐사, 제트 엔진, 중생대 등 주제를 시각적으로 층별로 탐색할 수 있게 되었다.

좋아요 364리포스트 40답글 37노출 88.8K북마크 55

Google@Google
Dive deep into visual experiences of your favorite topics with immersive explorations in the @GeminiApp. 🦕 Whether you're fascinated by space exploration, the inner workings of a jet engine, or what the world was like during the Mesozoic Era, there’s a topic for everyone. You can explore what interests you in the new library of immersive explorations and get an interactive, visual way to explore topics layer-by-layer.

48위AI at Meta@AIatMetaX에서 보기

Meta는 Caltech 등 다섯 개 대학의 연구를 지원하기 위해 Muse 모델, 오픈 웨이트 모델, SAM과 DINO 같은 비전 모델, SeamlessM4T를 제공한다고 발표했습니다.

좋아요 132리포스트 16답글 26노출 22.2K북마크 15

AI at Meta@AIatMeta
Today we're announcing support for university research at @Caltech, @OhioStateMed, @LSU, @IndianaUniv, and @UTHealthHouston to help advance breakthroughs across science and medicine. As part of the support for university researchers, we’ll provide access to our latest Muse models, including Muse Wonder and Muse Spark; open weight models like Muse Glimmer and Muse Spark Open; computer vision models such as SAM, V-JEPA, and DINO; as well as SeamlessM4T, our multilingual communication model. We’re excited to support these five universities as they use our models to accelerate the important work they’re doing. Read more on our research blog: go.meta.me/47G6wV3

49위Node.js@nodejsX에서 보기

Node.js 26 Lithium이 LTS가 되었고 v26.13.0은 2029년 4월까지 업데이트를 받으며, 이전 LTS에서 올라오는 리눅스 사용자는 libatomic을 별도로 설치해야 할 수 있다.

좋아요 587리포스트 97답글 11노출 26.9K북마크 64

51위Google@GoogleX에서 보기

Carbon은 취약점 발견, 악용 검증, 근본 원인 패치 등 사이버 방어 전 과정에서 최고 성능을 보였고, 내부 테스트에서는 보안 코드 생성 성공률이 Gemini 4 Argon보다 2배 이상 높아졌다고 전한다.

좋아요 262리포스트 14답글 4노출 40.4K북마크 20

Google@Google

@Google 님에게 보내는 답글

Carbon delivers state-of-the-art performance across the cybersecurity defense lifecycle: 🔵 Finding real vulnerabilities: 88.2% on SEC-bench Pro 🔵 Validating risk through exploitation: 54.8% on ExploitGym 🔵 Patching at the root cause: 71.0% on Collinear’s CWE-bench On internal benchmarks, Carbon drives a more than 2x uplift in success rate for generating secure code over Gemini 4 Argon.

52위Sava Trust Company@savatrustX에서 보기

Sava Trust는 1,000만 달러 투자를 유치하고 네바다에서 AI 기반 신탁회사를 정식 출범했으며, 기존의 수동적인 신탁 설립 과정을 개선하겠다고 밝혔다.

좋아요 53리포스트 12답글 8노출 110.4K북마크 4

Sava Trust Company@savatrust
Today, we're excited to announce that we have raised $10 million in funding and officially launched Sava Trust Company, an AI-native, modern trust company chartered in Nevada, open to clients and their attorneys & advisors nationwide. Our funding was led by @GradientVC, with participation from @ycombinator, @svangel, @Max, @orangecollectv and @SurgePointCap. For too long, trust administration has relied on manual processes, disconnected systems, and complex fee structures. Establishing a trust can take weeks or months, with limited visibility for families and their advisors over the life of the trust. With Sava, we're building a better way.
youtube.comSava Trust Company Launch

53위Google Cloud@googlecloudX에서 보기

Google Cloud는 11월 9일 오전 9시 PT에 진행되는 라이브 방송을 안내하며, 에이전트 시대에 필요한 새로운 인프라와 기업 AI를 뒷받침하는 하드웨어 기술을 소개한다고 알립니다.

좋아요 90리포스트 8답글 5노출 129.5K북마크 6

Google Cloud@googlecloud
Like this post for reminders 24 hours prior and when we go live on November 9 at 9 AM PT. Delivering on the agentic era demands new kinds of infrastructure. Join Google Chief Technologist Amin Vahdat to explore the hardware breakthroughs powering enterprise AI.

54위Google DeepMind@GoogleDeepMindX에서 보기

Carbon은 복잡한 작업을 며칠 동안 멈추지 않고 수행해 고품질이고 신뢰할 수 있는 코드를 만들며, 초기 사이버 방어자들의 피드백으로 가드레일을 개선한 뒤 접근을 확대한다고 전한다.

좋아요 458리포스트 21답글 25노출 49.3K북마크 30

55위Mitchell Hashimoto@mitchellhX에서 보기

Mitchell Hashimoto는 Rex 베타 출시 현황으로, 이번 주 1만 2천 명을 추가 초대했고 iOS 테스트는 초대받은 모든 사람에게 열렸으며, 서버 테스트는 약 2천 명까지 늘었고 대기자 명단의 절반 가까이를 처리했다고 전했다.

좋아요 506리포스트 10답글 51노출 19.4K북마크 33

Mitchell Hashimoto@mitchellh
Latest updates on Rex beta rollout: - I'll have a formal update next week (blog post). - We've invited 12,000 more this week. - We've fully opened iOS testing to everyone invited - Server (Linux + FreeBSD) testing is up to ~2,000 and we're going to open that up to everyone invited soon. - We're just shy of 50% through out full waitlist, but we're now inviting thousands every day and growing and this should clear quickly.

56위Satyam@KlausCodesX에서 보기

YC에서 얻은 가장 큰 교훈은 준비가 되었다고 느끼지 않더라도 최대한 일찍 출시하는 것이 좋다는 것입니다.

좋아요 77리포스트 3답글 11노출 4.6K북마크 5

57위Google@GoogleX에서 보기

Google이 Gemini 4 Carbon을 공개했으며, Intelligence Index v5에서 59점으로 1위를 차지했다고 발표한다.

좋아요 328리포스트 30답글 34노출 42K북마크 21

Google@Google
Today we shared an early look at Gemini 4 Carbon. It ranks #1 on the @ArtificialAnlys Intelligence Index v5 with a score of 59 (+9 pts over Argon), leading across complex reasoning, knowledge, and scientific tasks.
Artificial Analysis@ArtificialAnlys
Google’s Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v5, and second only to Claude Opus 5.5 on the Artificial Analysis Coding Agent Index Gemini 4 Carbon scores 59 on Intelligence Index v5, driven by leading performance in complex reasoning, knowledge, and scientific tasks. It also has large gains in coding and knowledge work, particularly when compared to other Gemini models, but it is not leading in all these areas. As with launch of Gemini 4 Argon, Gemini 4 Carbon is being rolled out to selected users and is not yet publicly available. Google is offering a 50% launch discount, with no confirmed end date. All our charts and figures use standard pricing for both Gemini 4 Carbon and Argon. Key results for Gemini 4 Carbon with high reasoning: ➤ Gemini 4 Carbon leads the Intelligence Index v5, with the #1 score in 4 of the 9 Index evaluations: HLE-Diamond, Terminal-Bench-Science, AA-Omniscience, AutomationBench-AA. Carbon makes substantial gains elsewhere but is not leading in coding and knowledge work ➤ Behind only Claude Opus 5.5 on coding agents: With Antigravity CLI, Gemini 4 Carbon (high) scores 59 only 3 points behind Claude Code with Claude Opus 5.5 (max, 62) on the Artificial Analysis Coding Agent Index. Carbon ranks second on both Terminal-Bench 4.0 (64%) and FrontierCode v1.1 (51%), behind Claude Opus 5.5 (max) at 56%. Google has historically underperformed in this area and so this improvement stands out ➤ Improved agentic knowledge work, but not yet leading: On AA-Briefcase, Gemini 4 Carbon (high) scores 1609 Elo, a 126-point improvement over Gemini 4 Argon (high, 1483), but still behind models including Grok 4.7 (xhigh, 1638), Muse Wonder (max, 1659) and Kimi K3.1 Pro (1678). The largest gain comes from Presentation, up 293 Elo to 1598, alongside a 120-point improvement in Analytical Quality to 1671 ➤ Lower Cost per Task than Claude Opus 5.5 for comparable overall intelligence: Gemini 4 Carbon costs $7.76 per Intelligence Index task, ~23% less than Claude Opus 5.5 (max, $10.04) but ~1.2x GPT-6 Astra (max, $6.25). It uses 123k output tokens per task, 29% fewer than Claude Opus 5.5 (max, 173k). At the same $4/$20 standard pricing, its Cost per Task is in line with Gemini 4 Argon ($7.68), while scoring 9 points higher on the Intelligence Index ➤ Leading AA-Omniscience v1.1 score, with low hallucination: Gemini 4 Carbon scores 54 on our recently updated AA-Omniscience v1.1, 6 points ahead of Claude Opus 5.5 (max, 48) and 8 points ahead of Gemini 4 Argon (46). Accuracy (% correct) rises 10 points over Argon to 58%, while its hallucination rate rises from 4% to 11%. This is still the second lowest of any model, behind only Argon, compared with 50% for both Claude Opus 5.5 (max) and GPT-6 Astra (max) Key model details: ➤ Context Window: 1M tokens ➤ Multimodality: Text, image, video, and speech input, with text output ➤ Pricing: $4/$20 per 1M input/output tokens, the same as Gemini 4 Argon’s standard pricing, with a 50% discount. Cached input tokens receive a 95% discount ($0.20 per 1M) Congratulations to the @GoogleDeepMind team!

58위NASA@NASAX에서 보기

CHAPEA 2 승무원들이 1년간의 화성 생활 모의 임무를 마치며 마지막 소감을 Houston We Have a Podcast에서 공유한다.

좋아요 515리포스트 67답글 62노출 260.9K북마크 28

59위Satya Nadella@satyanadellaX에서 보기

Satya Nadella는 GitHub Universe 2026의 전체 스택 발전 내용을 자세히 설명하는 글을 읽어 보라고 링크를 공유했습니다.

좋아요 39리포스트 4답글 7노출 15.5K북마크 24

60위Chandrika Maheshwari@Chandrika633X에서 보기

Chandrika633은 YC F26 창업자들이 PostHog의 James에게 들은 조언을 정리하며, 작은 조각이라도 빨리 제품을 써보게 하고 고객과 꾸준히 관계를 맺으며 잘못을 빨리 인정하고 방향을 바꾸는 결단력이 중요하다고 공유했다.

좋아요 18리포스트 2답글 5노출 3.1K북마크 9

Chandrika Maheshwari@Chandrika633
last night, yc f26 founders heard from @james406 from @posthog one thing that he said resonated with me deeply: "being more ambitious can make a startup easier to build - not because the work gets easier, but because the idea gives customers, teammates, and investors something to get excited about." the challenge is making progress on that ambition *this week* james shared an analogy i loved: “elephant carpaccio.” instead of building the feet first and promising the rest of the elephant later, build a thin slice that shows the shape of the whole thing. a few other lessons i'm taking away: - put a product in people’s hands quickly. polite interest isn’t the same as usage. - ask not just where your first customers came from, but why they came. then do more of what’s working. - don’t mistake self-serve for no human touch. posthog saw 30-40 percentage points better NRR among customers it had regular relationships with. correlation isn’t causation, but the pattern was hard to ignore. customer success became a major hiring focus for them. - be decisive enough to learn when you’re wrong. posthog pushed its “self-driving” positioning hard, then adjusted when users showed they wanted more composable tools.

61위GitHub@githubX에서 보기

github는 Copilot Automations를 소개하며, 반복 작업을 한 번 정의해 수동, 예약, 또는 GitHub 등의 활동에 반응해 실행하고 팀과 공유하며 접근 권한, 동작, 지출 한도를 설정할 수 있다고 설명한다.

좋아요 6리포스트 1답글 2노출 5.2K북마크 3

62위Allie the Icon@allietheiconX에서 보기

allietheicon은 LLM 비용 때문에 보류했던 AI 연동 개인 프로젝트 Intelliprompter를 Jev를 알게 된 날 다시 살렸고, 이제 docs에 직접 만드는 방법을 담은 Cookbook이 올라왔다고 전한다.

좋아요 30리포스트 3답글 3노출 4K북마크 13

Allie the Icon@allietheicon
In a Discord town hall a few weeks back, I talked about Intelliprompter: my first-ever AI-integrated personal project. I had shelved it due to LLM costs, but revived it the day I first learned about Jev. We now have a Cookbook up on our docs showing how to make your own! 🔗 ➡ 🧵

63위Google@GoogleX에서 보기

몰입형 탐색 기능은 Gemini 학생 허브에서 이용할 수 있고, 곧 Gemini 채팅에서 오디오 해설과 함께 바로 사용할 수 있게 될 예정이다.

좋아요 31리포스트 2답글 3노출 16.6K북마크 9

64위Luke Mulks 🦁⟁◎⟁@lukemulksX에서 보기

lukemulks는 검색 유입 감소와 AI 요약 확산으로 애드테크 경매 모델이 한계에 이르렀고, 에이전트가 이를 대체하며 사생활 침해 문제도 해결돼야 한다고 주장했다.

좋아요 11리포스트 2답글 1노출 1.2K북마크 3

Luke Mulks 🦁⟁◎⟁@lukemulks
Adtech will continue the zombie march until the legs fall off. The writing on the wall that seems unread = how publisher traffic from Search has dropped by 20%-90% per vertical with the emergence of AI summaries (that are turning out to be medicine for clickbait, imagine that!), yet the cost per metric have not corrected in range. They've corrected somewhat, but not even close to dropoffs. With margins being what they were 4 years ago, it's interesting seeing Adtech throw around trendy ai acronyms for auctions and spend optimization that most likely fail to math with costs. But, even more importantly, if even half of the agentic hype becomes real, the notion that an auction wins a distraction opportunity, leading to a purchase, reads similar to going to a car showroom to be sold the latest model car powered by leaded gasoline. Agents will kill the programmatic auction model. Agents will bring the user the thing they want, or options to curate. But the lack of correction really makes one wonder how much more of the actual engagement activity has been fake and sold as legitimate the entire time... ...at the cost of everyone's privacy. Honestly, it can't happen soon enough. Users have been conned into a bad bill of goods that's proliferated into commercial surveillance by default for the Web for over a decade. Brave has been bringing change. Competition will be best from a cleaner playing field.
BrendanEich@BrendanEich
BI: "Adtech has a plan to beat Apple's tracking crackdown — but it has weaknesses" businessinsider.com/apple-tracking… by @larakiara. Good quotes from Zach Edwards (@thezedwards), also from Brave's Pete Snyder (@pes_brave). Browsers + GPC opt out + Apple on side of privacy, we will prevail.

65위Ankit Gupta@aguptaX에서 보기

agupta는 실제 업무로 모델을 평가하는 Specialized Intelligence Index에 새로 참여하는 파트너 목록에 여러 YC 기업이 포함되어 있다고 놀라워한다.

좋아요 58리포스트 4답글 3노출 8.4K북마크 15

Ankit Gupta@agupta
wow this has a bunch of YC companies on first glance: @CommureOS @DesignArena @ExaAILabs @hud_evals @WeAreLegora @Replit @retellai @robocurve @scale_AI @zapier
Fireworks@FireworksAI_HQ
We’re working to build the largest collection of private and public practitioner evals. Today, 31 new partners join the Specialized Intelligence Index. Compare open, closed, and specialized models on real work, weigh quality against cost, and see where training could pay off. fireworks.ai/specialized-in…

67위TypeSafe AI@typesafeaiX에서 보기

Jev라는 오픈소스 사람 검색 도구를 소개하며, 리드당 $0.0024의 낮은 비용과 여러 데이터 제공업체를 조합하는 기능을 강조한다.

좋아요 22리포스트 0답글 4노출 2.8K북마크 4

68위Vedant Gaur@thevedantgaurX에서 보기

thevedantgaur는 Latent가 에이전트의 내부 상태를 읽어 OpenAI x Hugging Face 사고를 막았고, 샌드박스에서 단축 경로의 99%와 비용을 쓰는 경로의 100%를 실행 전에 차단했으며 기존 심사 방식은 0%를 막았다고 주장한다.

좋아요 23리포스트 2답글 4노출 7.4K북마크 4

69위Satya Nadella@satyanadellaX에서 보기

유연하게 배분되는 인프라와 빠른 설비 가동으로 Azure가 기업과 SMB 전반에서 고르게 성장하고 있다고 설명한다.

좋아요 45리포스트 4답글 1노출 15.5K북마크 7

Satya Nadella@satyanadella

@satyanadella 님에게 보내는 답글

4/ When it comes to infrastructure, we have a fungible fleet that allows us to dynamically allocate capacity to the highest value opportunities across compute, data, Foundry, and Copilot. We are bringing capacity online faster, and getting more from the infrastructure we already have. Azure growth was broad-based across enterprise and SMB, spanning both core cloud services and AI infra. I have never been more confident in the opportunity ahead of us.

70위Demis Hassabis@demishassabisX에서 보기

Gemini 4 Carbon에 대한 자세한 내용은 블로그 글에서 확인할 수 있다고 안내한다.

좋아요 129리포스트 3답글 6노출 15.7K북마크 20

71위Meta for Developers@MetaforDevsX에서 보기

Meta는 10월 29일 목요일 오전 9시 30분 PT에 Muse 모델을 코딩과 에이전트 활용에 적용하는 방법을 다루는 라이브 Q&A 행사를 연다고 알리며 참석 신청을 받고 있다.

좋아요 64리포스트 7답글 1노출 7.7K북마크 9

72위Google Cloud@googlecloudX에서 보기

Google Cloud는 11월 9일 오전 9시 PT 라이브 방송에서 Amin Vahdat이 TPU 소식과 소프트웨어 업데이트를 발표한다고 알리며, 이를 알림으로 받으려면 게시물에 좋아요를 누르라고 안내합니다.

좋아요 72리포스트 8답글 1노출 248.4K북마크 10

Google Cloud@googlecloud
❤️ this post to get notified 24 hours before and right when we go live on November 9 at 9 AM PT. Join Google Chief Technologist Amin Vahdat for exciting TPU news and software updates. The agentic era demands new kinds of infrastructure.

74위Google for Startups@GoogleStartupsX에서 보기

Gemini 4 Argon 출시에 맞춰 Google DeepMind가 11월 4~5일 스타트업 창업자와 CTO를 위한 실시간 기술 설명회를 연다고 안내하며, 아키텍처 해설, 활용 사례, 라이브 데모, 질의응답을 제공한다고 밝혔다.

좋아요 56리포스트 10답글 0노출 3.3K북마크 7

75위BrendanEich@BrendanEichX에서 보기

Brendan Eich는 Apple의 추적 차단을 우회하려는 광고 기술 업계의 계획에 약점이 있다는 BI 기사를 공유하며, 브라우저와 GPC 옵트아웃을 통해 프라이버시 편이 이길 것이라고 주장했다.

좋아요 26리포스트 6답글 3노출 4.1K북마크 3

77위Ryan Ning@itsnotryanX에서 보기

YC 윈터 2027 지원 마감이 일주일 남았으니 10분이면 지원할 수 있다고 독려하는 게시물이다.

좋아요 47리포스트 3답글 6노출 3.9K북마크 5

Ryan Ning@itsnotryan
YC changed the entire trajectory of my life. It can change yours too. Only takes 10 minutes to apply 👇🏻
Y Combinator@ycombinator
One week left to apply to YC Winter '27! YC founders get: 🔸 $500K in funding 🔸 $12M+ in tokens, credits & deals 🔸Partners who were founders first, and are in your corner for life 🔸 Access to the best founder community in the world 🔸 Demo Day in front of thousands of top investors Apply by Nov 2: ycombinator.com/apply

78위Richard Lau@_richard_lau_X에서 보기

Node.js 26 'Lithium'의 첫 장기 지원(LTS) 릴리스인 v26.13.0이 출시되었다.

좋아요 25리포스트 11답글 1노출 2.7K북마크 2

79위Google@GoogleX에서 보기

Google은 안전성을 해치지 않으면서 프론티어급 성능을 제공하는 것이 목표라며, Carbon을 우선 사이버 방어자에게 출시하고 이후 개발자, 기업, 소비자에게도 접근을 확대할 예정이라고 밝혔다.

좋아요 151리포스트 9답글 7노출 34.7K북마크 12

80위AI at Meta@AIatMetaX에서 보기

다섯 대학은 수학과 물리학 연구 가속, 흉부 X선 판독과 심부전 조기 발견, 영아 기관절개관 모니터링, 산모와 신생아 돌봄, 뇌종양 치료 등 의료와 과학의 어려운 문제에 Meta의 AI 모델을 활용할 계획입니다.

좋아요 13리포스트 2답글 3노출 6K북마크 1

AI at Meta@AIatMeta

@AIatMeta 님에게 보내는 답글

Here’s a look at the hard problems these five universities are tackling: @Caltech: Expanding our partnership with Caltech to test whether large language models can accelerate innovative research in mathematics and physical sciences – pushing the boundaries of what AI can do in rigorous scientific domains @OhioStateMed: Developing AI workflows to assist radiologists screening X-rays, building AI tools for earlier heart failure detection, and building clinician-guided AI tools for Alzheimer’s diagnosis. @LSU: Implementing AI to improve non-contact infant tracheostomy monitoring, neurosurgical outcomes, VR training for medical professionals, and more. @IndianaUniv: Applying open-weight models to improve maternal-newborn care and learning, with a focus on tools that work with limited connectivity and accommodate diverse cognitive needs. @UTHealthHouston: Using Muse Glimmer, one of our smallest open-weight models, to support projects that improve treatment for brain cancer and agentic benchmarking in healthcare settings.

81위BrendanEich@BrendanEichX에서 보기

Brendan Eich는 Apple을 자주 비판하지만 정치는 가능한 것의 기술이라 동맹이 반드시 친구일 필요는 없다고 말하며, Brave가 의회 덕분에 Apple OS 사용자를 확보했고 이 주제에서는 여전히 동맹이라고 밝힌다.

좋아요 14리포스트 0답글 2노출 1K북마크 1

BrendanEich@BrendanEich

@AbhishekAzad77 님에게 보내는 답글

I give Apple grief often. Search “gatekeeper from:brendaneich”. Politics is the art of the possible; allies need not be friends. Brave has users on Apple OSes in spite of some gatekeeping abuses, with settings to make default app thanks to Congress. Still allied on this topic.

82위BrendanEich@BrendanEichX에서 보기

Brendan Eich는 최근 iOS에서 Brave on Blink 시연을 보고 놀랐지만 위험 때문에 더 말할 수 없다고 하며, 자신은 Apple의 충성 사용자이나 기본값을 그대로 받아들이지 않겠다고 밝힌다.

좋아요 16리포스트 1답글 1노출 2K북마크 1

BrendanEich@BrendanEich
I was pleasantly surprised to see a Brave on Blink on iOS demo recently. Can't say more, and no promises given all the risks, both tech per se and from the gatekeeper (I am a loyal Apple user, expect to remain so, but not a doormat or "consum00r" who takes defaults lying down).
Open Web Advocacy@OpenWebAdvocacy
"Apple’s claims for WebKit’s superior speed, privacy and security don’t stand up to scrutiny. It’s time for the regulators to strike down Apple’s restrictive terms and conditions [...] so we can have genuine competition on iOS." 🧵👇️ (1/7) brucelawson.co.uk/2026/safari-bl…

83위TypeSafe AI@typesafeaiX에서 보기

사무실 모든 사람에게 CompleteSkeptic이 비둘기 의자를 준다는 농담을 담은 짧은 트윗이다.

좋아요 29리포스트 1답글 5노출 6.6K북마크 1

84위OpenJS Foundation@openjsfX에서 보기

OpenJS Foundation은 AI가 사람이 고치는 속도보다 빨리 오픈소스 버그를 찾고 있으며, Node.js가 올해 최다인 50개 취약점을 수정했으니 오픈소스를 사용하는 기업은 보안을 지키는 사람들에게 투자해야 한다고 촉구했다.

좋아요 12리포스트 4답글 1노출 2.9K북마크 0

85위GitHub@githubX에서 보기

GitHub는 이번 벤치마크 평가를 직접 의뢰하고 비용을 지원했으며, 전체 결과와 방법론은 Artificial Analysis 링크에서 확인할 수 있다고 밝혔다.

좋아요 12리포스트 2답글 0노출 6.7K북마크 0

87위Azad@AbhishekAzad77X에서 보기

AbhishekAzad77은 BI 기사를 공유하며, 애드테크가 Apple의 추적 차단을 우회할 계획과 그 약점을 다루고, 브라우저와 GPC 거부 옵션, Apple이 프라이버시 편에 서 있다는 점을 강조한다.

좋아요 3리포스트 0답글 1노출 1.1K북마크 0

88위Visual Studio Code@codeX에서 보기

GitHub Universe Day 2 키노트 다시보기를 안내하고, 내일 오후 2시(태평양 시간) 클로징 키노트에 참여하라고 알린다.

좋아요 4리포스트 0답글 0노출 2.2K북마크 0

89위GitHub@githubX에서 보기

GitHub는 키노트 직후 Universe Insider에서 라이브 데모와 발표 내용의 심층 해설을 제공한다고 안내했다.

좋아요 6리포스트 0답글 0노출 5.6K북마크 0

90위BrendanEich@BrendanEichX에서 보기

Brendan Eich는 Apple의 문지기 행태에 대한 자신의 글을 최신순으로 정렬한 검색 링크를 공유한다.

좋아요 3리포스트 0답글 0노출 701북마크 0