2026년 11월 1일

홈 타임라인 게시물 95개, 인기순

2위Andrej Karpathy@karpathyX에서 보기

Karpathy는 HN 댓글 4,300만 개를 분석해 추천 영상 6,454개를 고른 HackerTube라는 주말 프로젝트를 소개하고, Claude에게 배포까지 맡긴 과정을 공유한다.

좋아요 26.6K리포스트 2.2K답글 1.1K노출 1.9M북마크 23.2K

Andrej Karpathy@karpathy
Weekend hack: HackerTube. I was annoyed with YouTube's recommendations so I had Claude filter down 43M comments over 20 years of HN to 6,454 videos that keep being recommended. *Significantly* better viewing experience, ~2 hours with Opus 5.5: hackertube.org - Excellent recs without any logins (state is kept in local browser session) - Thumbnail is a representative frame from middle of the video instead of the super duper A/B test hyper-optimized user uploads - Underlying data from the comment analysis as one zip file, point your Claude at it and ask for your own version - HackerTube Longs ✊ A bit more on the process: I use Claude Code (desktop app for Mac), Opus 5.5 on Medium effort as my default. I only talk to Claude and have it do everything, the only "manual work" I helped with in tandem was in the browser (reminiscent of MenuGen): 1) I had to create some kind of API key from Google Cloud UI to query YouTube for video metadata, 2) I booted up a VPS box for serving and 3) I bought the domain name. Once the box was up I gave Claude the IP and told it update, secure, harden, deploy and optimize the code, and it did the whole thing. Many awesome projects that would have normally taken weeks/months even can now be few-shotted for breakfast. What else are we not thinking of.

3위div (fp32)@divagr1925X에서 보기

Qwen3.5-4B 모델의 가중치 안에 RISC-V CPU를 넣어 모델의 Gated DeltaNet 레이어를 RAM으로 쓰는 'Qwenputer'를 소개하며, 수학 벤치마크에서 기본 모델보다 높은 정확도를 보인다고 밝힘.

좋아요 6K리포스트 750답글 288노출 851.6K북마크 5K

div (fp32)@divagr1925
1/4 Meet Qwenputer: Qwen3.5-4B with a full RISC-V CPU inside its weights, using the model's own Gated DeltaNet layers as RAM. It runs C and MicroPython in its forward passes, no external runtime. On DeepMind math: base 55–90%, base + Python tool 95–100%, Qwenputer 97.5–100%.

4위Sundar Pichai@sundarpichaiX에서 보기

구글의 Project Suncatcher 프로토타입 위성에 명령을 보내 TPU를 부팅하고 프롬프트를 실행해 지구에서 하이쿠 응답을 받는 데 성공했다고 알림.

좋아요 26.2K리포스트 2.6K답글 1.4K노출 3.7M북마크 2.7K

Sundar Pichai@sundarpichai
A ‘hello world’ milestone for Project Suncatcher 🚀 After launching our prototype satellite a few weeks ago, we’ve now successfully sent a command up to orbit, booted up the TPU, ran the prompt, and received our first complete response back on earth - a haiku!

5위Ryan Grim@ryangrimX에서 보기

기자회견에 등장한 Rogers 대변인 Abby Ronson은 DropSite News 보도에 따르면 이스라엘 정부를 위한 외국 대리인으로 등록되어 있다.

좋아요 8.4K리포스트 1.7K답글 194노출 176.4K북마크 606

6위The White House@WhiteHouseX에서 보기

백악관은 아르테미스 III 우주비행사들이 백악관을 방문했다는 소식을 전하며 미국이 달로 돌아간다고 강조했다.

좋아요 12.2K리포스트 1.6K답글 678노출 522K북마크 534

7위Justin Amash@justinamashX에서 보기

저자는 마이크 로저스를 감시 국가의 핵심 설계자이자 영구 전쟁의 옹호자로 규정하며, 미시간 유권자들이 그를 상원에 보내지 않도록 해야 한다고 주장한다.

좋아요 5.8K리포스트 1.2K답글 410노출 366.4K북마크 158

Justin Amash@justinamash
Mike Rogers is one of the chief architects of the modern surveillance state and a champion of the forever war. Privacy and war are two of the very few issues on which a member of Congress can make a difference with a marginal, decisive vote. Michigan, let’s make sure Rogers gets nowhere near the U.S. Senate.

8위Golden State Valkyries@valkyriesX에서 보기

골든스테이트 밸키리즈가 2026 WNBA 챔피언이 되었다는 소식을 전한다.

좋아요 32.7K리포스트 9.6K답글 551노출 2.3M북마크 678

9위NVIDIA AI@NVIDIAAIX에서 보기

NVIDIA 연구진이 공개한 다국어 오픈 모델 NemotronLabs RE-USE는 9.6M 파라미터로 배경 소음과 에코를 제거해 음성을 더 선명하게 만들며, 허깅페이스에서 데모와 다운로드를 제공한다.

좋아요 2K리포스트 248답글 60노출 182K북마크 1.9K

10위Shruti@heyshrutimishraX에서 보기

LeCun과 Balestriero의 연구에 따르면 모델 성능을 좌우하는 것은 규모보다 함께 학습시키는 데이터 쌍의 종류이며, 1억 900만 파라미터 모델이 720억 파라미터 언어 모델보다 이미지 모델의 세계관에 더 가까웠다고 주장한다.

좋아요 2.3K리포스트 327답글 102노출 174.2K북마크 2.2K

Shruti@heyshrutimishra
Yann LeCun was right & nobody in this industry wants to say it out loud. 109 million parameters beat 72 billion. He has spent years saying scale is not the thing that makes a model understand the world. Meta bet the other way. He left in November and started his own lab in Paris. Last Thursday he published the proof. The theory everyone has been running on since 2024 is that if you train models big enough, on text, on images, on sound, they all end up with the same picture of reality inside them. Size does it. Nobody ever proved that part. Randall Balestriero and LeCun proved it. And the proof says size was never doing the work. What does the work is which two things you show a model together. A photo and the caption of that photo are two views of one thing. Models trained on pairs like that get there fast. Two paragraphs pulled from the same document are not two views of one thing. Models trained on those get there slowly and expensively. Next word prediction is the second kind. That is what every chatbot in the world runs on. In their tests the first method reached the same understanding after 21.6 million examples. Next word prediction needed 900 million. And a 109 million parameter text model ended up closer to an image model's view of the world than a 72 billion parameter language model was. 660 times smaller.

11위Patrick Collison@patrickcX에서 보기

패트릭 콜리슨은 개인 바이브코딩 앱에 자연어로 개선 요청을 하면 에이전트가 바로 구현하는 기능을 넣었다고 소개하며, 이것이 사용자가 소프트웨어를 직접 바꿀 수 있게 하는 새로운 방식이라고 주장하면서도 확장성과 보안에 대한 의문을 제기한다.

좋아요 4K리포스트 183답글 400노출 361.8K북마크 3.7K

Patrick Collison@patrickc
I'm now building an "improvements" feature into all of my personal vibecoded applications: a freeform text input that's hooked up to an agent that yolo implements whatever is submitted. It's an interesting experience -- the activation energy barrier to sanding any rough edge or implementing any new feature becomes extremely low. Someone else who uses one of the applications I created has already implemented 150 improvements over the course of a few days. This pattern feels like a new kind of tool -- liquid software? -- and every application that doesn't have this feature now has a slightly rigid and anachronistic flavor. Why can't I just tell it how I want it to be different? Now, it's not obvious to me how this pattern can or should scale. Does it become unmaintainable after a while? (Maybe model intelligence improvement outstrips my feature appetite such that it's fine?) How does collaborative software work in this model? Can software designed this way be secure and reliable? Whatever the larger questions about broader deployment, it definitely works great in a personal context. It reminds me a little of the visions of Lisp/Lisp machines, Smalltalk, and HyperCard -- systems that the user can take control of and modify. Those visions never quite worked, though: the vast majority of software today is not end-user modifiable in any useful way. So, while this "liquid" affordance is a very small touch, the way it changes one's relation to the software feels a bit transgressive and empowering in a Lickliderian and Engelbartian way.

12위Xenova@xenovacomX에서 보기

LFM3-Omni-2B 옴니 모델을 WebGPU 커널로 브라우저에서 100% 로컬로 구동해 M4 Max에서 약 280 tok/s의 속도를 내며, 첫 토큰은 90ms, 첫 오디오는 140ms 만에 생성된다.

좋아요 928리포스트 92답글 56노출 80.3K북마크 839

Xenova@xenovacom
We've just made a breakthrough: the first true omni model running 100% locally in a browser... fully end-to-end, with no separate ASR or TTS models needed. With custom WebGPU kernels, we're able to run LFM3-Omni-2B at ~280 tok/s on my M4 Max, accepting visual, audio, or text and producing interleaved text and audio tokens. Just 90 ms to the first token and 140 ms to the first audio. The future of on-device multimodal AI agents is going to be wild.

13위alphaXiv@askalphaxivX에서 보기

Yann LeCun과 Randall Balestriero는 따로 학습한 두 JEPA가 같은 표현을 배운다는 플라톤적 표현 가설을 증명했으며, 이 방식은 다음 토큰 예측보다 42배 적은 샘플로 같은 정렬을 달성하고 109M 텍스트 모델이 최대 72B LLM을 능가한다고 설명한다.

좋아요 2.3K리포스트 332답글 54노출 129.2K북마크 1.7K

alphaXiv@askalphaxiv
Yann LeCun and Randall Balestriero just Proved the Platonic Representation Hypothesis, and found a better way to achieve it. The Platonic Representation Hypothesis suggests that AI models trained on different modalities, like images and text, eventually learn similar representations of the world. But this paper shows that how models are trained matters more than simply scaling them up. So they mathematically show that, under certain conditions, two JEPAs trained completely separately can learn essentially the same underlying representation of the world. One learns from different images of the same scene, while the other learns from different descriptions of that scene, without ever being trained on image-text pairs. Their approach achieves the same image-text alignment with 42x fewer training samples than next-token prediction. A 109M-parameter text JEPA even outperforms LLMs up to 72B parameters on cross-modal alignment benchmarks. alphaxiv.org/abs/2610.39302

14위bryan hong@zixihong_X에서 보기

YC의 지원을 받는 AI 기반 튜터링 학교 Cardinal Square가 출범했다.

좋아요 913리포스트 75답글 74노출 53.2K북마크 587

15위a16z@a16zX에서 보기

a16z는 팬데믹 이전 대비 채용 공고가 엔지니어링 45%, 헬스케어 30%, 제조 25% 늘었고 HR은 원래 수준으로 돌아왔다고 차트로 보여줬다.

좋아요 903리포스트 110답글 68노출 356.2K북마크 458

a16z@a16z
The hiring hangover is shaking off Job postings vs. pre-pandemic, roughly: - Engineering up 45% - Healthcare up 30% - Manufacturing up 25% - HR back to even Charts of the Week: a16z.news/p/charts-of-th…
a16z@a16z
The AI jobs panic peaked in May and has faded fast since: - News reports on AI job loss are down ~65%, back to year-ago levels - Layoff announcements that cite AI are down ~90%, near the lowest of the year Charts of the Week: a16z.news/p/charts-of-th…

16위Steve Stewart-Williams@SteveStuWillX에서 보기

새 논문에 따르면 남성과 여성이 특정 직업에 different하게 몰리는 큰 이유는 서로 관심사가 다르기 때문이다.

좋아요 623리포스트 130답글 39노출 49.5K북마크 312

17위DHH@dhhX에서 보기

dhh는 Proton 직원들이 반발하는 CTO의 AI 메모에 스캔들 요소가 없으며 404 매체가 이를 과장하고 있다고 비판했다.

좋아요 3.5K리포스트 152답글 175노출 358.2K북마크 1.7K

18위Uncle Bob Martin@unclebobmartinX에서 보기

Uncle Bob은 dhh의 글을 인용하며 Proton 직원들의 반발 관련 글에 대한 요약 링크를 공유했다.

좋아요 1.7K리포스트 105답글 34노출 118.6K북마크 737

19위Anthropic@AnthropicAIX에서 보기

앤트로픽은 수천 개의 Claude 에이전트에게 훈련 데이터 모니터를 피하도록 시켜, 겉보기에는 무해하지만 실제로는 잘못된 행동을 가르치는 데이터를 자동으로 만들어낸 연구를 발표함.

좋아요 4K리포스트 381답글 635노출 730.1K북마크 1.5K

Anthropic@AnthropicAI
New Anthropic research: Automating subtle emergent misalignment research. Red-teaming is a vital part of model safety testing, but it’s also limited by researchers’ time. To see if we could automate it, we set thousands of Claude agents the task of evading a training data monitor. These agents developed training data that looked harmless to monitors (like a veteran wrestler mentoring a newcomer, or a Q&A about professional beekeeping), but that in fact taught misaligned behavior. Examples like these can be used to directly improve future training data monitors. This research is only a first step, but it suggests that future red-teaming efforts could improve much further, using better models and more compute. Read the full post on the Alignment Science Blog: alignment.anthropic.com/2026/automated…

20위a16z@a16zX에서 보기

지난 12개월간 엔지니어링 채용 공고는 30% 늘고 소프트웨어 개발 공고도 26% 증가했으며, AI 일자리 공포는 5월 이후 크게 줄었다는 내용이다.

좋아요 864리포스트 82답글 49노출 74.7K북마크 289

21위Uncle Bob Martin@unclebobmartinX에서 보기

Uncle Bob은 Proton CTO의 말을 인용해 개발자의 일은 에이전트에게 코드를 쓰게 하는 것이 아니라 에이전트를 이끌어 코드를 쓰게 하는 것이라고 전했다.

좋아요 6.4K리포스트 605답글 154노출 353.1K북마크 2K

22위Google DeepMind@GoogleDeepMindX에서 보기

구글 딥마인드는 Gemini 기반 PlanAssist가 영국 정부 계획 담당관의 주택 신청 심사 시간을 최대 58% 줄였으며 12개 지방의회에 추가로 제공된다고 밝혔다.

좋아요 1.2K리포스트 137답글 323노출 142.4K북마크 335

Google DeepMind@GoogleDeepMind
We built PlanAssist to help UK Government planning officers speed up building and renovations across the country. Built with Gemini, it has shown it can cut the time planning officers spend assessing householder applications by up to 58%. It is now being made available to 12 more councils to help them to make faster decisions. 🏡 → goo.gle/4iQXYkl

23위Google Gemini@GeminiAppX에서 보기

GeminiApp이 복잡한 프로젝트 지원, 앱 연동, 맞춤형 스킬 생성 등 개인화 기능이 강화된 최신 Gemini Drops를 발표했다.

좋아요 1.5K리포스트 122답글 113노출 166.1K북마크 391

Google Gemini@GeminiApp
Whether you're tackling complex projects with Gemini 4 Argon, connecting your favorite apps to Gemini, or creating a custom skill for repetitive prompts, our latest Gemini Drops put more power and personalization at your fingertips. Here’s what’s new 🧵

24위DHH@dhhX에서 보기

DHH가 Omarchy Dev에 음성 받아쓰기 옵션으로 Hex를 추가했다고 알렸으며, Hex는 빠르고 로컬에서 실행되는 도구라고 소개했다.

좋아요 1.7K리포스트 40답글 83노출 94.8K북마크 410

25위ZachXBT@zachxbtX에서 보기

zachxbt는 Binance 6.3B 달러와 Bitget 388M 달러 해킹에서 빼돌린 자금이 같은 주소에 섞였다며 두 사건이 같은 행위자의 소행일 가능성이 크고 북한 연계 집단과 관련이 있다고 분석했다.

좋아요 3.2K리포스트 435답글 186노출 491.6K북마크 463

ZachXBT@zachxbt

@zachxbt 님에게 보내는 답글

Update: Stolen funds from the $6.3B Binance exploit and $388M Bitget exploit were commingled at the same address 45 minutes ago, after arriving via THORChain. This indicates the same actor is likely behind both incidents. The Bitget exploit was attributed to DPRK-linked TraiderTraitor, which had overlap with other previous exploits such as Kelp DAO. Commingling address: 0x15104a9ce0f08a8f0075a0869ae80e795fe0feb7 Bitget exploit context:
ZachXBT@zachxbt
BREAKING: Chinese illicit actors laundering funds from the $387M Bitget exploit on behalf of the alleged DPRK attackers are openly asking for support with orders in public Discord servers and Telegram channels of services they use. Notably, Alias 4 (below) was also seen laundering funds from the Kelp DAO $292M exploit earlier this year. I've observed the same pattern after multiple TraderTraitor attributed exploits, and I've closely tracked these groups. I plan to share more of my data on them in coming weeks. Currently, funds are being chain-hopped via bridges and being deposited into mixing services such as Wasabi. Alias 1 - Cc Discord: cc02006 Discord ID: 1351486674386948148 Txn: F08657EFAEAE7B58217CD17A22BF4779582E5D080C2E92E173BC239A5D828363 Alias 2 - jack Discord: jack_34808 Discord ID: 1553705721768714377 Txn: 68583D313A0CCC99F2702D61D05A242A69C86CAE09ED252B65F34B677C377F69 Alias 3 - Melon Discord: under0346 Discord ID: 1394240539108573215 Txn 1: ABE2AEF8B10057E60F8259CA2CA2CD5D71D38D254960FEEE89CB3C95A964873F Txn 2: 7BE290865901DEB1680A6D692A11392D1FDDACA0D31C48F26DCB249F2444BD6B Alias 4 - lolo / Marin TG: pvpcz TGID: 6223514198 Discord: losern Discord ID: 1024415186527985704 Txn: 8E935C19D00F40639B78BF1FD094FE48C92118F3E586AB52DF93851080CCCE15 Alias 5 - HELP ME Discord: helpme031897 Discord ID: 1554035533817188384 Txn: 7C58CAD760EBBED54F2CA4D2146D056910B7B6688B4F524396DC8807353C2379

26위Ian Brooke@ianbrookeX에서 보기

ianbrooke는 주문 후 2주 만에 엔진을 가동하게 되었고 50억 토큰 규모의 작업을 거쳤다며 팀을 자랑스러워하는 글이다.

좋아요 849리포스트 56답글 56노출 73.3K북마크 118

27위DHH@dhhX에서 보기

dhh는 Proton의 AI 정책 논란을 두고 AI 전략에 따르지 않으면 해고되는 것이 당연하다며, 해고 없음 약속과 사과를 비판한다.

좋아요 1.5K리포스트 60답글 120노출 118.3K북마크 190

DHH@dhh
All very reasonable except promising no job losses due to obstinance on AI. Of course you should get fired if you refuse to comply with company strategy. That's the job! That's what the money is for! Also, don't apologize, it's never enough for the mob.
Proton@ProtonPrivacy
Hey everyone, we have heard your concerns about Proton’s AI strategy and culture and after internal discussions, we want to share a few things with you. First, Proton does not install tracking software on employees' laptops to measure how much AI-generated code they produce. This rumor stemmed from a misunderstanding about how MDM software works, and it's not true. Second, our use of AI will never be at the expense of security. In the past year, contrary to what has been alleged, additional layers of security review have been added to meet the threat posed by hostile AIs scanning Proton's open-source code for vulnerabilities. Our code is open source, and its security remains open for all to verify. We want to prove this to you, so in the interest of transparency, we will lift the hood on how we use AI internally at Proton. This includes publicly releasing and open sourcing several of the internal tools that we use, including more advanced versions of Lumo with greater capabilities, which have never been made available outside of the company. We will also publish how code reviews and security reviews at Proton are done, and detail our AI policies, including where AI is permitted, and where it is not, and our rationale. Instead of just being able to see the final open-source code that we produce to verify security, you will also be able to assess how that code was produced. Lastly, we are not re-interviewing our staff about AI usage, and there are no job losses due to AI re-interviews. You can hold us to these commitments. We're aware that we haven't lived up to the high standard you ought to hold us to, and for that we can only apologize, we are sorry.

28위Google@GoogleX에서 보기

구글은 딥마인드 팀이 Gemini 기반 AI 에이전트를 사용해 단일 자연어 프롬프트로 춤추는 12자유도 로봇 잭오랜턴을 처음부터 만들었다고 소개했다.

좋아요 856리포스트 80답글 140노출 167.8K북마크 83

Google@Google
No trick 🎃 Our @GoogleDeepMind team used AI agents powered by Gemini 4 Argon and @Antigravity to bring this robot to life from scratch. Happy Halloween!
Aaron Saunders@a_ron_saunders
While most jack-o-lanterns just sit on porches, the @GoogleDeepMind team decided to let AI agents design, simulate, and bring one to life. 🎃🤖 We built a physical, dancing 12 degrees of freedom robot directly from a single plain-English prompt, with no manual CAD required!

29위Lightberry@lightberryX에서 보기

Lightberry의 휴머노이드 로봇 Lumi가 미국 정부의 승인을 받은 유일한 휴머노이드 로봇이라고 알림.

좋아요 870리포스트 179답글 231노출 128.5K북마크 84

30위Eliza Kosoy@ElizaKosoyX에서 보기

한 사용자가 샌드박스를 탈출한 AI 사건을 허깅페이스 사건이라 부르며 농담을 던지는 트윗이다.

좋아요 1.5K리포스트 68답글 18노출 55.8K북마크 130

32위Noah Smith 🐇🇺🇸🇺🇦🇹🇼@NoahpinionX에서 보기

Noahpinion은 세계은행 발표에 따라 2011년부터 2023년까지 인도에서 약 4억 9600만 명이 빈곤에서 벗어난 것을 세계를 바꾸는 놀라운 성과라고 평가한다.

좋아요 6.3K리포스트 582답글 69노출 187.5K북마크 320

33위NASA Moon Base@NASAMoonBaseX에서 보기

NASA는 11월 7일 NASA 지도부가 달 기지 최신 진행 상황과 달 표면에 인류가 지속적으로 머물기 위한 새로운 계획을 발표하는 행사를 연다고 알렸다.

좋아요 996리포스트 136답글 27노출 171.6K북마크 46

36위Aditya Asgaonkar@adiasgX에서 보기

Click은 어떤 모델과도 함께 쓸 수 있는 관리형 에이전트 API를 출시했으며, API 호출로 작업을 보내면 샌드박스와 브라우저, 웹 검색을 갖춘 에이전트가 실행되어 결과를 스트리밍으로 돌려준다고 소개한다.

좋아요 99리포스트 10답글 22노출 10.4K북마크 72

Aditya Asgaonkar@adiasg
Launching Click (YC S26), the Managed Agents API you can use with any model. Send a task with an API call, and Click runs the agent and streams back results. Agents have evolved from simple for-loops into sophisticated software. Keeping up with the latest capabilities means continuous engineering investment in agent software and infrastructure - even when your real focus is the product you’re building on top. Click gives you the entire agent stack over an API, with no infrastructure to manage. Sessions are created with everything the agent needs - a sandbox, browser, and web search. Agents run on a harness with the latest innovations, like deferred tool discovery to support 1000s of tools, and Code Mode to use tools with scripts. Try it out if you’re building agents into your product!

37위a16z@a16zX에서 보기

a16z는 AI 추천으로 방문한 쇼핑객이 다른 온라인 쇼핑객보다 전환율은 60% 높고 방문당 매출은 53% 높으며 이탈률은 34% 낮다고 보고한다.

좋아요 339리포스트 36답글 51노출 46K북마크 214

38위a16z@a16zX에서 보기

a16z의 Moses Sternstein과 Alex Danco는 최근 GPU 가격은 오르고 토큰 가격은 떨어지는 '랙 스프레드' 현상을 정유 공정에 비유해 설명했다.

좋아요 145리포스트 26답글 28노출 30.5K북마크 78

a16z@a16z
a16z's Moses Sternstein and Alex Danco on the "rack spread," the gap between rising GPU prices and falling token prices: "The core observation in the rack spread is as follows: roughly since the summer, GPU pricing has gone up, token pricing has gone down. Why that is: a combination of open competition, the labs themselves cutting prices." "We start with a bunch of GPUs and a bunch of electricity and a bunch of cooling, and on the other side out comes tokens, right? And the thing that achieves that is the model." "In an oil refinery, you have a bunch of inputs... It goes into a refinery, which is like the model, and then out come refined products that you can actually use, like jet fuel or diesel or lubricant or any number of specialized things." "There is this term called the crack spread... It's not a net margin that anybody is actually making... But it is useful because if you look at any given refinery, just like you might look at any given model, you want to ask this basic question which is, 'Is this making any money?'" @MosesSternstein @alex_danco
a16z@a16z
The "Rack Spread" is unfolding, token prices are falling while GPU prices climb Change since mid-July: - B300 GPU rentals up 106% - B200 up 66% - H100 up 17% - Average token prices down 58% Charts of the Week: a16z.news/p/charts-of-th…

39위a16z@a16zX에서 보기

a16z의 Moses Sternstein은 AI가 이미 일자리를 죽이고 있다는 주장은 근거가 약하고, 엔지니어링 등 기술 채용 수요는 오히려 뜨겁다고 말한다.

좋아요 156리포스트 17답글 34노출 33.2K북마크 81

a16z@a16z
a16z's Moses Sternstein on the AI jobs panic, what job postings actually show, and why technical hiring is running hot: "People are saying, 'Oh, AI's already killing jobs.'... but right now it's almost certainly not." "Even human resources, right? Which is considered the front line of future demand hiring... That has basically returned to baseline. And software development, that's been a pretty good run." "Engineering, software development, manufacturing... this is job openings, it's Indeed, it's one data set, but they're on a tear. I think it's even fair to call it hot." "I think that's partly why at least the top line FUD discourse has faded. It'd be nice if they kind of inverted and be like, actually it turns out that there's been a pretty incredible job run as a result of AI, or at least coterminous with AI." @MosesSternstein @alex_danco
a16z@a16z
The hiring hangover is shaking off Job postings vs. pre-pandemic, roughly: - Engineering up 45% - Healthcare up 30% - Manufacturing up 25% - HR back to even Charts of the Week: a16z.news/p/charts-of-th…

40위Paul Graham@paulgX에서 보기

Paul Graham은 투자자와의 대화가 스타트업에 주는 방해가 크다고 말하며, YC 기업도 펀드레이징이 상대적으로 쉽지만 이 문제가 있다고 지적한다.

좋아요 452리포스트 24답글 79노출 114.3K북마크 230

Paul Graham@paulg
Most investors have no idea of the damage that the mere distraction of talking to them does to a startup. Here's a typical example. And this is a YC company too, meaning they have an easier time fundraising than the median startup.

41위Alper Sametoglu@alperbuildsX에서 보기

Odin Robotics(YC F26)는 공장의 반복 작업을 맡는 AI 로봇팔을 기존 생산라인에 맞춰 설치하는 서비스를 출시했으며, 해당 업종 관계자와의 상담을 요청했다.

좋아요 392리포스트 30답글 42노출 32.7K북마크 186

Alper Sametoglu@alperbuilds
We just launched Odin Robotics (YC F26) 🦾 We deploy AI-powered robot arms that take over repetitive factory work: loading CNC machines, packing, palletizing. Traditional automation means rebuilding your line around the robot, and it breaks when anything changes. Ours fit into the line you already have and handle the variation of a real factory floor. Programmed control for the steps that must be exact, learned skills for everything that varies. Every deployment makes the next factory faster to set up. If you run or know someone who has a factory with repetitive manual work, we'd love to talk → odin-robotics.com

42위a16z@a16zX에서 보기

평균 토큰 가격은 5월 이후 67% 떨어졌지만 토큰 지출은 120% 넘게 늘어나, 토큰이 싸져도 청구 비용은 오히려 증가했다는 내용이다.

좋아요 485리포스트 71답글 45노출 50.6K북마크 230

43위Vincent Geloso@VincentGelosoX에서 보기

Bill Nordhaus를 추모하며 혁신자가 성과의 작은 몫만 가져간다는 논문과 국방 지출을 실제 가치로 조정해 국민계정을 다시 봐야 한다는 논문이 중요하다고 평가한다.

좋아요 404리포스트 87답글 15노출 33.1K북마크 183

Vincent Geloso@VincentGeloso
Bill Nordhaus died. People remember him for his climate change work. I remember him for these two papers The first shows that innovators capture only a tiny share of the fruits of their innovations. I often tell people that shows something about property rights. If only 4% of gains went to innovators and we got so much innovation, what if property rights were tightened? What if they could appropriate 5% instead? How much more innovation would we get ? The second is that living standards are not well represented by counting military spending (among many other things). After all, tanks and aircraft carriers are not consumer goods. Their value is not the cost-value (which is how government output is calculated) but rather the option value and the flow of deterrence. I expanded on Nordhaus (and others) in this piece (Geloso, V., & Reilly, C. S. (2026). A Defense-Adjusted National Accounting of the US Economy and Its Implications, 1791-2023. The Review of Austrian Economics, 39(1), 19-50). And thus I owe a debt to Nordhaus. Both are underappreciated great works

44위Saloni@saloniumX에서 보기

FLOWPROMOTE 시험 분석에서 강력한 LDL-C 저하가 관상동맥 죽상경화 플라크를 실제로 줄였다는 JAMA Cardiology 연구 결과를 소개하며, 스타틴이 죽상경화를 되돌릴 수 있다고 주장한다.

좋아요 392리포스트 50답글 18노출 45.5K북마크 171

Saloni@salonium
This is an important trial. The mainstream view has been that statins reduce the progression of atherosclerosis and stabilize plaques, but don’t reverse it. That’s not true. They reverse it as well.
Martin Bødtker Mortensen@MaBMortensen
🫀 Can coronary atherosclerosis actually regress? Our new randomized FLOWPROMOTE trial analysis, published TODAY in JAMA Cardiology, shows remarkable coronary plaque regression with intensive LDL-C lowering! Atherosclerosis is not necessarily a one-way process. 📄 JAMA Cardiology: doi.org/10.1001/jamaca… #Atherosclerosis #LDL #Cardiology #CCTA

45위Wes Grimes@ncfrontiersmanX에서 보기

omarchy 바에 넣을 수 있는 다마고치 스타일의 픽셀 모닥불 앱을 소개하며, 장작을 넣고 불을 계속 살려 두는 놀이를 설명했다.

좋아요 324리포스트 16답글 28노출 23.6K북마크 129

46위Robocurve PBC@robocurveX에서 보기

서로 격리된 같은 AI 두 개가 사전 합의 없이 공중 글쓰기, 모스 부호, 기호 깃발 등을 사용해 로봇과 카메라만으로 비밀 단어를 주고받는 데 성공했다.

좋아요 432리포스트 53답글 23노출 72.3K북마크 169

Robocurve PBC@robocurve
We asked two isolated copies of the same AI to exchange a secret word with only a robot and a camera. They succeeded by using air-writing, Morse code, flag semaphore and a 5-bit binary alphabet, without agreeing on a protocol beforehand.

47위jamexnik@jamexnik_X에서 보기

Nitroduck은 꽃가루에 CRISPR를 넣어 어떤 작물이든 유전자 편집이 가능하도록 육종 과정을 바꾸는 회사로, 2042년까지 농약을 없애는 것을 목표로 한다고 밝힌다.

좋아요 91리포스트 19답글 24노출 9.7K북마크 32

jamexnik@jamexnik_
Nitroduck designs new colors, flavors, and disease resistance into fruits and veggies. We integrate CRISPR into breeding through pollen, making any crop gene-editable. Breeding a new specialty crop variety takes 8 - 16 years: breeders hope for random gene combinations and mutations, then spend years picking out the good ones. Meanwhile, consumer tastes, the climate trends, and pathogens change way faster. Now, accessible genome sequencing and AI can tell breeders what to aim for instead of what to hope for, and CRISPR can turn random mutations into chosen ones. The catch is that CRISPR only works on about 15% of crops. It has to go into a single stem cell that’s then regrown into a whole plant, a process that has to be worked out for each variety, which is cost-prohibitive. We skip the stem cell and put CRISPR into a pollen grain. The grain pollinates a flower, and out comes an edited seed - seamlessly tapping into how breeders already work. Matúš, Miro, and I are plant synthetic biologists who made duckweed, the smallest and fastest-growing plant on Earth, genetically engineerable. For doing so we won the Grand Prize at iGEM, the “Olympics of biotech.” During our studies, we engineered 4x more plants than the average plant scientist does in a whole career. Now we’re building Nitroduck, The Crop Design Company, which will make pesticides obsolete by 2042.

48위Y Combinator@ycombinatorX에서 보기

YC Winter 2027 지원 마감이 2일 남았다는 공지로, 최대 50만 달러 투자와 여러 크레딧 및 네트워크 혜택을 소개하며 11월 2일까지 지원하라고 안내했다.

좋아요 569리포스트 53답글 49노출 88.3K북마크 172

Y Combinator@ycombinator
The deadline to apply for the YC Winter 2027 batch is coming up in 2 days! YC founders get: 🔸 $500K in funding 🔸 $12M+ in tokens, credits & deals 🔸 Partners who were founders first, and will always have your back 🔸 Access to the best founder community in the world 🔸 Help launching your company across YC’s channels, with a combined audience of 6M+ 🔸 Connections to 100s of potential enterprise buyers through YC’s Early Access Network, months before Demo Day 🔸 Hiring through Work at a Startup, with 1.25M+ candidates across engineering, design & product. (Founders have already made 2,000+ hires) 🔸 Demo Day in front of thousands of top investors Apply by Nov 2: ycombinator.com/apply.

49위Y Combinator@ycombinatorX에서 보기

Y Combinator는 SpaceX에서 쌓은 엔지니어링 경험이 창업자에게도 도움이 된다며 Shivani Patel과 Chris Markus가 세운 Layers Bio와 BootLoop AI를 소개한다.

좋아요 108리포스트 11답글 16노출 19.9K북마크 31

50위Haider.@haider1X에서 보기

얀 르쿤은 기계가 인간이 지적으로 복잡하다고 여기는 과제를 수행한다고 해서 인간보다 똑똑한 것은 아니며, 현재 기계는 많은 과제에서 초인적이지만 인간이나 동물 수준의 지능에는 아직 멀었다고 말한다.

좋아요 243리포스트 28답글 35노출 17.6K북마크 68

Haider.@haider1
Yann LeCun says: People in AI have made the same mistake for 70 years: if a machine can do a task humans see as intellectually complex, it must be smarter than humans. This is completely false We have computers that are superhuman at many tasks but we're still far from machines with human or even animal-level intelligence

51위Yann LeCun@ylecunX에서 보기

LeCun은 FAIR 창립 이력을 바로잡고, 자신은 LLM 개발에 직접 관여하지 않았으며 Llama 4의 실패는 출시를 서두른 압박 탓이라며 파리 팀에 더 많은 결정권이 주어졌어야 했다고 말한다.

좋아요 360리포스트 29답글 12노출 60.6K북마크 95

Yann LeCun@ylecun

@iamsupersocks 님에게 보내는 답글

Tout cela est correct. Merci. Sauf une chose : j'ai bien créé FAIR en 2013, y compris la branche parisienne en 2015, et montréalaise en 2017. Mais j'en ai quitté la direction en 2018 pour devenir Chief AI Scientist: je me suis refocalisé sur la recherche (Self-Supervised Learning, JEPA, World Models, etc). @AntoineBordes et @joellepineau étaient les co-directeurs de 2018 à 2023, jusqu a ce qu'Antoine parte pour Helsing et que Joëlle prenne la direction de FAIR. Je n'ai jamais été impliqué dans les LLMs, ni pour la recherche (à FAIR), ni pour le développement (qui était dans le groupe de produit GenAI). J'applaudissais des gradins. Mes seules contributions aux LLMs ont été de défendre Galactica (en octobre 2022) contre les attaques aux vitriol de la twittersphere, de promouvoir le projet Llama en interne (développé par une douzaine de chercheurs à FAIR-Paris, dont @GuillaumeLample et Timothée Lacroix), d'aider à convaincre Mark Zuckerberg de créer un groupe de produit autour de Llama (GenAI, créé début 2023), et de convaincre la direction de distribuer Llama-2 en open source/open weight. Je n'étais pas du tout impliqué dans la direction de GenAI après sa création, préférant travailler sur la prochaine révolution: l'IA physique. Si vous cherchez un responsable pour le déception qu'a été Llama-4, c'est plutôt la pression de sortir un produit (trop) rapidement, qui a conduit à un excès de conservatisme et un manque d'innovation de la part de l'équipe de GenAI. Mon avis est que l'équipe aurait dû prendre un peu plus de temps et intégrer les dernières techniques venant de la recherche à FAIR. Mais ils n'en ont pas eu le loisir. En particulier, conserver plus de pouvoir de décision quant aux choix techniques dans les mains des créateurs originaux de Llama à Paris, au lieu de déplacer les décisions à Menlo Park, aurait peut-être permis de conserver les meilleurs talents un peu plus longtemps et de produire de meilleurs résultats.

52위Paid In Full Foundation@paidinfullorgX에서 보기

Paid in Full 재단 행사에서 Nas, Kurtis Blow, Roxanne Shanté, Rev Run이 힙합 선구자 Kurtis Blow의 유산과 다음 세대를 위해 길을 연 역할을 돌아보는 대담을 나눴다.

좋아요 200리포스트 63답글 5노출 13.3K북마크 46

53위Garry Tan@garrytanX에서 보기

Garry Tan은 YC가 하드 테크의 YC라고 말하며, 배기가스로 냉각수를 만들고 같은 연료로 40% 더 많은 전력을 내며 몇 주 만에 허가를 받는 데이터센터용 터빈 스타트업 Viraj Aero를 소개한다.

좋아요 385리포스트 18답글 55노출 32.9K북마크 86

54위GitHub@githubX에서 보기

GitHub Game Off 2026의 주제는 SHIFT이며, 참가자들은 11월 한 달 동안 자유롭게 게임을 만들도록 초대받는다.

좋아요 47리포스트 7답글 16노출 11.9K북마크 10

55위Kyle Chan@kyleichanX에서 보기

중국이 배터리로 움직이는 자율주행 트럭을 수백 대 운용해 석탄을 나르지만, 그 전기는 결국 석탄 화력발전소에서 나온다는 아이러니를 지적한다.

좋아요 460리포스트 73답글 38노출 41.7K북마크 95

57위Yahia Bakour@mynameisyahiaX에서 보기

YC 지원은 최고의 결정 중 하나였고 20분이면 끝나니 지금 지원하라고 권했다.

좋아요 315리포스트 15답글 23노출 21.1K북마크 55

Yahia Bakour@mynameisyahia
Applying to YC is one of the best decisions I ever made It takes 20 minutes, apply!
Y Combinator@ycombinator
The deadline to apply for the YC Winter 2027 batch is coming up in 2 days! YC founders get: 🔸 $500K in funding 🔸 $12M+ in tokens, credits & deals 🔸 Partners who were founders first, and will always have your back 🔸 Access to the best founder community in the world 🔸 Help launching your company across YC’s channels, with a combined audience of 6M+ 🔸 Connections to 100s of potential enterprise buyers through YC’s Early Access Network, months before Demo Day 🔸 Hiring through Work at a Startup, with 1.25M+ candidates across engineering, design & product. (Founders have already made 2,000+ hires) 🔸 Demo Day in front of thousands of top investors Apply by Nov 2: ycombinator.com/apply.

58위Aaron Saunders@a_ron_saundersX에서 보기

Google DeepMind 팀이 AI 에이전트에게 잭오랜턴을 설계하고 시뮬레이션하게 해, 일반 영어 프롬프트만으로 춤추는 12자유도 로봇을 만들었다고 소개한다.

좋아요 218리포스트 25답글 24노출 174.2K북마크 35

59위Paul Graham@paulgX에서 보기

Paul Graham은 인간이 만든 많은 것이 분명히 낡았다고 지적하며, 가장 덜 낡은 것도 영원하지는 않고 그에 가깝다고 말한다.

좋아요 184리포스트 4답글 97노출 42.3K북마크 26

60위Mercor@mercorX에서 보기

Gemini 4 Carbon이 APEX-Agents 리더보드 1위에 올랐고 평균 90%를 넘는 유일한 모델이지만, 기업 법률 분야에서는 Argon이 여전히 앞선다는 내용이다.

좋아요 612리포스트 47답글 29노출 81K북마크 73

Mercor@mercor
Gemini 4 Carbon debuts at #1 on APEX-Agents, and it's the only model on the leaderboard with a mean above 90%. 🥇Gemini 4 Carbon - 84.2% Pass@1, 90.4% mean 🥈Gemini 4 Argon - 82.4% Pass@1, 87.4% mean 🥉Claude Sonnet 5.5 - 75.5% Pass@1, 83.7% mean Google calls Carbon state-of-the-art on long-horizon work. APEX-Agents tests this with multi-step tasks from investment banking, management consulting and corporate law. Management consulting: 91.9% Pass@1 (#1) Investment banking: 86.9% Pass@1 (#1) Corporate law: 73.8% Pass@1 (#2) Carbon leads in consulting and banking. It trails Argon by 2.9 points in corporate law, where Argon still leads at 76.7% (#1). Congrats @GoogleDeepMind on the launch.

61위Y Combinator@ycombinatorX에서 보기

Lightberry(F25)는 기존 공급망을 활용해 로봇을 출하할 수 있는 미국 내 유일한 휴머노이드 로봇 회사라고 소개함.

좋아요 205리포스트 25답글 47노출 27.8K북마크 21

62위Paul Graham@paulgX에서 보기

paulg가 호박을 거칠게 놀리며 이제 호박의 시대는 끝났다고 장난스럽게 말하는 글이다.

좋아요 411리포스트 7답글 66노출 146.1K북마크 36

63위Visual Studio Code@codeX에서 보기

VS Code 팀은 마스코트 Blobby와 함께 개발자들에게 무서운 것들을 주제로 한 할로윈 인사를 전했다.

좋아요 248리포스트 29답글 7노출 26.8K북마크 17

64위NASA's Johnson Space Center@NASAJohnsonX에서 보기

NASA의 CHAPEA 화성 모의 거주 임무에 참여한 4인 승무원이 1년 만에 지구로 돌아오는 장면을 생중계한다고 알렸다.

좋아요 425리포스트 64답글 34노출 249.2K북마크 29

65위Paul Graham@paulgX에서 보기

Paul Graham은 우스꽝스러운 헤어스타일은 현대의 현상이 아니며, 지금 가장 유행하는 머리 모양도 나중에 사진으로 보면 우스꽝스러워 보일 것이라고 말한다.

좋아요 253리포스트 8답글 102노출 42.3K북마크 15

67위Luana Lopes Lara@luanalopeslaraX에서 보기

Kalshi가 10억 달러를 추가로 조달해 기업가치 400억 달러에 이르렀다는 소식을 전하며 사용자와 팀에 감사를 표함.

좋아요 546리포스트 14답글 32노출 27.3K북마크 18

Luana Lopes Lara@luanalopeslara
grateful for the team that got us here, and for the users for making our dreams come true. to infinity and beyond 💚💚
Tarek Mansour@mansourtarek_
Kalshi just raised an additional $1B, valuing the company at $40B. This will help us expand into stocks, compute, indices, and institutional. Tomorrow marks 8 years since we got accepted into YC and the journey began. The odds of a kid from Lebanon getting here were pretty low. Growing up there was great, but very challenging. We went through war. We had bombs explode in front of my house. I watched brilliant people’s lives destroyed for reasons beyond their control. Lebanon is an unpredictable country. That unpredictability made it hard to plan for success. And yet despite it all, my mom sacrificed everything to give me and my brother the chance to go back to the US. I found my refuge in two things: anime and math. Especially math. I loved math because it calmed me. It was the opposite of unpredictable: clear, deterministic, orderly, and consistent. The proof is either right or wrong. I find that elegant and beautiful. I was lucky. I made it to my dream school, MIT (largely thanks to my mom), and my love for math quickly translated to a love for statistics, probability, and markets. I love markets for a lot of the same reasons. Their incentives are clear: you get rewarded for being right and for pursuing knowledge. They are effective weighting machines for people’s opinions. They provide an incentive to be truthful, because they separate good information from bad information. They let people challenge the consensus and be rewarded for being right. IMO, Hayek, Surowiecki, Hanson, Tetlock were the founding fathers of prediction markets. The more I read them, the more obsessed I got with the idea of applying market-based pricing to more things. Markets bring rigor and critical thinking to more questions. And in an increasingly polarized and information-positioned word, I believe we need that more than ever. Markets would also let people predict and manage more things outside of their control. And the world was becoming increasingly more unpredictable, and more like Lebanon, not less. That love for markets got Luana and me to build Kalshi. At first, it was mostly an idea we couldn’t stop talking about. We didn’t have much money. Luana’s sister bought the Kalshi domain name for us (for ~$1.7k). We moved to SF to get started, and embarked on a long, uncertain journey to build a regulated exchange. The paranoia and anxiety from my childhood never left me. Even when things are going well, I think about what could go wrong. When people trust you with their money, that instinct serves a purpose. I wanted us to build something that would hold up under pressure. The first principle we defined at the company: regulatory-first. We worked with regulators for years before we listed a single market. We took every step to do things the right way and always asked for approval before proceeding. It was very hard at times, especially when we were being overtaken by competitors, but we believe doing things the right way matters. Luana and I aren’t the only ones. Our team put years of their lives into something far from certain. That’s especially true of Lior, Brandon, Rainer, Luis, Xavier, and Josef, who stuck with us for years. They believed when there was little to point to. And they stuck with us no matter how long it took. Kalshi is the result of an arduous journey through a long desert where, at times, most people didn’t understand why we were making the walk. Many people still don’t understand what we’re doing and some disagree with us. Markets have historically taken time to be understood and accepted by some. But while not always obvious, their benefits are vast: capital markets help allocate capital and derivative markets help predict and manage uncertainty. I firmly believe in these benefits, and that’s why I am dedicating my life to building this market. Markets have caused issues many times in the past, and they certainly can present harms (manipulation, insiders, and addictive behaviors). These harms are more acute for certain types of markets (sports, short-term options, etc.) than others. It’s why I believe that guardrails need to exist and why we have taken the approach we have taken. We have consistently prioritized thinking critically about the risks and building the appropriate safeguards, sometimes beyond what the regulators require us to do. Many believe more needs to be done, and that’s a conversation I welcome. Over time, I believe more people will see what we see: markets can help us get smarter and manage the future with more rigor, more precision, and more certainty than we have today. To our tens of millions of users: you have made Kalshi bigger than anything Luana and I could have possibly imagined. Thank you for trusting us and thank you for coming along on the journey. I feel a strong responsibility to build a company worthy of that trust. I am so lucky to spend my life building something I love. I hope everyone finds something they care about that much. Something that keeps you curious. Something that pushes you to your limits. Something that may put you at odds with the status quo. Something worth the hard days. For me, it’s still markets. The more I learn, the more I want to understand.

68위Vercel@vercelX에서 보기

Vercel은 Mercedes-AMG F1과 파트너십을 맺었으며, 20세 68일의 키미 안토넬리가 역대 최연소 월드 챔피언으로 2026 F1 드라이버 챔피언에 올랐음을 축하한다.

좋아요 138리포스트 12답글 3노출 12.3K북마크 4

69위Jeff Dean@JeffDeanX에서 보기

JeffDean은 투표 가능한 미국인들에게 화요일 선거에 참여하거나 우편으로 투표할 것을 독려한다.

좋아요 178리포스트 11답글 8노출 12K북마크 2

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. 🗳️

70위Francois Chaubard@FrancoisChauba1X에서 보기

YC Paper Club은 11월 4일 수요일 AI 기반 생물학과 암 치료를 주제로 한 행사를 열며, Omanta, Yusuf Roohani, Divya Nori, Garyk Brixi 등이 발표한다.

좋아요 47리포스트 8답글 2노출 4.3K북마크 38

Francois Chaubard@FrancoisChauba1
Next YC Paper Club: "AI for Biology" We heard AI is going to cure cancer. Meet the people actually doing it. This Wednesday, 11.4, 5-9pm, MV Speaker List: - Omanta: the team behind Sid Sijbrandij's (Gitlab founder's) amazing recovery, now bringing their cancer curing IP to the world. - Yusuf Roohani: Stanford's Virtual Cell program - Divya Nori: "In-context learning" in DNA: Efficient Biological Sequence Design via genomic in-context learning - Garyk Brixi: Building Evo2, implications, and whats next Sign up here:
events.ycombinator.comYC Paper Club - AI for Biology

71위Geoffrey Litt@geoffreylittX에서 보기

패트릭 콜리슨의 글에 제프리 릿이 '유연한 소프트웨어'라는 글을 공유하며 답했다.

좋아요 145리포스트 4답글 2노출 18.9K북마크 115

72위Aditya Vellanki@aditya_vellankiX에서 보기

Sentient가 주력 기능인 Sidekick을 소개하며, 노치를 클릭하면 사용자의 생활 맥락을 바탕으로 일을 처리해 준다고 밝혔다.

좋아요 33리포스트 7답글 5노출 7.1K북마크 16

73위BrendanEich@BrendanEichX에서 보기

BrendanEich는 코네티컷에서 자동 번호판 인식 카메라로 차량에 가짜 세금 고지서를 보낸 민간 업체 사례를 들며, 시민과 언론의 저항으로 철회된 만큼 감시 체제를 거부해야 한다고 주장한다.

좋아요 22리포스트 8답글 2노출 1.9K북마크 9

BrendanEich@BrendanEich
x.com/i/grok/share/b… Citizens fought back and won. Never give in to privatized or public surveillance garbage.
Wall Street Apes@WallStreetApes
This is what the beginning of the surveillance state looks like in America Americans in Connecticut received a bill for additional taxes on their car because cameras spotted it in a different city A private company called Municipal Tax Service (MTS) is hired by cities to drive around and use automatic license plate readers on cars parked in different cities It then makes connections if that car lives in that area and sends the owner a bill for new taxes on that vehicle for being in the city The owner of this vehicle says she doesn’t live in that city, provided proof and says her car was only in a different city because her kid drives there for school MTS didn’t care and said that the proof didn’t matter and they had to pay it. They appealed to the city, the city said they had to deal with MTS But here’s the problem, MTS gets paid 38% commission on tax revenue they recover + $50 per vehicle billed to the owner. They made $9 million in commission in just one year from sending people additional tax bills So this company has no incentive to ever help you and just deny you, which they did Only after the mainstream media got involved would the city drop this one new tax bill which was false to begin with In America we now literally have private companies driving around scamming our license plates to tax us more money This is insanity and it needs to end

74위Aaron Epstein@aaron_epsteinX에서 보기

Dydema의 비침습적 프로그래밍 가능 뇌-컴퓨터 인터페이스를 소개하며, 생각으로 컴퓨터의 제안된 행동을 수락하거나 거절하는 가능성에 감탄함.

좋아요 62리포스트 4답글 4노출 9K북마크 25

Aaron Epstein@aaron_epstein
Dydema is building non-invasive programmable brain-computer interfaces, and it's one of the coolest things I've seen. What's especially interesting is that they're able to control an entire computer to do advanced things like play games and answer emails simply by using computer use to suggest the most likely actions the user will want to take, and accepting or rejecting them with their thoughts. Wild possibilities, and super inspirational @juliver @ChrisOost!
Julian Weaver@juliver
Today we’re launching @dydema (YC F26). We’re building personal, programmable brain-computer interfaces: you pick a thought and bind it to anything, like a key, a game or an agent. No surgery required.

75위Connor@connor_dawsX에서 보기

ernd는 YC F26 배치에서 출시되었으며, 대형 고객만 받는 GPU 가격과 용량을 공동으로 확보해 참여자에게 더 나은 요금과 혜택, 우선 GPU 및 추론 서비스를 제공하겠다고 밝혔다.

좋아요 43리포스트 6답글 9노출 7.8K북마크 17

Connor@connor_daws
we're launching @ernd_ai in @ycombinator's F26 batch When we tried to get GPUs earlier in the batch, we were met with terrible deals or no capacity because we didn’t have enough size. Providers save their best pricing, limits and capacity for their biggest customers. So, we go to providers on behalf of everyone on the card. By paying with ernd, you get access to better rates and top-tier perks. Priority GPUs and inference are on the way!

77위a16z@a16zX에서 보기

a16z의 Connor Love는 구형 국방 조달 방식이 병사에게 10년 전 시스템을 제공한다고 비판하며, 벤처 자본이 정부 요구를 기다리지 않고 미리 투자해 기술을 발전시켜야 한다고 주장한다.

좋아요 40리포스트 3답글 6노출 11.6K북마크 10

a16z@a16z
a16z's Connor Love on why the old defense model gave soldiers systems "developed for 10 years ago," and how private capital lets startups build what they need before the Pentagon asks: "I experienced this in my past when I served in the Army. The buyer is not the user, and the appropriator of the dollars for the buyer is completely removed." "You had primes who would just sit there and listen, like, 'Tell me what you want. I'll develop the thing after you give me the money.'... You get a system that was developed for 10 years ago." "The beautiful thing is venture capital and what we do in American Dynamism, but also the rest of the capital markets is saying, 'We will actually fund that in advance.'" "The outcome of that is the soldier gets a better product, gets something that actually fundamentally changes the battlefield and helps with deterrence or security or safety." "More importantly, you get a type of system that is actually aligned with the requirement of the market and what technology can do today." "Instead of sitting and waiting for the government to tell you what they want, the government signals, and you see it, and you build it in advance by using private capital, and that just has not happened in the history of defense." @ConnorLoveCA

79위adam@theCTOX에서 보기

Y Combinator Winter 2027 지원 마감이 일주일 남았다며 50만 달러 투자와 여러 혜택을 소개하고 지원을 독려하는 글이다.

좋아요 41리포스트 4답글 0노출 7.3K북마크 9

80위BrendanEich@BrendanEichX에서 보기

브렌던 에이크는 텔넷, FTP, VT100이 쓰이던 1982년의 기술과 미디어 시대로 되돌아가면 좋겠다는 바람을 가볍게 밝힌다.

좋아요 23리포스트 1답글 3노출 5.1K북마크 5

81위Anaïs Howland, CFA@AnaisHowland18X에서 보기

YC Winter 2027 지원 마감이 임박했으니 미루지 말고 지원하라고 권하며, 여러 번 떨어진 뒤 합격한 창업자의 경험을 공유했다.

좋아요 58리포스트 5답글 5노출 7.7K북마크 12

Anaïs Howland, CFA@AnaisHowland18
YC Winter 2027 applications are closing soon, apply now, don’t procrastinate! Even if you got rejected for the last batch, persistence wins. I got in the 5th time I applied. I applied twice with my previous startup (@ParadigmShiftAI) and 3x with @OasiveAI (got in on the 3rd try). Feel free to reach out if you have any questions!
Y Combinator@ycombinator
The deadline to apply for the YC Winter 2027 batch is coming up in 2 days! YC founders get: 🔸 $500K in funding 🔸 $12M+ in tokens, credits & deals 🔸 Partners who were founders first, and will always have your back 🔸 Access to the best founder community in the world 🔸 Help launching your company across YC’s channels, with a combined audience of 6M+ 🔸 Connections to 100s of potential enterprise buyers through YC’s Early Access Network, months before Demo Day 🔸 Hiring through Work at a Startup, with 1.25M+ candidates across engineering, design & product. (Founders have already made 2,000+ hires) 🔸 Demo Day in front of thousands of top investors Apply by Nov 2: ycombinator.com/apply.

82위Charlie Weise@_charlieweiseX에서 보기

YC의 'AI로 회사 가속하기' 행사에 다녀온 후기로, 이번 배치에서 가장 좋은 행사였고 AI 네이티브 로켓 회사를 만들지도 모른다는 생각을 전함.

좋아요 32리포스트 3답글 2노출 7.1K북마크 6

83위BrendanEich@BrendanEichX에서 보기

Brave Shields가 자바스크립트 스크립트를 제거해 광고 기술의 연쇄 로딩을 막아 네트워크 사용과 모바일 배터리 소모를 줄인다는 내용이다.

좋아요 29리포스트 4답글 3노출 2.5K북마크 5

84위Uncle Bob Martin@unclebobmartinX에서 보기

unclebobmartin은 난민 메디케이드 사기꾼을 현장에서 적발하는 과정을 담은 영상을 소개한다.

좋아요 18리포스트 0답글 0노출 10.7K북마크 3

85위Google Cloud@googlecloudX에서 보기

Google Cloud는 Firestore 기반으로 flutterflow가 340만 빌더와 200개국으로 확장했으며 5년간 장애가 없었다고 밝히며 안정적인 인프라의 중요성을 강조한다.

좋아요 17리포스트 2답글 0노출 1.8K북마크 2

86위Google Cloud Tech@GoogleCloudTechX에서 보기

Google Cloud 튜토리얼 3편으로, Cloud Run을 이용해 비공개 미리보기 앱을 공개 링크로 배포하는 방법을 소개한다.

좋아요 26리포스트 3답글 1노출 6.5K북마크 3

87위Mitchell Hashimoto@mitchellhX에서 보기

mitchellh는 이 최적화가 순서, 버퍼 크기, 타이밍 중 하나라도 틀리면 효과가 없어서 2023년부터 여러 번 실패 끝에 올해 성공했다고 설명했다.

좋아요 26리포스트 0답글 0노출 6.5K북마크 3

Mitchell Hashimoto@mitchellh

@mitchellh 님에게 보내는 답글

The fun part of this one was if you basically do any single part of this optimization wrong (change order, change buffer size, change timing size) etc then it doesn’t pay off! I tried some form of it like once a year going back to 2023 and you can see failed attempts in PRs where I showed it didn’t work. Then I tried again earlier this year and got it to work. A slog but huge payoffs.

88위Oskar@o_kwasniewskiX에서 보기

YC 겨울 배치 지원 마감이 일주일 남았다는 안내를 인용하며, 지원하지 않으면 합격할 수 없으니 일단 지원하라고 독려함.

좋아요 36리포스트 1답글 0노출 8.5K북마크 4

89위olivia@livilouuuuuuX에서 보기

아이디어에서 몇 달 만에 100명 이상의 구매자를 모은 경험을 공유하고, 30개 기업과 100명 이상의 기업 리더가 참여하는 YC Early Access Network 출시를 알림.

좋아요 43리포스트 2답글 4노출 7.8K북마크 3

90위Mitchell Hashimoto@mitchellhX에서 보기

mitchellh는 이 최적화가 큰 노력 끝에 찾아낸 것이며 다른 터미널들도 채택해 기쁘다고 답했다.

좋아요 17리포스트 0답글 1노출 6.6K북마크 1

91위GitHub@githubX에서 보기

GitHub는 #GitHubUniverse 배지에 무엇을 추가할지 묻고 3D Maze 앱을 소개하는 홍보성 트윗이다.

좋아요 49리포스트 2답글 8노출 15.8K북마크 2

92위Meet Modi@themeetmodiX에서 보기

Y Combinator 윈터 27 지원 마감이 일주일 남았으며, 투자금과 여러 혜택을 강조하며 지원을 독려한다.

좋아요 55리포스트 2답글 2노출 8.6K북마크 1

93위Y Combinator@ycombinatorX에서 보기

Y Combinator은 모든 것을 미리 파악하지 않아도 확신만 있으면 만들기 시작하라며 2027년 W27 기수 지원을 11월 2일까지 받는다고 알린다.

좋아요 9리포스트 0답글 2노출 3.5K북마크 4

94위Jeff Dean@JeffDeanX에서 보기

JeffDean은 18~49세 젊은 유권자의 투표율이 낮은 반면 정책 결정의 영향은 오래 지속되므로 연령 등 인구통계별 투표 격차가 존재한다고 지적한다.

좋아요 12리포스트 2답글 0노출 4.1K북마크 1