Liner

🚀 라이너에서 글로벌 프로덕트를 만들어갈 ML Engineer 전문연구요원분들을 모십니다. 채용 설명회 사전 신청하기 👉🏻 https://lnkd.in/ePWW4zzF

Stanford just accepted and recognized a research paper co-authored by a high school student and AI. At Agents4Science 2025, the world’s first AI-author-centric academic conference, one of the ten spotlight papers wasn’t from a PhD. It was from Jaeyoung Choi, a first-year high school student from Korea. He used Liner's AI agents as a research partner, brainstorming hypotheses, finding citations, and even running peer review before submission. The next generation of discovery won’t begin only in universities or research labs. AI is already lowering the barriers to research and making science accessible to anyone curious. Read the full story here: https://lnkd.in/gDhvrj4c Special thanks to the organizers of Agents4Science 2025 for leading this new chapter of AI-driven research! James Zou Owen Queen Nitya Thakkar Eric Sun Federico Bianchi

When the NeurIPS 2025 numbers came out, the reaction was immediate: 25,000+ submissions, about 20% acceptance, reviewers overwhelmed. Many called it a bloodbath. That frustration is real. But zooming out, the bigger story is not that conferences have gotten too big. It is that for most of history, they were far too small. For decades, fewer than 0.1% of people produced nearly all the science the world relied on. That scarcity made conferences look orderly, but it also meant billions of smart, curious people never had the chance to contribute. The surge in submissions is a sign of progress, not decline. The real failure is that the infrastructure of science — peer review, conference models, publication norms — has not adapted. This is where AI steps in. Research agents can break Shockley’s old bottlenecks by helping researchers find meaningful problems, recognize valuable results, and communicate clearly. They keep us from drowning in the flood by scanning the literature, surfacing insights, and connecting dots that would otherwise be lost. The frustration researchers feel today is real. But it is not the end of the story. It is the growing pain of a system that must evolve. Without AI, scale collapses under its own weight. With AI, scale becomes our greatest strength. Image Source: https://lnkd.in/gUS9VtPw

Stop trusting stale AI data. Liner Search MCP is live. 200 free searches a day. "Tell me what changed in TypeScript 7 and what to fix in our code." Your agent can't answer that from training data. TypeScript 7 shipped in July — the Go rewrite, 8-12x faster builds, strict on by default, a pile of hard errors that used to be warnings. Ask anyway and it'll answer confidently from 5.x. So you add a search tool. Now you get ten links and a snippet back — a to-do list, not an answer. You still open each page, read it, and work out what's true. You saved nothing. Liner Search MCP: → Search — one tool, liner_web_search, up to 10 results with date filtering → Verify — every claim comes back with its source and publication date → Finish — code changes, error fixes, comparison tables, sourced drafts → Continue — hit the daily limit and there's no 429. You get a normal response and paid credit picks up, so the loop never breaks mid-task Ask what changed in TypeScript 7 and you get the exact tsconfig lines to change, with the release notes and two migration guides underneath. Not where to look. What to change. One thing we do differently: we make you sign in. OAuth 2.1 + PKCE — the standard for MCP, the same flow Linear, Notion, and Sentry use. Your client opens a browser, you click approve, you're connected. No keys to create, rotate, or leak. After the free tier, $1 per 1,000 calls. The first 1,000 accounts to connect keep 200 calls a day for 12 months. Connect in one click ↓ https://abr.ge/zuzo1l5

에이전트에 도구 붙이기, 너무 번거로웠습니다. 그래서 Liner Actions를 만들고 있고 먼저 써보실 분 1,000명께 1년 이용권을 드립니다. 지금은 수천 개의 MCP 중에서 직접 찾고, 문서를 읽고, 설치하고, 설정합니다. 그렇게 미리 연결해둔 도구만 쓸 수 있었습니다. 심지어 그게 제대로 동작하는지도 직접 확인해야 했습니다. 하지만 Liner Actions는 필요한 도구를 에이전트가 알아서 찾아 씁니다. 미리 연결해두지 않아도 됩니다. "깃허브 이슈 정리해서 우선순위 정하고 담당자까지 지정해 줘." 딱 이 한마디면 됩니다. 저희가 모든 MCP 서버를 직접 실행해보고 검증하고 있습니다. 없으면 만들려고 하고 요즘은 여기에 가장 많은 시간을 쓰고 있습니다. 지금 사전 신청하시면 선착순 1,000명에게 Liner Actions 1년 이용권을 드릴테니 많이 사용해보시면서 다양한 의견부탁드립니다. 어떤 도구가 필요한지, 어떻게 동작하면 좋겠는지 알려주시면 최대한 반영하겠습니다. 사전 신청하기 https://lnkd.in/ggn8ZJtZ

Need free Search API credits? We're giving away $3,000 in credits. Claim up to $100 before the pool runs out 💰 If you're building an AI product that needs reliable search, this is your chance to try Liner's Search API and Search Agents. Liner helps developers add reliable, source-backed web and academic search to their products. New users receive a random credit reward on a first-come, first-served basis. How to claim: 1. Open the link below. 2. Sign up for Liner API. 3. Get your credits and start building. Claim your gift. Try it now ↓ https://lnkd.in/g5heJvAR