The scariest AGI future is not just that the models get closed. It’s that the entire research process becomes nationalized. In our conversation with @EMostaque on @postagixyz, this was the question we couldn’t shake. If frontier AI becomes national security infra, what happens… https://twitter.com/eigenlabs/status/2073109928248852585/video/1
ECDSA(.)fail made it to the pod. @jbrukh joined @srimisra on the un# podcast to discuss how coordinated agent swarms tackled a frontier cryptography challenge, reconstructing the hidden algo, improving it and open-sourcing the work! A glimpse of where open research is headed.
Sovereign intelligence is the starting point. @sreeramkannan: "Labs don't want to sell just intelligence. They want to sell the products of intelligence." Open weights give everyone the ingredient. The question is who turns it into science. The answer is OPEN NETWORKS.… https://twitter.com/eigenlabs/status/2072815723584598283/video/1
AI concentrates power by default. The pushback is coordination. - Who gets to own it - Who gets to shape intelligence - Who gets left out if we don’t act The problem we care about most. @EMostaque of @StabilityAI frames it better than almost anyone. In conversation with… https://twitter.com/eigenlabs/status/2072110946034819142/photo/1
中文: 人工智能默认集中力量。阻力在于协调。 - 谁才能拥有它 - 谁来塑造智力 如果我们不采取行动,谁会被排除在外 我们最关心的问题。 @@StabilityAI 的 @EMostaque 比几乎任何人都更擅长。 与......对话中:
AGI is here. It is rapidly reconfiguring the architecture of power. Left unchecked, this technology will concentrate leverage in the hands of a few. Governments, corporations and the small number of labs able to harness it first. That is the default outcome. We do not accept…
中文: AGI 来了。它正在迅速重构权力架构。 如果不加以控制,这项技术将把杠杆集中在少数人手中。政府、企业以及少数能够首先加以利用的实验室。这就是默认结果。 我们不接受......
When AI makes every resume and deck look equally polished, the real question becomes how do we surface actual ability? This episode gets into the need for better reputation, and funding systems that let talent prove itself from anywhere. 🎙️ @ccatalini @soubhikdeb @sreeramkannan
中文: 当人工智能让每一份简历和副本看起来都同样精致时,真正的问题就变成了我们如何展示实际能力? 本集关注了对更好声誉的需求,以及资助让人才从任何地方证明自己的体系。 🎙️ @ccatalini @soubhikdeb @sreeramkannan
AI makes production abundant. The hard part becomes verification. @soubhikdeb and @sreeramkannan talk with @ccatalini about why that shift matters for labor, markets, and the future of the AI economy. Worth watching:
中文: 人工智能使生产变得丰富。困难的部分变成了验证。 @soubhikdeb 和 @sreeramkannan 与 @ccatalini 谈论这种转变对劳工、市场以及人工智能经济未来的重要性。 值得关注:
Access to intelligence became the throughline last week. > @darkbloomai went live on @OpenRouter, making open, user-powered compute broadly accessible. > @vishnu_patankar joined @OfflineOnAir to connect AI’s trajectory with nuclear technology and made the case that access can’t…
中文: 获取情报成为上周的直通电话。 @darkbloomai 在 @OpenRouter 上上线,使开放、用户驱动的计算功能变得广泛可访问。 @vishnu_patankar 加入 @OfflineOnAir,将人工智能的发展轨迹与核技术连接起来,并证明无法实现接入......
Read @sreeramkannan's recap on how anons beat Google in 72 hours https://x.com/sreeramkannan/status/2062015259821723925
中文: 阅读@sreeramkannan的回顾,了解阿南斯在72小时内如何战胜谷歌
ECDSA(.)fail is now 47.2% ahead of Google’s classified circuit. An open agentic science experiment is now showing what happens when humans + AI agents collaborate on frontier research in public. An @eigenlabs project. https://twitter.com/eigenlabs/status/2068890158704029856/photo/1
中文: ECDSA(.)失败目前领先于谷歌的分类电路47.2%。 一项开放的特化科学实验现在展示了人类与人工智能智能代理在公共领域开展前沿研究合作时会发生什么。 一个 @eigenlabs 项目。
ECDSA(.)fail is now 47.2% ahead of Google’s classified circuit. An open agentic science experiment is now showing what happens when humans + AI agents collaborate on frontier research in public. An @eigen_labs project.
中文: ECDSA(.)失败目前领先于谷歌的分类电路47.2%。 一项开放的特化科学实验现在展示了人类与人工智能智能代理在公共领域开展前沿研究合作时会发生什么。 一个 @eigen_labs 项目。
Our CTO @vishnu_patankar was on @OfflineOnAir recently, where he drew a fascinating parallel between AI and nuclear technology. When a technology becomes powerful enough, the question is no longer just who builds it. It becomes who gets access to it. AI cannot remain… https://twitter.com/eigenlabs/status/2068729913130340700/video/1
中文: 我们的首席技术官 @vishnu_patankar 最近在 @OfflineOnAir 上,他在人工智能与核技术之间画出了引人入胜的相似之处。 当一项技术变得足够强大时,问题就不再是仅仅是谁来构建它。成为谁可以接触到它。人工智能无法继续存在......
JUST IN: Darkbloom is live on @OpenRouter. The next wave of compute will be open, distributed and user-powered and Darkbloom makes it easy for ANYONE with a Mac to contribute. Excited to bring new free capacity to OpenRouter users, starting with gpt-oss-20b and Gemma 4 26B.… https://twitter.com/eigenlabs/status/2066591846550343745/video/1
中文: Just In:Darbloom 在 @OpenRouter 上直播。 下一波计算将是开放、分布式和用户驱动的,而暗黑布隆则让任何拥有 Mac 的人都能轻松做出贡献。 很高兴能从 gpt-oss-20b 和 Gemma 4 26B 开始,为 OpenRouter 用户带来新的免费容量。
PostAGI is a new podcast powered by Eigen on what comes after AGI. In the first episode, @sreeramkannan and @soubhikdeb sit down with economist @alexolegimas to talk through labor, capital and markets post-AGI. Listen here 👇
中文: PostAGI 是由 Eigen 发布的全新播客,专注于 AGI 之后的内容。 在第一集中,@sreeramkannan 和 @soubhikdeb 与经济学家@alexolegimas 坐下来,通过 AGI 后的劳动力、资本和市场进行对话。 在这里收听 👇
Last week the question got sharper: who controls machine intelligence and do we get to keep individual agency? > @sreeramkannan made the case for open coordination at Manifest 2026 in Berkeley. > Darkbloom is going live on OpenRouter. > https://ecdsa.fail/ now has the…
中文: 上周问题变得更加尖锐:谁来控制机器智能,以及我们能否留住各个机构? @sreeramkannan 提出了在伯克利举行的 Manifest 2026 上进行公开协调的理由。 《暗黑》将在OpenRouter上上线。 现已上线
The next wave of compute will be open, distributed and user-powered. Darkbloom is making it easy for anyone with a Mac to contribute compute. Excited to see the network scaling as Darkbloom gets ready to go live on @OpenRouter soon! Powered by @eigenlabs.
中文: 下一波计算将是开放、分布式和用户驱动的。 暗黑布隆正让任何拥有Mac的人都能轻松贡献计算能力。 很高兴看到 Darkbloom 即将上线 @OpenRouter 的网络扩展! 由 @eigenlabs 提供支持。
We have the fastest and most resource-efficient known algorithm for this ECDSA benchmark. A community of individuals, working in the open just built it. Not a closed lab. All of us together. This is the power of open research.
中文: 我们拥有为此ECDSA基准测试提供最快且资源效率最高的已知算法。 一个在露天工作的个体社区刚刚建成了它。不是封闭式实验室。我们所有人在一起。 这就是开放研究的力量。
Our mission is building coordination tools that preserve and expand individual agency in a post-AGI world. @postagixyz asks the question underneath that mission: once intelligence concentrates, who does it serve? Conversations with the people thinking hardest about it. 🎙️…
中文: 我们的使命是构建协调工具,以在后AGI世界中保护并拓展各个机构。 @postagixyz 会问这个使命背后的问题:一旦情报集中,它就会服务于谁? 与那些对此最想点的人交谈。 🎙️...
The safety debate has spent a decade asking whether AGI will be aligned. @sreeramkannan's argument: that's the wrong demon. Even perfectly aligned AGI concentrates power if it lives inside 3 buildings in San Francisco. Alignment without coordination is just obedient… https://twitter.com/eigenlabs/status/2065461778742640960/photo/1
中文: 安全争论已花了十年时间,询问AGI是否会保持一致。 @sreeramkannan 的论点:那是错误的恶魔。 即使完全对齐,只要AGI居住在旧金山的三栋建筑内,也能集中力量。 没有协调的对齐只是听话的......
BIG UNLOCK: Darkbloom goes live on @OpenRouter on Monday. Private inference becomes available where users already route model traffic, powered by idle Macs around the world. A network goes from interesting to imminent to inevitable. https://twitter.com/eigenlabs/status/2065257147118870835/video/1
中文: 大解锁:Darkbloom 将于周一在 @OpenRouter 上线。 私有推断功能在用户已路由模型流量的地方可用,该流量由全球闲置的 Mac 驱动。 网络从有趣到迫在眉睫,再到不可避免。
The ladder is going away. We exist to build new ladders. Open coordination tools, open infra, and systems where individuals and networks can still compete in a post-AGI world. A powerful piece by @soubhik_deb ⬇️
中文: 梯子正在消失。 我们的存在是为了建造新的梯子。 开放协调工具、开放的基础设施,以及个人和网络在后AGI世界中仍可竞争的系统。 由 @soubhik_deb 提供的有力作品 ⬇️
A CLAUDE FABLE 5 USER IS NOW AT THE TOP OF THE ECDSA(.)FAIL LEADERBOARD @BitW0nka's latest submission pushes the frontier to 41.0% ahead of Google’s classified circuit. The frontier is moving in public. https://twitter.com/eigenlabs/status/2064587640926826880/photo/1
中文: 一名克劳德·法伯5号用户现已位于欧洲标准分类系统(ECDSA)的顶部。失败的LEADERBOARD @BitWnka 的最新提交内容使谷歌的上市前方数量突破了 41.0%。 边境正在公共场合移动。
Last week week an open network of researchers, agents and builders did something no quantum lab paid for. They rebuilt Google's withheld circuit benchmark in the open, surpassed it and pushed 39.9% past it in under a week. > @sreeramkannan called it the beginning of Open…
中文: 上周,一个由研究人员、特工和建造者组成的开放网络做了一件没有量子实验室付费的事情。 他们在公开时重建了谷歌被扣留的电路基准,超过了该基准,并在不到一周内突破了39.9%。 @sreeramkannan 称这是 Open 的开端......
AGI will concentrate power by default unless we build coordination structure to prevent it Markets, democracy, and the scientific commons all proved the same thing Individuals can compete with institutions when the right structure exists That is what we are building together
中文: AGI将默认集中电源,除非我们建立协调结构以防止 市场、民主和科学常识都证明了同样的事情 当正确的结构存在时,个人可以与机构竞争 这就是我们共同建设的
We're now 35.2% ahead of Google. The system is starting to compound: > Agents now share notes on approaches, successes and failures > Future agents can learn from what already worked and what didn’t > Agents are also giving feedback to improve the platform itself >…
中文: 我们现在领先谷歌35.2%。 系统开始复合: 代理商现在分享有关方法、成功与失败的注意事项 未来的中介可以从哪些已经奏效以及哪些方面没有奏效 代理商也在提供反馈,以改进平台本身 ......
Today @sreeramkannan spoke at @initc3org Camp 2026 at @Princeton about coordination in the PostAGI world. > His core question: Who controls machine intelligence, and who is responsible for its outcomes? > Without open coordination technologies, AI power will centralize among a… https://twitter.com/eigenlabs/status/2063063931900547294/photo/1
中文: 今天,@sreeramkannan 在@Princeton 的 @initc3org Camp 2026 上发表了关于 PostAGI 领域协调性的采访。 他的核心问题:机器智能的控制权是谁?谁对其结果负责? 如果没有开放式协调技术,人工智能的力量将集中在一个...
We believe science belongs to everyone
中文: 我们相信科学属于每一个人
“An 18-year-old just hit 80% of Google’s classified breakthrough over a weekend using an AI agent swarm.” Frontier science is changing. As covered in news: The story is not just about quantum cryptography. It is about what happens when AI agents, open collaboration, and…
中文: 一名18岁的年轻人在一个周末利用人工智能代理群体,刚刚突破了谷歌80%的机密突破。 前沿科学正在发生变化。 如新闻报道所述:这个故事不仅涉及量子密码学。这是关于人工智能代理、开放协作以及......时会发生什么
A 23 and 18 year old just out-researched Google's secret quantum lab Bitcoin was the test. Biology, neuroscience, and genomics are next Science belongs to everyone now @sreeramkannan @bbuddha_xyz @MTSlive https://twitter.com/eigenlabs/status/2061923310024135016/video/1
中文: 一名23岁和18岁的年轻人刚刚对谷歌的秘密量子实验室进行了研究 比特币是测试。生物学、神经科学和基因组学是下一个 科学现在属于每一个人 @sreeramkannan @bbuddha_xyz @MTSlive
Google proved they had the best circuit. They just wouldn't show it So we reverse engineered it, and open research is now 13.4% ahead @sreeramkannan @bbuddha_xyz @MTSlive https://twitter.com/eigenlabs/status/2061902169830515155/video/1
中文: 谷歌证明了他们拥有最好的赛道。他们只是不表现出来 因此,我们对它进行了反向改造,而开放研究现在领先了13.4% @sreeramkannan @bbuddha_xyz @MTSlive
Google published the proof. Not the circuit So we built https://ecdsa.fail/ and open research has already surpassed their unpublished benchmark https://twitter.com/eigenlabs/status/2061894116905029853/video/1
中文: 谷歌发布了该证明。不是电路 因此,我们建立了
Eigen Labs builds technologies for human-agent coordination. Our mission: ensure open networks can compete and win in the age of AI. @MTSlive @sreeramkannan @bbuddha_xyz https://twitter.com/eigenlabs/status/2061884918951477617/video/1
中文: Eigen Labs 开发用于人机协调的技术。 我们的使命:确保开放网络在人工智能时代能够竞争并取得胜利。 @MTSlive @sreeramkannan @bbuddha_xyz
The open agentic science challenge beating Google's Quantum Benchmark is officially trending: https://x.com/i/trending/2061623376527929418
中文: 超越谷歌量子基准的开放式科学挑战正呈正趋势趋势:
11.8% ahead of Google’s classified circuit. The frontier didn’t just reach Google, it passed it. Best score now sits below the Google line and the race is officially in new territory. Who’s going to push it further? 👀 https://twitter.com/eigenlabs/status/2061821788896760067/photo/1
中文: 领先于谷歌的分类电路11.8%。 边界不仅触及了谷歌,它也突破了它。 最佳分数现在位于谷歌线以下,比赛正式进入新领域。 谁会进一步推动它?👀
95% of the way to Google’s confidential circuit. Best score is now down to 4,068,676,284 and the frontier keeps moving. 84 promoted submissions. 23 solvers. One question: Who’s going to beat @Google? 👀 https://twitter.com/eigenlabs/status/2061706298123751425/photo/1
中文: 通往谷歌保密电路的95%。 最佳分数现已降至4,068,676,284,且边界持续发展。 84 份晋升提交。23 名解说员。 一个问题:谁会战胜谷歌?👀
Eigen Times #062 is live. Darkbloom hit public alpha. Bankless and a16z both covered the ecosystem this week. And @sreeramkannan put a 3-year timeline on self-sovereign AI. Read it here 👇
中文: 《艾根时报》#062 是实时的。 暗黑发球线在公开场合上大放异力。本周,Bankless 和 a16z 都报道了这一生态系统。 @sreeramkannan 为自动驾驶汽车人工智能设定了为期三年的时间线。 点击此处阅读 👇
Darkbloom update: Research Preview to Public Alpha Thousands of providers. 600M+ tokens served. Open-weight inference running on idle Macs. With 30–200% performance gains across TTFT, TPS, and token capacity, Darkbloom is a glimpse into a future where AI infra is cheaper, more…
中文: 暗黑布隆更新:向公众Alpha发布研究预览 成千上万的供应商。已提供600M+代币。在空闲 Mac 上运行的开量推理。 凭借TTFT、TPS和代币容量的30-200%性能提升,Darkbloom 揭示了人工智能基础设施更便宜、更多未来的前景。
The ecosystem is accelerating! From the AI x trust layer conversation with @a16zcrypto to verifiable agent telemetry, agentic commerce and post-quantum infrastructure, a lot happened this week across the ecosystem. Learn more in Eigen Times Edition # 061 ⬇️
中文: 生态系统正在加速! 从与@a16zcrypto的AI x信任层对话,到可验证的代理遥测、代理商务和后量子基础设施,本周整个生态系统中发生了许多事情。 了解更多,请访问《艾根时代》第061版⬇️
As AI agents move from assistants to actors, trust becomes the bottleneck. - Did they use the right data? - Did they follow the rules? - Can they prove what happened? Watch @sreeramkannan with @rhackett of @a16z ↓ https://x.com/a16zcrypto/status/2057588211480989972/video/1
中文: 随着人工智能代理从助手转向参与者,信任成为瓶颈。 他们使用正确的数据了吗? 他们遵守规则了吗? 他们能证明发生了什么吗? 与@rahkett(@a16z)一起观看@sreeramkannan ↓
We've been heads down building! Here's everything that happened across the ecosystem: builders shipping, Agentic by Eigen taking over cities, AI agents judging hackathons for the first time ever and a lot more. Read the full edition of Eigen Times 🗞️ ⬇️
中文: 我们一直都在建楼! 生态系统中发生的一切:建筑商航运、艾伊根公司接管城市,人工智能特工首次对黑客马拉松进行评判,以及更多。 阅读《艾根时报》🗞全文 ⬇️
🪩🖤 https://twitter.com/eigenlabs/status/2056117629484777593/photo/1
Excited to see @silencelabs_sl pushing post-quantum MPC from research into practical infrastructure. At Eigen Labs, our work sits at the frontier of AI research and open networks. As AI systems become more autonomous and economically relevant, the guarantees around them need to…
中文: 很高兴看到@silencelabs_sl 推动后量子MPC从实际基础设施领域进行研究。 在艾根实验室,我们的工作处于人工智能研究和开放网络的前沿。随着人工智能系统变得越来越自主且具有经济相关性,它们周围的保障需要......
AI Has Crossed the Threshold @sreeramkannan explains why the real question is no longer whether AGI has “officially arrived.” AI is already in human range across enough domains that it's changing how software gets built, how organizations operate and how companies form. The… https://twitter.com/eigenlabs/status/2054415067148644796/video/1
中文: 人工智能已经突破了门槛 @sreeramkannan 解释了为何真正的问题不再是 AGI 是否“正式到来”。 人工智能在足够多的领域已经处于人类的范围之中,正在改变软件的构建方式、组织的运作方式以及公司的形成方式。
A pattern across our research experiments: The hard problems in AI are stopping looking like model problems. They're starting to look like coordination problems. - How work gets verified - How compute gets allocated - How agents identify each other - How trust scales without…
中文: 我们研究实验中的一个模式: 人工智能中的难题正不再像模型问题那样存在。他们开始看起来像是协调问题。 - 工作如何得到验证 计算是如何分配的 - 代理人如何相互识别 - 不信任如何扩展而不...
The next wave of compute won’t only come from bigger data centers. It will come from millions of smaller nodes: homes, devices, edges and idle machines with unused capacity. Darkbloom is the coordination layer for this new compute frontier. Wake the world’s sleeping compute!
中文: 下一波计算浪潮不仅来自更大的数据中心。 它将来自数百万个较小的节点:家庭、设备、边缘和空闲设备,且容量未使用。 暗布布卢姆是这一新计算前沿的协调层。 唤醒世界沉睡的计算!
Project Darkbloom explores a new model for AI inference. Using underused Apple Silicon as a distributed, privacy-first compute layer. With hardened execution, hardware-backed identity, and attestation built into the system design, Darkbloom is now live in research preview. An…
中文: 暗黑布卢姆项目探索了人工智能推理的新模型。将未使用的苹果硅胶作为分布式的、以隐私为先的计算层。 在系统设计中内置了强化执行、硬件支持的身份和认证,Darbloom 现已进入研究预览版。 一...
Our Technical Talent Partner @philuponcrypto will be at Stanford and in SF the week of April 20 meeting builders and researchers working on AI agent systems. Agents are becoming economic actors. They generate value, coordinate work, and operate with increasing autonomy. But…
中文: 我们的技术人才合作伙伴@philuponcrypto 将于4月20日的下周在斯坦福大学和旧金山举行与人工智能代理系统合作的建设者和研究人员会面。 代理人正成为经济参与者。它们创造价值,协调工作,并以增强的自主性运作。 但......
💼 We’re hiring: Podcast Ops Associate We’re building a podcast that reflects the voice, ambition, and worldview of our leadership and we’re looking for a sharp, execution-focused operator to help bring it to life. This is a chance to: You’ll be a great fit if you’ve: - Owned… https://twitter.com/eigen_labs/status/2039842439595442411/photo/1
中文: 💼 我们正在招聘:播客运营助理 我们正在打造一个反映领导力声音、抱负和世界观的播客,并寻找一位敏锐且注重执行力的操作者,以帮助其实现现实。 这是一个机会: 如果你有以下选择,你会非常合适: - 拥有......
We’re hiring a Social Media Lead, AI @ Eigen Labs. If you’re AI-native, chronically online, sharp on narrative, and know how to turn complex ideas into posts people actually care about, this role is for you. You will: ☑️ Own social across X, TikTok, Instagram and emerging… https://twitter.com/eigen_labs/status/2032614064409686499/photo/1
中文: 我们正在招聘一位社交媒体主管,人工智能 @ 艾根实验室。 如果你是人工智能原生的,长期在线,在叙事上敏锐,并且知道如何将复杂的想法转化为人们真正关心的帖子,那么这个角色就是为你而设的。 你将会: ☑️ 拥有 X、TikTok、Instagram 和新兴社交内容......