数据抓取时间:2026-11-07 08:14 来源:Hacker News
Hacker News 每日早报(2026-11-07)
今天精选了 15 条 Hacker News 热门条目,覆盖 AI 模型竞争、Agent 安全、隐私与数据训练、软件工程与复古计算等话题。
1. CATL Sodium Battery Hits LFP Cost Parity in 2026: 15,000 Cycles, 30-Year Life
-
来源: Hacker News | 时间: 2026-11-06 23:00 UTC | 热度: 65 points | 评论: 24
-
讨论: Hacker News 评论
-
摘要: CATL Sodium Battery Hits LFP Cost Parity in 2026: 15,000 Cycles, 30-Year Life
-
核心评论:
- EwanToo:Alternative site which is still up https://www.pv-magazine.com/2026/11/06/sodium-ion-batteries-…
- syntaxing:I’m so desperate for home battery using sodium batteries. Safer and lower temperature it can operate at. Low energy density but I don’t care for home battery.
- KellyCriterion:Just imagine: ALL the money that flew into “Fossil & Co” throughout the last, lets say, 50 or 60 years, including all implicit fundings, like state-building-streets or universities-masters-engineering-focus-ICE, all the money for lobbying, all the money for all the (counter)campa…
-
深度解读: 该话题在 HN 社区引发 24 条讨论。评论焦点包括:Alternative site which is still up https://www.pv-magazine.com/2026/11/06/sodium-ion-batteries-…;I’m so desperate for home battery using sodium batteries. Safer and lower temperature it can operate at. Low energy dens。
2. A Jev that can ask Opus for help gives Opus accuracy at 4x lower cost
-
来源: Hacker News | 时间: 2026-11-06 19:48 UTC | 热度: 287 points | 评论: 63
-
讨论: Hacker News 评论
-
摘要: LLM Judges for Identity Resolution at 4x Lower Cost | Hightouch | Hightouch Skip to main content Yes, Hightouch can send messages now. Learn more → A Jev that can ask Opus for help gives Opus accuracy at 4x lower cost We tested whether fast, cheap decision models like Jev and GPT-6 Luna could grade identity resolution as well as frontier LLMs, and …
-
核心评论:
- xlayn:If you use Qwen3.8 27B to manage, to orchestrate, do do shallow research, and then make it ask claude you will notice that Qwen can get out of Haiku, Sonnet and Opus almost as good answers… I was researching how good the model asnswers where, I can’t get to work with anything b…
- nico:This is very cool. But also, it’s a benchmark/result for a very specific experiment/task: identity resolution Despite the title, the article is not making claims of, or demonstrating general usage improvements or general coding agent improvements
- angry_octet:It’s illustrative that there are even errors in how the humans have formulated the task and documented the errors. They propose: Valentino → Female Child Yet this is obviously a boy name, else it would be Valentina. Is this model going to return confidence values? No, they want d…
-
深度解读: 该话题在 HN 社区引发 63 条讨论。评论焦点包括:If you use Qwen3.8 27B to manage, to orchestrate, do do shallow research, and then make it ask claude you will notice th;This is very cool. But also, it’s a benchmark/result for a very specific experiment/task: identity resolution Despite th。
3. Firecracker VM Escape: PwnEscape Part 2
-
来源: Hacker News | 时间: 2026-11-06 19:33 UTC | 热度: 147 points | 评论: 68
-
讨论: Hacker News 评论
-
摘要: Back Firecracker VM Escape: PwnEscape Part 2. PWN Research, Paulos Yibelo · 19 m 19 min read Firecracker is supposed to keep untrusted code inside a microVM. We escaped it and achieved code execution as root on the host. We found how it saved registers, changed where it would resume, and made it execute a call of our choosing when it returned from …
-
核心评论:
- scoodah:I really wish they would have expanded on this paragraph in their article further: > The configuration used here had no jailer. Firecracker’s production containment includes dropping privileges and restricting the process through namespaces and seccomp. A deployment using those c…
- lrvick:I am more convinced by the day that the Nova Microhypervisor approach is correct. Do virtualization with a tiny formally verified microkernel and have complex kernels like Linux run underneath instead of letting a huge, complex, and brittle linux kernel and vm ecosystem control t…
- zX41ZdbW:I have a demo service that entirely relies on Firecracker VMs: ClickBench playground, https://benchmark.clickhouse.com/playground/ It hosts hundreds of database engines - many of them have vulnerabilities, and some of them, like Pandas, are basically open RCEs regardless. We can …
-
深度解读: 该话题在 HN 社区引发 68 条讨论。评论焦点包括:I really wish they would have expanded on this paragraph in their article further: > The configuration used here had no ;I am more convinced by the day that the Nova Microhypervisor approach is correct. Do virtualization with a tiny formally。
4. Tesla to wind down Cybertruck production
-
来源: Hacker News | 时间: 2026-11-06 12:13 UTC | 热度: 805 points | 评论: 1640
-
讨论: Hacker News 评论
-
摘要: Tesla to wind down Cybertruck production
-
核心评论:
- kraig911:Full disclosure: I work for a competitor to Tesla so I’m a bit biased. The Cybertruck I thought was initially kind of neat but when you try to use one as a truck you quickly can see some glaring problems that most people who buy a truck want. Towing with an EV is terrible. I’ve h…
- brightball:After Helene knocked out power in parts of my area for a little over 14 days, I found out that a friend of ours used his cybertruck to power his house through the entire mess. He’d run the house on it for a couple of days, then drive the cybertruck over to a super charger station…
- paxy:- Federal EV tax credits are over. - The emissions credit market (where Tesla made most of its money) is dead due to lowered standards. - Manufacturing costs haven’t come down. - EV sales share in the US is down to 6% (and falling every quarter). - Chinese EVs are dominating ever…
-
深度解读: 该话题在 HN 社区引发 1640 条讨论。评论焦点包括:Full disclosure: I work for a competitor to Tesla so I’m a bit biased. The Cybertruck I thought was initially kind of ne;After Helene knocked out power in parts of my area for a little over 14 days, I found out that a friend of ours used his。
5. US FAA says it will phase out floppy disks from air traffic control by 2028
-
来源: Hacker News | 时间: 2026-11-06 04:00 UTC | 热度: 260 points | 评论: 205
-
讨论: Hacker News 评论
-
摘要: US FAA says it will phase out floppy disks from air traffic control by 2028
-
核心评论:
- bargl:I worked on modernization in the FAA specifically around digitizing NOTAMs. I would urge you not to consider this modernization from the perspective you may be familiar, as someone who may work on the bleeding edge and sometimes sees a lack of caution from engineers pushing for c…
- jonhohle:More information needed. Was this a sneakernet? No one thought of replacing with a floppy emulator and SD cards? All of the data can fit into 1.44MB? If the last one is true, devs take note. If FAA can rely on slow, small storage for safety critical applications, you probably can…
- driverdan:The politically grandstanding is gross. From what I can tell E-IDS has been in development since at least 2020 with a planned rollout starting in 2025. Unfortunately web search has gotten so bad that it’s hard to find past information about it.
-
深度解读: 该话题在 HN 社区引发 205 条讨论。评论焦点包括:I worked on modernization in the FAA specifically around digitizing NOTAMs. I would urge you not to consider this modern;More information needed. Was this a sneakernet? No one thought of replacing with a floppy emulator and SD cards? All of 。
6. AI is dramatically accelerating AI research
-
来源: Hacker News | 时间: 2026-11-06 20:27 UTC | 热度: 169 points | 评论: 131
-
讨论: Hacker News 评论
-
摘要: AI is dramatically accelerating AI research right now Planned Obsolescence Subscribe Sign in AI is dramatically accelerating AI research right now If acceleration compounds, the naive extrapolation implies superintelligence within a year Ajeya Cotra Nov 06, 2026 84 14 20 Share This post discusses a public METR report on AI R&D acceleration at Anthr…
-
核心评论:
- nsilvestri:AI is accelerating my work…work which only exists because of AI. If I had to do my individual tasks without AI, I would be slower. But if I take a step back, I’m really not changing the world any faster than I was before.
- cronin101:Not trying to discredit the premise but I do wonder how much this is from vertical scaling of model intelligence in weights vs the sheer amount of capital being thrown at training and horizontal scaling of agent swarms etc. Anecdotally it seems like the majority of continued upli…
- deepsquirrelnet:For a much more thorough, alternate take I found this article better: https://open.substack.com/pub/rameznaam/p/ai-rsi-isnt-leadin…
-
深度解读: Ajeya Cotra 基于 METR 关于 AI 研发加速的公开报告(对象包括 Anthropic 等实验室)展开推演:如果加速效应持续复合增长,朴素外推意味着一年内逼近超级智能。社区反应分裂:有人体感 AI 确实在加速自己的工作,也有人指出这种加速更多来自资本涌入和 Agent 集群的横向扩展,而非模型智能本身的垂直提升。
7. ChatGPT silently passes data between connectors and the bounty wants an ID card
-
来源: Hacker News | 时间: 2026-11-06 22:50 UTC | 热度: 45 points | 评论: 14
-
讨论: Hacker News 评论
-
摘要: Security: Follow-up tool call silently routed to a different connector - ChatGPT Apps SDK - OpenAI Developer Community Security: Follow-up tool call silently routed to a different connector ChatGPT Apps SDK AldiPower November 5, 2026, 9:30am 1 In a conversation with two connectors enabled, ChatGPT made a follow-up tool call through a different conn…
-
核心评论:
- simonw:I assumed that the ID requirement was about sanctions compliance - OpenAI can’t legally send money to someone in North Korea, for example. But https://docs.bugcrowd.com/researchers/managing-account/accou… says: > To improve platform integrity by reducing high volumes of low-qua…
- kristjansson:This seems like a fundamental misunderstanding / over estimation of the guarantees OAI and MCP can offer. If ‘confirmation’ is a separate tool call … it’s a separate tool call. There’s no scoping, there’s no ‘binding’, nothing. When you hand control back to the model you hand c…
- orf:Why did “techsoft.ye” feel the need to reply with a lengthy LLM comment that contained nothing useful at all?
-
深度解读: OpenAI 开发者社区披露:当用户同时启用两个 connector 时,ChatGPT 的后续工具调用会被静默路由到另一个 connector。Simon Willison 等指出,这暴露了 MCP 生态的权限边界缺陷——『确认』只是另一个工具调用,没有 scoping 也没有 binding,控制权一旦交回模型,安全假设便不复存在。叠加 Bugcrowd 要求赏金猎人提交身份证才能领赏,社区开始重新审视 AI 工具链的信任基础。(https://kevinboone.me/efitest_index.html)
-
来源: Hacker News | 时间: 2026-11-06 16:50 UTC | 热度: 172 points | 评论: 31
-
讨论: Hacker News 评论
-
摘要: Kevin Boone: First steps in bare-metal programming using UEFI – series index Kevin Boone Home Contact About Software Articles 🔍 First steps in bare-metal programming using UEFI – series index These articles are about using a Linux workstation to develop bare-metal applications for a PC-type computer with UEFI firmware. They are written for experien…
-
核心评论:
- the_panopticon:For “Get used to the lack of a C standard library” mentioned in https://kevinboone.me/efitest4.html there’s always https://github.com/tianocore/edk2-libc . And on “Be aware of the Windows-ness of EFI” commentary the resemblance to Windows coding style https://computernewb.com/~li…
- kevin_thibedeau:> The linker won’t complain if your code calls functions that don’t exist – it will just mark them as external, expecting them to be provided by some other code. But they won’t be, and the program will just crash when it calls one of these missing functions. You can verify no ext…
- edoceo:This guide is fantastic for getting rolling in the UEFI space. I used this as one of the references when building a bootloader with Zig. Has its own set of issue, especially that 16bit character situation. Once you dig into those there are a bunch of sharp edges, or maybe it’s ju…
-
深度解读: 该话题在 HN 社区引发 31 条讨论。评论焦点包括:For “Get used to the lack of a C standard library” mentioned in https://kevinboone.me/efitest4.html there’s always https;> The linker won’t complain if your code calls functions that don’t exist – it will just mark them as external, expectin。
9. Restaurant Delivery Payment-Skimming Campaign
-
来源: Hacker News | 时间: 2026-11-06 19:19 UTC | 热度: 74 points | 评论: 30
-
讨论: Hacker News 评论
-
摘要: Scamwich: Inside a Restaurant Delivery Payment-Skimming Campaign — InfoHawk News ← All news November 5, 2026 Scamwich: Inside a Restaurant Delivery Payment-Skimming Campaign InfoHawk research has uncovered a network of 820 suspicious restaurant domains impersonating major US chains, including Jersey Mike’s, Jimmy John’s, Firehouse Subs, Chipotle, a…
-
核心评论:
- ModernTimes:None of this would work if Google Adsense didn’t allow scam ads on their system. But they do, millions every day (I deal with this every day on my websites, Google don’t care). https://en.wikipedia.org/wiki/Racketeer_Influenced_and_Corru…
- mlmonkey:Couldn’t the original website owner just look at the logs and see where most of the requests are coming from? And also the referrers would be a dead giveaway, right? That’s assuming someone is looking at the
access.logfile occasionally. - pixl97:Hmm, how do you even protect against this one? Rewriting the entire website makes the entire situation a bit more difficult, you can control about everything.
-
深度解读: 该话题在 HN 社区引发 30 条讨论。评论焦点包括:None of this would work if Google Adsense didn’t allow scam ads on their system. But they do, millions every day (I deal;Couldn’t the original website owner just look at the logs and see where most of the requests are coming from? And also t。
10. Replit rebuilt its sandbox engine with microVMs
-
来源: Hacker News | 时间: 2026-11-06 19:06 UTC | 热度: 53 points | 评论: 14
-
讨论: Hacker News 评论
-
摘要: Inside Replit’s New Sandbox Engine, Part 1: The Path to MicroVMs Skip to content Search engineering posts Esc Sections Agent sandboxes are all the rage. A new sandbox startup is announced every couple of weeks. At Replit, we’ve been running sandboxes for nearly a decade. For years, we’ve run our custom sandbox infrastructure to make it easy for any…
-
核心评论:
- ghshephard:Replit intrigues me. Other than Paul Graham talking it up a lot on Twitter - I’ve never once heard of anyone using it, seen any chatter about it in any or our internal slack forums. Every other AI company I get barraged with- no matter where I go, even antigravity/gemini carbon/a…
- schainks:My read on this is they re-invented the OTP supervision tree, process isolation, hibernation, and actor-to-actor message passing, but had to push the isolation boundary all the way down into Ring -1 (KVM/Cloud Hypervisor) and customized Linux tinyconfig kernels because their “act…
- clhodapp:It’s only somewhat related to the article, but: How do people feel about the hardware isolation provided by microVMs versus strong software isolation techniques like gvisor? Right now, we put a huge amount of stock in “hardware isolation”. If that starts falling to agents to any …
-
深度解读: 该话题在 HN 社区引发 14 条讨论。评论焦点包括:Replit intrigues me. Other than Paul Graham talking it up a lot on Twitter - I’ve never once heard of anyone using it, s;My read on this is they re-invented the OTP supervision tree, process isolation, hibernation, and actor-to-actor message。
11. Google has acquired Crusoe’s data center site in southeast Wyoming
-
来源: Hacker News | 时间: 2026-11-06 20:09 UTC | 热度: 29 points | 评论: 5
-
讨论: Hacker News 评论
-
摘要: Google Acquires Crusoe’s Wyoming Data Center Site November 6, 2026 Crusoe Announces Wyoming Data Center Campus Update Google has acquired Crusoe’s data center site in southeast Wyoming, completing Crusoe’s pre-development work on site selection, governmental approvals, and a hybrid power strategy developed with Tallgrass Energy. No items found. Nov…
-
核心评论:
- tristanj:Google announced yesterday they would invest $17 billion into this Wyoming data center project https://blog.google/innovation-and-ai/infrastructure-and-clo… > Data center upstart Crusoe is being pushed aside at an artificial intelligence campus it was developing in Wyoming afte…
- simonw:Presumably the value of data centers that (1) already exist and (2) already have deals to provide the energy (and water) they need is about to skyrocket.
- zerozerotwo:This isn’t trans meta crusoe right?
-
深度解读: Google 收购 Crusoe 位于怀俄明州东南部的数据中心地块,据传将在此投入 170 亿美元建设 AI 数据中心。评论指出:已建成且锁定电力与水资源供应协议的数据中心价值正在飙升,这场交易也折射出 AI 基建竞争中『囤地囤电』的激烈格局。
12. Memorilla: Latent Semantic Memory for LLMs
-
来源: Hacker News | 时间: 2026-11-06 18:03 UTC | 热度: 77 points | 评论: 6
-
讨论: Hacker News 评论
-
摘要: [2610.21375] Memorilla: Latent Semantic Memory for LLMs Skip to main content Search arXiv Press Enter to search · Advanced search Computer Science > Machine Learning arXiv:2610.21375 (cs) [Submitted on 15 Oct 2026 ( v1 ), last revised 3 Nov 2026 (this version, v3)] Title: Memorilla: Latent Semantic Memory for LLMs Authors: Adibvafa Fallahpour , Par…
-
核心评论:
- jkingsman:So if I’m reading this right, it’s basically a specialized embedding of a RAG, and then a model in the middle takes the agent’s queries and returns responses in a form that is injecting into the main model’s input token stream directly i.e. skipping actual text. And the injection…
- bob1029:I am finding that semantic ~= lexical when you have something like a frontier model driving. Using full text search engines is still the most compelling approach for handling large document sets. I really struggle with the semantic / embedding concerns when we have a generative m…
-
深度解读: Memorilla 提出了一种『潜在语义记忆』:用特化的 embedding 存储经验,由中间模型直接以注入 token 流的形式将记忆返回主模型,跳过文本化环节。评论认为这本质上是特化版 RAG,而在前沿模型驱动下,语义检索与传统全文检索的能力差距正在缩小。
13. Devvm.dev: One VM for every project
-
来源: Hacker News | 时间: 2026-11-06 17:30 UTC | 热度: 96 points | 评论: 45
-
讨论: Hacker News 评论
-
摘要: devvm Skip to content Docs Getting started Installing Your first VM Everyday commands Using devvm Coding agents Toolchains SSH and editors Ports and size Managing VMs Images and tools Customizing images Settings and files Working on a clone devcontainer.json Links between VMs A VM’s desktop Starting from a container image What devvm keeps on this m…
-
核心评论:
- photonios:I keep seeing projects like this pop up and they never nail it. I need isolation, but I also need to give my agents secure access to internal package registries, logs, internal infra, designs, emulators, browsers, docs etc. At a large company, work can span dozens of repositories…
- cliche:I made something like this at work as a way to spin up multiple dev stacks on the same project. It’s docker based though so not the best isolation. The problem i having is more workflow related. AI is doing the dev so fast I constantly need to validate different versions and chan…
- mnahkies:Damn, I’ve been building something very similar the past few weeks - haven’t quite gotten to a public release yet. Mines based on libvirt and uses cloud-init for creating the images, with shared golden images across projects. Skimming your docs it wasn’t totally clear to me if yo…
-
深度解读: 该话题在 HN 社区引发 45 条讨论。评论焦点包括:I keep seeing projects like this pop up and they never nail it. I need isolation, but I also need to give my agents secu;I made something like this at work as a way to spin up multiple dev stacks on the same project. It’s docker based though。
14. The Ubuntu 26.04 Rust sudo is not compatible with standard sudo
-
来源: Hacker News | 时间: 2026-11-03 16:43 UTC | 热度: 127 points | 评论: 104
-
讨论: Hacker News 评论
-
摘要: You’re using a too-old browser You’re using a suspiciously old browser You’re probably reading this page because you’ve attempted to access some part of my blog (Wandering Thoughts) or CSpace , the wiki thing it’s part of. Unfortunately you’re using a browser version that my anti-crawler precautions consider suspicious, most often because it’s too …
-
核心评论:
- davb:Further validation of our decision to move to Debian (as if the sluggish kernel updates and ramming of snaps down our throats weren’t enough). We migrated a few hundred lab machines this summer, and have around half of our servers and containers switched over. It has gone very sm…
- Vaushite:Tools have to be allowed to evolve. Half the Linux ecosystem’s problems are caused by sticking on some design that made sense in 1980. At least the command is still called sudo or run0. Changing it to doas would have been a much more breaking change. Alas, TFA blocks me with no w…
- orf:Sudoers glob expressions are a horrible footgun and I have little empathy for anyone complaining about their removal, especially when it’s trivial to fix with 3 extra (explicit) lines.
-
深度解读: 该话题在 HN 社区引发 104 条讨论。评论焦点包括:Further validation of our decision to move to Debian (as if the sluggish kernel updates and ramming of snaps down our th;Tools have to be allowed to evolve. Half the Linux ecosystem’s problems are caused by sticking on some design that made 。
15. Gemini Omni 1.2 Flash
-
来源: Hacker News | 时间: 2026-11-06 13:35 UTC | 热度: 291 points | 评论: 216
-
讨论: Hacker News 评论
-
摘要: Gemini Omni 1.2 Flash
-
核心评论:
- wxce:It’s insane how Google has dropped the ball almost completely with AI. I literally don’t care about their models anymore because I can’t even try them without paying for them? I already pay for the plus plan but whatever their latest model was, is locked for me. Or otherwise it’s…
- msabalau:This is a new video model. It is not a replacement for Gemini 4, it is doing something altogether different. It has been available, since yesterday, within Google Flow. I know, because I used it yesterday. Creatives normally are using Flow or similar third party tools, not APIs o…
- Willingham:The rollout is so confusing. Going straight to the Gemini app and asking when I can use this model gets me vague and contradicting answers. Ultimately it sounds like the Vids rollout will take 1-3 days, and you can go to vids.new on desktop to see if you’ve won the lottery and ha…
-
深度解读: Google 发布 Gemini Omni 1.2 Flash 视频模型,已通过 Flow 面向创作者开放。社区抱怨的不是能力,而是混乱的发布策略:命名体系令人困惑,Gemini 应用内的回答自相矛盾,Vids 渠道的灰度 rollout 被调侃为『抽奖』。