从 65 条内容中筛选出 5 条重要资讯。
已抓取正文称，澳大利亚公司 Apate.ai 部署 35 万个 AI 伪装成易受骗者与骗子聊天，累计对话 250 万次，据称耗掉骗子约 17 年时间；澳大利亚运营商 TPG 将约 59 万通可疑来电转接给这些 AI；澳洲联邦银行公开报告的诈骗损失降低 76%；该公司今年 8 月获 815 万美元种子轮融资。以上为厂商说法，未见独立核实。原标题提及国内年轻人用 AI 视频反制 AI 卖惨账号，正文未给出细节。
aihot · 公众号：数字生命卡兹克 · 10月20日 00:26 · 中文阅读
谷歌官方博客宣布，Gemini 已在全球范围的 Android Auto 及内置 Google 的汽车上线，并开始推送 Personal Intelligence，博客标注日期为 2026 年 10 月 20 日。用户驾驶时可以语音免手操作提问；经授权后，Gemini 可调用 Gmail、谷歌日历、Google Photos、谷歌地图等 Google 应用信息，给出更个性化的回答。谷歌称用户可自行决定 Gemini 能访问哪些应用，并可在 Gemini 设置中随时开启或关闭。上述均为厂商说法，覆盖范围与实际体验尚待核实。
rss · Google Official Blog · 10月20日 20:00
fal 于 X 平台宣布，与 Black Forest Labs 合作推出后训练视频模型 FLUX 3 Max。fal 称该模型在第三方独立人类偏好评测中获 1139 Elo，图生视频质量与 Seedance 2.5 并列第一梯队。据 fal 介绍，模型约 12 秒可生成一段 10 秒、720/768p 的视频，未来三周将降价 75%，并继续后训练以改进质量与性能。以上评分、排名及价格信息均由 fal 公布，尚待独立核实。
aihot · X：fal (@fal) · 10月20日 17:42 · 中文阅读
AIBase 10 月 20 日 AI 日报汇总多条动态：腾讯开放元宝搜索，称索引达千亿级、接入 157 个行业信源，检索平均耗时 8 毫秒，支持 Skill、MCP 与 API 调用，每月免费 2000 次；蚂蚁百灵开源 Ming3.0 语音系列五款模型，覆盖语音合成、音频理解与对话、语音识别；MiniMax 更新 M3.1Flash，官方称速度接近前版 3 倍、成本降低 70%。
web · AIBase 中文 · 10月20日 09:09
Mistral 发布 Voxtral 3，含新一代语音转文字与语音合成模型及语音代理 SDK，厂商称支持 50 多种语言的实时多步对话与工具调用，可通过其欧盟托管 API 使用或自行部署。官网演示了代理接听法语转诊电话并完成预约。评论有人质疑该演示不现实，法国很难这么快约到皮肤科；也有开发者称已让上一代 Voxtral 模型在消费级硬件离线运行。上述未获独立核实。
hackernews · ilreb · 10月20日 16:02 · 社区讨论
以下内容已成功抓取，因评分、数量限制或同题合并未进入精选。AI 判断仅供参考，你可以继续阅读并自行挑选；各来源切换是全部列表的子集。
openai/codex released rust-v0.164.0-alpha.8
github-actions[bot] · 10-20 22:15 UTC
OpenAI 的 Codex 命令行工具发布了 Rust 版 v0.164.0-alpha.8 预发布版本。
AI 判断：标题初筛（未依据正文）：这只是开发者命令行工具的 alpha 版本号更新，标题未体现任何面向普通人的新功能或可演示效果，按技术类低分处理（仅基于标题判断）。
Release 0.164.0-alpha.8
TechCrunch：AI（RSS） · 10-20 21:25 UTC
仅从标题看，Anthropic 与 AWS、Google、微软的云合作关系出现变化，这些云厂商将 Anthropic 视为竞争者。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，该消息主要涉及云厂商与企业间的竞争关系，缺少普通人可直接使用的新功能或可演示效果，相关性有限。
据泄露给华尔街日报的 S-1 文件，Anthropic 难以从 AWS、Google 和 Microsoft 获得足够云算力，这些厂商日益将其视为竞争者而非客户。
anthropics/claude-code released v2.1.305
ashwin-ant · 10-20 20:46 UTC
GitHub 发布信息显示，anthropics/claude-code 发布了 v2.1.305 版本。
AI 判断：标题初筛（未依据正文）：仅依据标题看，这是开发者命令行工具的常规版本更新，未体现普通人的实用功能，应归为技术类并低分。
## What's changed
- Added `sandbox.allowNestedUserNamespaces` so sandboxed commands on Linux can create user namespaces (Chromium, rootless podman); this lets them get around the sandbox's file protections
- Added `CLAUDE_CODE_RETRY_WATCHDOG_USAGE_LIMIT_RECHECK_MS` so unattended retry mode (`CLAUDE_CODE_RETRY_WATCHDOG`) notices a usage limit that was lifted before its reset
- Fixed every request failing on Vertex, Bedrock and other providers when the endpoint or a gateway rejects tool search, and behind gateways with ToolSearch disallowed or left out of `--tools` (a 2.1.304 regression): Claude Code now retries without it
- Fixed `/effort` and `/autocompact` settings not applying when the model name differed from a `modelOverrides` value by letter case or a `[1m]` tag
- Fixed new worktrees in a bare repository sometimes starting from local HEAD instead of the default branch, and unchanged subagent worktrees sometimes not being cleaned up
- Fixed Bash commands on a full disk failing with a raw `ENOSPC` file path: the error now says the disk is full and that nothing ran
- Fixed an MCP server passed by a cloud session or by the SDK's `setMcpServers` being replaced later in the session by a same-named server from your own settings
- Fixed Claude naming the wrong model for a subagent's work after the subagent switched to a fallback model
- Fixed cloud sessions and teammates not telling Claude when a compaction had to leave the oldest messages out of its summary to fit
- Fixed more cases of scheduled Cowork cloud sessions waiting five minutes for each subagent whose WebFetch needed a URL approval that nobody was there to give
- Fixed Workflow resume re-running already-completed `pipeline()` and `parallel()` agents instead of serving them from the run's journal
- Fixed Remote Control leaving empty "(session)" entries in `/resume` after switching conversations with `/resume`
- Fixed `/usage` rows being cut off in narrow terminals
- Fixed startup warnings printing as raw text above the logo: they now go to the debug log, and permission rules that restrict nothing show as warning rows under the banner
- Fixed screen reader mode: search boxes now say "Search:" instead of a symbol VoiceOver read as "telephone recorder", and highlighted prompt text, such as Ctrl+R search matches and "ultrathink", no longer has spaces inside words
- Self-hosted runner: Fixed a start script that shares only its stdout with the runner or orchestrator failing with `Resource temporarily unavailable` when `TERM` is set (a 2.1.300 regression)
- Self-hosted runner: Fixed repository access failing since 2.1.296 when a git credential helper or askpass program reads its settings through `GIT_CONFIG`
- Self-hosted runner: Fixed the orchestrator reporting a spawn-runner hook as timed out when a runner it started in the background kept its output open
- Improved `/usage`: each warning is now a one-line fix that names its source, and `f` applies the fixes; the same fixes appear in VS Code and `/doctor`
- Improved MCP consent dialogs so an Enter meant for the prompt is less likely to be taken as Accept
- Improved a mod's `Image` to draw its picture in iTerm2 3.5 and later with fullscreen rendering, where it showed only its `alt` text
- Remote Control: Improved the spawn-mode messages: pressing `w` now says the new mode is saved for the folder, and a later `claude remote-control` start that uses the saved mode says where it came from
- Changed `allowedMcpServers` to stop filtering `managed-mcp.json` stdio servers that use `${VAR}` only in `env`: one your allowlist left out now loads on upgrade; use `deniedMcpServers` to keep it off
- Changed plugins that Claude Desktop mounts by path to no longer keep their `allowed-tools` when your organization's managed settings set `allowManagedPermissionRulesOnly`
- Changed the `/artifacts` delete confirmation to take 1 or Enter to keep, 2 to delete, and Esc to cancel; `y` and `n` no longer answer it
- Changed a mod's docked pane to stay within 45% of the terminal's width, including a width kept from a wider terminal
- [VSCode] Added a "Running start-up hooks" spinner status, with a screen reader announcement, while SessionStart or Setup hooks run
- [VSCode] Fixed the spinner cutting off "Compacting..." when your `spinnerVerbs` setting replaces the built-in verbs with short ones
- [VSCode] Changed the message shown when a session does not start in time and no hook is running: it now leads with the output logs and lists causes instead of pointing to sign-in and network first
- [Cloud sessions] Fixed the Cloud environments admin page wrongly refusing to save an existing environment with an "installs packages" error
- [Cloud sessions] Improved deleting a self-hosted environment so the dialog and error messages say what blocks the delete and what it will affect
- [Claude Tag] Added public Slack channels Claude hasn't joined to the Claude Tag admin Add place search, so admins can set one up from there
- [Claude Tag] Fixed Claude Tag access bundles so a repository that moved to another GitHub account can be added again
- [Claude Tag] Fixed a routine set up in a Slack DM thread no longer posting after another routine in that thread restarted the conversation; both now keep posting there
- [Claude Tag] Fixed routines set up in a Slack thread reporting success without posting after the conversation ended; those runs are now skipped and the creator is told
- [Claude Tag] Improved Claude Tag admin pages in large organizations so plugin, skill and bundle edits work while the channel list is still loading
- [Code Review] Fixed Code Review sometimes not reviewing the newest commit when several pushes landed on a pull request during a review
- [Code Review] Changed Code Review re-reviews so they no longer open with LGTM while earlier blocking findings are still open
TechCrunch：AI（RSS） · 10-20 20:40 UTC
有报道称 SpaceX 财报显示其 AI 业务表现超出预期，占公司收入的 52%。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是公司财报类消息，没有普通人可用的功能或可演示画面，而且“SpaceX 的 AI 业务占收入 52%”这一说法本身可疑、未经核实，存在不确定性。
SpaceX 发布第三季度监管文件，披露营收 146 亿美元、净利润 44 亿美元，扭转去年同期 10 亿美元亏损。其更名的“Super Intelligence”业务线（含 Cursor、Grok 和 neocloud 算力业务）表现好于投资者预期，AI 目前占公司收入 52%，算力规模达 1.7 GW。
Battle Drones - a vibe coded RTS with audio-based multiplayer
Simon Willison · 10-20 19:34 UTC
Simon Willison 介绍了一款用 AI 辅助编码（vibe coding）做出的即时战略游戏 Battle Drones，其多人模式基于音频实现。
AI 判断：标题初筛（未依据正文）：仅凭标题可知这是一个 AI 辅助编码的个人游戏项目，缺少普通人可直接上手的功能、价格与可核实细节，实用价值有限。
My latest vibe coded game experiment is a Command & Conquer style RTS (Real Time Strategy) game called
Battle Drones
(
play it here
). Unlike
my previous experiments
, this one is
actually fun to play
. I'm somewhat embarrassed to admit that I got there by telling Claude to "Make it a LOT more fun". It also supports multiplayer between nearby devices using audio to synchronize game state! Here's what I've built so far, and some of the things I learned along the way.
Battle Drones
How I built this
v1: the initial prototype
v2: full sprites
v3: varied terrain
v4: the fun update
v5: Battle Drones goes 3D
v8: Archipelago
v10: Goliath, and agent battles
Agent battles
v11: Random Maps
v12: Multiplayer via sound
Some thoughts
Battle Drones
My initial goal was to build an RTS game I could play on my phone, in the style of
Command & Conquer
. I have fond memories of the 1995 original (I liked Red Alert too) and wanted to recapture a little bit of the magic.
After 12 major iterations (and
321 commits
by coding agents, mainly Claude Code) I now have a browser-based game that works on mobile and desktop, has multiple play modes, several maps (including a random map generator) and even supports multiplayer for two devices within microphone range of each other.
Here's a video demo of Battle Drones running in a desktop browser:
Features so far include:
9 buildings, 7 ground vehicles, 6 aerial vehicles, 3 infantry units - all 3D models, which you can
browse here
Fog of war
Different weapons are needed to target aerial or ground vehicles
Six maps
, across five biomes - temperate, desert, arctic, volcanic, tropical
A random map generator
Skirmish mode
- red vs. blue on a rotationally symmetrical map
Last Stand mode
- one player holds out against as many waves of enemies as possible
Area effect weapons: Barrage, EMP, Drop Pod, Sky Hammer
Crate drops
with bonus resources
In-game music that responds to events - it gets more intense as fighting breaks out
Different in-game
AI personalities
(Balanced, Rush, Air Swarm, Boom, Skyforce, Ironclad)
AI vs. AI
mode, to watch those personalities play each other
The
Goliath
- a giant stomping mech with laser eyes and arm cannons and the ability to establish a new base
Mutators
that add extra rules or twists
A "Daily Op" challenge
Ranks and achievements (
Medals
)
Agent vs. Agent
mode where two LLMs can play Battle Drones against each other
A
multiplayer mode
based on two nearby phones communicating via audio messages, an extension of my earlier experiment with
Chirp Pong
How I built this
Battle Drones is heavily vibe coded - I've reviewed almost none of the code, and it's built in a language (TypeScript) that I've hardly used myself. I've playtested it quite a bit though, and it's been through a large number of iterations since I started the project.
Here's a rough history of the project so far, including key prompts and links to transcripts.
v1: the initial prototype
I started out with Claude Fable 5 back in July:
We are going to build a mobile friendly RTS game - command and conquer style - called Battle Drones, in which two armies of drones (both flying and ground-based) battle for territory and resources
Let's start with a MVP design, ask me some questions
Here's a
shared transcript
. I'd given Claude Code for Web an environment variable with an OpenAI API key and told it to use the OpenAI
Image Generation API
for assets, which got me some basic sprites.
You can
try v1 here
- it looked like this:
It's recognizable as an RTS, has building mechanics and the ability to select and direct units, and a basic AI to play against.
This first version was barely a proof of concept, and it wasn't any fun to play.
Claude used the
Phaser
2D game framework for these early versions. I didn't even realize this until several iterations later!
v2: full sprites
The second version (
try v2 here
) added sprites for all of the vehicles and the "AI vs. AI" mode, so I could watch the game playing without having to directly control it myself.
Claude proved reasonably competent at prompting OpenAI's image models to generate sprites. You can browse the v2 sprites in the
public/assets/sprites
folder.
v3: varied terrain
It still wasn't much fun, and the graphics were boring. It was now October, and I had upgraded to Claude Opus 5.5. I told it:
Massively improve this game. Make the graphics and gameplay better (keep it mobile friendly). Render new graphics using the latest OpenAI GPT image API (lookup details) with the API key in the environment variable. Add new unit types and balance the gameplay better. Also add new buildings and better, more varied terrain.
This resulted in five new maps with generated background tiles in a variety of styles,
visible here
. Opus also regenerated the sprites, making them much more visually appealing. You can
try v3 here
- it looks like this:
This is starting to feel like a real game! It still wasn't very fun to play though - I found myself getting bored pretty quickly.
v4: the fun update
So far my Battle Drones experiment was confirming my
previous belief
that coding agents can build a thing that looks like a game but can't actually make something that's genuinely compelling to play.
I shared this theory with someone over dinner, and then thought to myself that I should give the models a proper chance to prove me wrong. I prompted Opus 5.5 on my phone:
This game isn't fun enough yet. Make it a LOT more fun.
This time I was running in
Ultracode mode
, where Claude Code (in this case Claude Code for Web) runs "dynamic workflows", described in their docs as:
A dynamic workflow is a JavaScript script that orchestrates many
subagents
at once.
This kicked off
many hours
of work. I started it in the evening and didn't get a playable result back until late into the next morning. But the results were spectacular.
You can
try v4 here
- it looks like this:
I had Claude
write its own release notes
- here are the highlights:
Every skirmish and Daily Op now opens with a choice, gives the side that's losing a way back, and pulls the armies into the middle of the map. There are also two new ways to play.
Openers
: pick Blitz, Industry or Air Wing before the clock starts.
Call-down powers
: Barrage, EMP, Drop Pod and the Sky Hammer orbital strike.
Objectives and salvage
: Uplink relays, supply drops and wrecks worth scrap.
Feel
: combat juice, adaptive music, unit voices and an announcer.
Command
: battle alerts, game speed, saved squads and touch placement.
New modes
: Last Stand survival waves and an 8-mission Ops campaign.
Progression
: heroes, an after-action report with awards, a career, 7 mutators and a Daily Op.
Fights run longer and swing more. In 1,200 AI-vs-AI test matches on Normal, games still undecided after 25 minutes fell from 14% to 0.9%. The share of wins by a side that had been 800+ scrap behind rose from 24% to 46%.
Part of the reason Claude took so long was that it ran over a thousand simulated playtest sessions, and used the results to rebalance the existing units, add new ones, and introduce new features.
v5: Battle Drones goes 3D
After the success of Ultracode at "Make it a LOT more fun", I decided to be more ambitious:
Port this game to three.js with 3D models for all of the buildings and vehicles and landscapes
That was another overnight run. I woke up to
this v5 version
:
The new 3D models were entirely constructed by Claude in TypeScript code.
v6 (
play v6 here
) added some more visual polish. v7 (
play v7
) added robot infantry (troopers, rocket squads, snipers) and a "Condor" troop carrying aircraft.
v8: Archipelago
I was beginning to feel guilty at how little actual design input I had into the game - it really had been a case of me saying "make it more fun" and "add visual polish" for the past few releases.
I'd noticed that the AI didn't use the new infantry and troop carrying aircraft from v7 very often - there wasn't a great strategic reason to use them.
It would be neat if we had a new map with a new game mechanic that actively encouraged players to load up Condors with infantry robots and send them out into the map.
This resulted in my longest prompt yet:
Design a new map. Goal is to make different strategic incentives that encourage use of Infantry and Condors and aerial vehicles.
To do this the map will have limited harvester resources but several more uplinks - but uplinks will be on islands that are inaccessible by land vehicles or infantry. These uplinks can also not be claimed using air vehicles. The ONLY way to get them is to fly infantry there in a Condor.
Armies always start with a Condor and a single squad. If an army has no Condor a new one will be provided for free, flying in from the edge closest to their HQ and landing there. That free condor will come with a single free squad on board.
The two HQs should be entirely separated by water, but still accessible by vehicles thanks to two bridges across the map in different places
I want to encourage a battle over uplinks.
The map should be the times the size of other maps - still square
First research and then critique this design. Only build after we have agreed on any modifications
You get to design a new biome for this
The result was the tropical
Archipelago
map, with seven island uplinks that players can only capture by flying out infantry in a Condor. It's definitely the most fun map to play, once you've got the hang of the basic game mechanics from the simpler maps. The
v9 edition
made that map a little more interesting - it now looks like this:
v10: Goliath, and agent battles
I decided to add giant robot mechs to the game - think Jaegers from
Pacific Rim
, or Titans from Warhammer 40,000.
Here are some of the prompts I used to brainstorm with Claude:
Help me brainstorm ideas for adding giant robots to this game - think dreadnoughts or titans from W20K - how could we do this in a way that enhanced gameplay, kept things balanced, added unique twist? Don't overthink this yet
Let's keep brainstorming. I like the idea that it's a mobile base - that can accept harvested scrap and can have buildings built near it (even if it later moves on) and can produce infantry units too. An interaction with condor would be neat but I don't know what that could be.
Idea: as it gets damage it loses the ability to fire some weapons - by default it has eye laser (like lancers) and arm canons (like titan) and two shoulder mounted mortars, plus machine guns similar to a squad of troops
Damage knocks out mortars and then laser - it always has guns. You can't build next to a building that was only put there because of the mech, you need the mech there for off-base building and only within three squares of it
[...] Lets do two condors can carry the mech
The end result is
really cool
.
Your browser does not support HTML video.
Goliaths can be constructed using a Gantry - a building that can only be created after first building a Fab, Bay, Airpad, and Tech Lab. You can see the Goliath under construction, and it then walks out from the Gantry and can march around the map.
Goliaths can "anchor", which digs them into the ground. In their anchored state they can act as a new HQ, constructing infantry units and additional buildings in a new area.
Two Condor aircraft can hook onto a Goliath, at which point it can be flown around the map!
The game now ends when you destroy the enemy's HQ building
and
any Goliaths they might have built.
There are also a couple of mutators relating to Goliaths:
Goliath Armada
starts each side with four airborne Goliaths (each with a pair of Condors) and no other units.
Robot Love
adds a special rule: if enemy Goliaths get close to each other on the battlefield they fall in love and the game ends in a draw. (Thank Natalie for that one).
Agent battles
The other v10 feature is
agent battles
. In musing about how multiplayer could work I had an idea: what if different LLMs could fight it out in the game?
Fun extra challenge: think about how we might make multiplayer available to coding agents. It would be neat if we could set up a battle between eg GPT-6 Sol and Claude Opus 5.5 where each model has access to a tool which gives them information about the game (presumably non-visual, but only the information they could see through fog of war etc) and lets them control their units etc [...]
Consult ~/dev/llm/docs and include notes in that file about how LLM could be used here via a new tools plugin
Opus 5.5 went to work on this, producing
this AGENT-BATTLES.md proposal
and then implementing and testing it. The feature also includes the ability to record and replay these battles - here's
a replay of Claude Haiku 5.5 vs. GPT-6 Luna
, and here's that same battle
in a separate debug tool
that shows the underlying prompt responses.
The prompt that drives agent battles
can be seen here
.
There are a few more replays like this
linked in the README
. I hope to turn this into a full league table at some point in the future.
v11: Random Maps
This version exists purely because I glanced at my
Claude Usage
page one day and saw that I'd only used 75% of my weekly allowance and it was resetting in 45 minutes. I decided to see if I could get an extra Battle Drones feature from my remaining tokens, and picked "random map" as the target:
Design and implement a mechanism for generating a random map using any of the available biomes - it should be the size of the largest map and rivers and bridges and rocks and forest and scrap and crystals are all in play. Make them rotationally symmetric and use a seed to allow them to be selected or reused later. You have 45 minutes to build this! Be creative.
Here's
the transcript
, and an image Claude generated to show me its work:
v12: Multiplayer via sound
I only
just
got this working, so I've not had much time to actively test it yet. The latest Battle Drones now supports multiplayer via two devices sat within microphone distance of each other!
This feature started out as my
Chirp Pong
experiment from last week. I pointed Claude Code at that working proof of concept:
Clone Simonw/tools from GitHub and look at recent branches for the example pong game that uses web microphone to implement multiplayer between two phones via audio signals
Build a version of that for battledrones
Opus built
this Radio Probe tool
to measure how much bandwidth we could get between two devices and, based on positive results from that, implemented a full audio modem system. I've tried it between laptops, iPhones and iPads so far and it seems to work really well!
Why use sound for this? Because I
love static HTML
that costs nothing to run, and I don't want to spin up a server somewhere and then stay responsible for keeping it operating for as long as anyone might want to play the game.
Some thoughts
I know almost
nothing
about game development. This makes it a great opportunity to try out coding agents in a field where my existing software engineering experience has less impact on the results.
Games are also super low stakes. The worst that can happen is one of my silly game experiments suffers a bug.
This morning (
via Hacker News
) I found
Steam Fans Are Begging for a No-AI Filter as Steam Next Fest Is Filled With Slop
. I wouldn't personally dream of submitting Battle Drones to an event like this - it's still way too low effort to warrant consideration as a legitimate indie gaming project.
... but I can now see a path to vibe coded games like this being taken seriously. Battle Drones is fun, it looks good, it has innovative features (the multiplayer-over-audio thing is a first for an RTS as far as I can tell).
I'd love to see what someone with better taste than me and a driving passion to produce something truly exceptional could build!
A few other observations from this project:
Static HTML and JavaScript is a fantastic medium for games development. WebGL works great, and
Three.js
(and
Phaser
before it) are excellent platforms to build on that the current models know inside out.
This also lets them run as
Progressive Web Apps
- I didn't ask for it, but Battle Drones works offline. I was able to play it on a plane despite never having directly installed it - my phone had it cached.
Giving Claude an OpenAI API key and telling it to generate images using OpenAI's models works really well. It can view the images it has generated and iterate on the prompts.
Opus 5.5 is spectacular at 3D modeling using Three.js (and Blender, and no doubt other 3D applications as well). The
same is true
of GPT-6 Astra. They are absolutely capable of building models for 3D game projects.
For features like replay and multiplayer it's useful to have a deterministic simulation engine. At one point Claude had to replace calls to
Math.sin()
and
Math.cos()
and friends with
custom versions
because different browsers (Chrome vs. Safari, for example) returned very slightly different numbers, desynchronizing the replay.
Opus 5.5 can write music! There are no audio files in this project - the soundtrack is entirely generated through code, varies across different maps, and even gets more intense as the battle heats up. See
Scrimshaw Jukebox
for a previous experiment with Claude for music composition.
I still can't get over how well the prompt "Make it a LOT more fun" worked.
Tools are effectively free. I prompted up the
Radio Probe
, the
Model browser
, the
Agent vs. Agent debug tool
, and Claude Code
built a bunch of tools
for itself that I haven't even explored myself.
Audio-based multiplayer between static HTML games works! I hope to see a whole lot more games use this same pattern - in digging around today I've not found much prior art beyond
Siner
by Alon Horesh and
ggwave
by Georgi Gerganov (also of
llama.cpp
fame).
Claude Code "Ultracode" burns tokens like there's no tomorrow, but it can also solve complex problems (like adding audio-based multiplayer) from a single prompt.
I'm still not quite over the
guilt
of building software in this way. I don't feel like I've earned the right to create something this good, in a field in which I have so little prior experience! At this point I've shown it to enough people in private that it felt wrong not to publish it for everyone else to see and learn from.
Tags:
game-design
,
projects
,
ai
,
generative-ai
,
llms
,
claude
,
vibe-coding
,
coding-agents
,
claude-code
Grok Bot's IncidentWatch & IncidentSweep Automate Support Queue Incident Detection
Future Tools · 10-20 19:00 UTC
据标题，Grok Bot 的 IncidentWatch 与 IncidentSweep 可自动检测支持队列中的事件。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这是面向支持或运维团队的自动化事件检测，属于技术向选题，普通用户关联弱且缺少可验证细节。
Tools & Products · 4h ago Grok Bot's IncidentWatch & IncidentSweep Automate Support Queue Incident Detection xAI has published details on two Grok Bot AI teammates built to automate support queue incident detection and resolution. IncidentWatch scans tickets every five minutes, flagging spikes in similar reports before internal engineering alerts fire. IncidentSweep handles post-incident cleanup, identifying affected tickets and closing them in approved batches. During a September 18 incident, IncidentSweep cleared 204 tickets in roughly 10 minutes, a task that previously required hours of manual replies or imprecise bulk actions.
SimCity, but with Permits
pcai · 10-20 18:29 UTC
标题显示这是一个把《模拟城市》与许可证审批结合的趣味项目，具体内容和 AI 关联未知。
AI 判断：标题初筛（未依据正文）：仅凭标题看，它未体现 AI，也无法确认普通人的实用价值或可演示效果，因此暂按低优先级趣味项目处理。
--- Top Comments ---
[bruce511]: I suppose this is the result of a hundred years of democracy, aka having to keep everyone happy all the time. Municipal rules, zoning, community feedback and so on all operate in the space of keeping people happy. No-one wants noise, so we have noise restrictions. Oops, construction is constrained. We all want no traffic, and lots of parking, so low density please, but that creates housing shortages and limits the usefulness of public transport. So a rule is added here, a process there, and e...
[zabriel_goss]: The original EPCOT as imagined by Walt Disney was a master-planned city, with all the pleasant amenities and cleverest design you could imagine in the 1960's: Walkable areas, clean public transit, no traffic congestion, homes furnished with the latest in technology and appliances. It was his greatest ambition at the end of his life to make it a reality. But it didn't happen. Namely, because of how much central control would be given to one man and/or his company to pull it off....
[ryandrake]: Kind of like this one[1] which made the rounds on HN a few weeks ago. Looks like the social commentary from them are similar: "Real life is worse than SimCity because of rules and the public getting a say." and/or "Why can't we in real life let the mayor use the Bulldozer Tool to just wipe out any property he doesn't like and put up whatever property he likes in its place?" 1: https://news.ycombinator.com/item?id=50036864
[BatmansMom]: A little too much going on all at once in the beginning. For a complex game like this I think you have to start small and slowly introduce functionality. Why are so many buttons available to me from minute 1? I'm not using them.
[goda90]: The various decisions/petitions that pop up can break the immersion by mentioning details that don't fit the city(petitions with signatures from >250% of the town population, traffic issues at the "4-way intersection in front of the feed store" when no such place exists, or the new building causing shadows when there's nothing around it). It would be great if more text was turned into templates that pull game data for accuracy, and if the simulation were able to fe...
OpenAI acquires Orchestra Bio to accelerate AI-driven drug development
Future Tools · 10-20 18:00 UTC
OpenAI 宣布收购 Orchestra Bio，旨在加快 AI 驱动的药物研发。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这是一则企业收购新闻，缺少面向普通人的具体用法或可演示效果，实际相关性有限。
Company News · 5h ago OpenAI acquires Orchestra Bio to accelerate AI-driven drug development OpenAI has acquired Orchestra Bio, a platform that helps biopharma teams coordinate drug development workflows. The acquisition merges Orchestra's technology with Rosalind, OpenAI's existing life sciences suite that includes GPT-Rosalind and Rosalind Workbench. Together, they aim to build AI agents that connect scientific findings with studies and decisions needed to advance medicines. Drug development typically takes 10 to 15 years from target discovery to approval, and OpenAI sees AI as key to compressing that timeline for partners like Novo Nordisk and Moderna.
What shall I teach my son?
Curiositry · 10-20 17:25 UTC
标题询问作者该教自己的儿子什么。
AI 判断：标题初筛（未依据正文）：仅凭标题无法看出与 AI 相关，缺乏可演示的教程、新功能或可核实信息。
--- Top Comments ---
[PeterWhittaker]: If lifestyles support it, and parents are so inclined, I'd worry less about teaching the first few years and more about reading to them and with them, esp. with some silliness and improvisation, hiking, camping, and shopping as a family, especially at local markets. Spend meaningful, loving time together, and out in the world, as much of the small scale natural world as you can. Listen to their questions, answer when you can, learn the answers together when you can't. If they are ac...
[fxwin]: > My economically valuable skillset – programming – suddenly feels less valuable. What shall I teach my son? Maybe that's just me, but this strikes me as a bit odd. When i think about my parents and what i appreciate about them and how they raised me, hard skills are not very high up the list. Things like encouraging curiosity, agency and freedom to pursue things that I showed interest in even if it meant spending many weekends and many eurodollars on me that they would no doubt have ...
[dijit]: My mum was a mechanic, my dad was a mechanic/truck driver. My grandfather and uncle were mechanics... I am not, in fact my understanding of mechanics is extremely basic. If I go back into my family tree, I can find three generations of Sugar House Labourers in Scotland in the early 1800s.. a profession that no longer exists in the UK as far as I can tell, and I'm certain they had a specialisation. We'll be fine, your children aren't you, they'll find their own passion...
[weirdhacker42]: Hear, O my son, and receive my sayings; and the years of thy life shall be many. ¹¹I have taught thee in the way of wisdom; I have led thee in right paths. ¹²When thou goest, thy steps shall not be straitened; and when thou runnest, thou shalt not stumble. ¹³Take fast hold of instruction; let her not go: keep her; for she is thy life. ¹⁴Enter not into the path of the wicked, and go not in the way of evil men. ¹⁵Avoid it, pass not by it, turn from it, and pass away. ¹⁶For they sleep not, excep...
[ThrowawayR2]: > " ...Bill Gates were college dropouts, but the market deemed them good. " Gates had excellent grades and was credited in an academic paper during his undergraduate years. Not exactly like the average dropout. > " Now that I think of it, apparently Einstein was a mediocre student, but his contributions to science were his way out.
" That's a myth. Einstein got top grades in science and mathematics. As for the original question, " If you wish to build a ...
Claude Designs 17 Protein Binders Against Scorpion Neurotoxins in 64-Hour AI Campaign
Future Tools · 10-20 17:00 UTC
据标题，Claude 在一项 64 小时的 AI 行动中设计出 17 种针对蝎子神经毒素的蛋白质结合剂。
AI 判断：标题初筛（未依据正文）：标题显示为专业蛋白质设计研究，普通人难以直接使用或演示，偏技术类，且仅凭标题无法确认实际效果与可用性。
Research · 6h ago Claude Designs 17 Protein Binders Against Scorpion Neurotoxins in 64-Hour AI Campaign Anthropic's internal Claude research model designed protein minibinders against scorpion neurotoxins in a hour autonomous campaign, producing 17 confirmed binders across four dangerous toxins: Cn2 (Mexico), Ts1 (Brazil), AaH I and AaH II (North Africa), and LqhIII (Middle East). Working from one initial prompt and four follow-ups developed with Citrini Research, Claude generated 318,484 designs using 12 tools and ~10,000 GPU hours. Partners Adaptyv Bio and Twist Bioscience lab-tested 211 designs, confirming 17 met binding criteria. All 211 designs are being released publicly.
Google open-sources TorchTPU, a native PyTorch back end for TPUs
ilreb · 10-20 16:40 UTC
谷歌开源了 TorchTPU，这是一个让 PyTorch 原生支持 TPU 的后端项目。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向开发者的框架后端开源项目，与普通人的工作生活场景没有直接关系。
--- Top Comments ---
[pcmoritz]: For people interested in this: I've been working on an open-source libtpu implementation for https://tenstorrent.com/ accelerators as a side project, it can already run most of torchtpu, see https://github.com/pcmoritz/libtt (based on the tenstorrent software stack). It shows all the things involved in lowering torch to tile based accelerators.
[npalli]: Interesting, Amazon doing the same for AWS Trainium. https://github.com/aws-neuron/torch-neuronx
[dalemhurley]: Google and Apple are going to move consumers and ordinary business users to edge inference. This is more disruptive to the upcoming IPOs than most investors understand.
[fancyfredbot]: Cool! XLA is pretty good + potentially a better compiler infrastructure than inductor which will lead to improved performance on all hardware. Also I guess this is the final nail in the coffin of Tensorflow.
[ismailmaj]: Amazing news, I remember the days when you had to fight XLA to make PyTorch run on TPUs, it seems we finally got first-party support! Since HW specifics live mostly in the kernel layer (Pallas or CUDA), I imagine the Jax vs PyTorch question is more of a style preference now?
X：Aravind Srinivas（Perplexity CEO） (@AravSrinivas) · 10-20 16:34 UTC
据标题所述，Perplexity CEO 表示 Perplexity Computer 现在可以在 Mac 上完全本地运行，不需要云端推理。
AI 判断：标题初筛（未依据正文）：仅凭标题无法核实硬件要求、可用范围和实际体验，内容偏向本地推理部署，普通用户能否直接受益尚不明确。
Perplexity CEO Aravind Srinivas 宣布，Perplexity Computer 现可在 Mac 上完全本地运行，无需云端推理，也不用 WiFi。模型和智能体都在本机运行，可操作 Mac 原生应用和本地文件，敏感数据不必发送到云端，现已在 Perplexity Mac 应用中可用。
PyTorch：Blog（RSS） · 10-20 16:00 UTC
据标题信息，Google 与 Meta 联合发布了 TorchTPU，为 PyTorch 提供原生的 TPU 后端支持。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向开发者的框架与硬件后端工程进展，普通用户既无法直接使用也缺乏可演示效果。
Google 与 Meta 发布开源项目 TorchTPU，为 Google TPU 提供原生 PyTorch 后端，模型迁移只需把 .to("cuda") 改为 .to("tpu")。
Google Gemini Launches Business Hub With Financial Insights and Brand Tools for SMBs
Future Tools · 10-20 16:00 UTC
标题显示 Gemini 为小企业主上线了包含财务洞察和品牌工具的 Business Hub。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向小商户的新功能发布，可能具备实用演示价值，但具体功能、价格与地区开放信息尚未知，需谨慎核实。
Tools & Products · 7h ago Google Gemini Launches Business Hub With Financial Insights and Brand Tools for SMBs Google has launched a dedicated Business Hub within the Gemini app, offering small business owners a central dashboard to manage branding, finances, and customer insights. New features include on-brand visual generation powered by the Nano Banana 2.1 image model, real-time financial analysis that converts spreadsheets and receipts into interactive graphics, and integrations with Google Business Profile, Google Ads, Canva, Shopify, Xero, and Adobe. The tools are rolling out globally for all Gemini app users in the coming days.
Ask a Scientist: How can AI help doctors spend more time on collaborative care?
Google Official Blog · 10-20 16:00 UTC
文章以“问科学家”形式探讨 AI 如何帮助医生把更多时间用于协作护理。
AI 判断：标题初筛（未依据正文）：标题仅显示医疗 AI 临床研究问答，未提供普通人可直接使用或验证的功能细节。
Two Googlers in a virtual interview with text to the left of them that says 'Ask a scientist about patient care research'
Google Official Blog · 10-20 16:00 UTC · 同链接合并评分
标题显示 Gemini 为小企业主上线了包含财务洞察和品牌工具的 Business Hub。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向小商户的新功能发布，可能具备实用演示价值，但具体功能、价格与地区开放信息尚未知，需谨慎核实。
"Your brand, your tools, all in one place" banner featuring three feature highlights: "Create image styles for your brand," "Customize Gemini with apps and skills," and "Create content with templates."
Sixteen Years After Deepwater Horizon, Deep Sea Life Shows Few Signs of Recovery
speckx · 10-20 15:59 UTC
耶鲁环境360的一则摘要指出，深水地平线漏油事件过去十六年后，深海生物依旧几乎没有恢复的迹象。
AI 判断：标题初筛（未依据正文）：该选题为海洋生态新闻，与 AI 无关，且仅有标题信息，不构成面向普通人的 AI 教程或趋势解读。
--- Top Comments ---
[WarOnPrivacy]: In 1979, Ixtoc I released 3.3M barrels into the Bay of Campeche (DwH released 4.9M). By the time DwH happened we had already learned that exposed, high-energy habitats bounce back;
sheltered and deep-sea habitats do not.
ref: https://www.frontiersin.org/journals/marine-science/articles... ref: https://beneaththehorizon.org/ixtoc.html
[jsbisviewtiful]: It’s truly exhausting to know the damage we’ve done to the ecosystem and how many bad actors only concerned with profit are out there preventing it from happening again.
[superxpro12]: The Deepwater disaster recovery very effectively demonstrated the phenomenon of "out of sight out of mind". they dumped almost 2 million gallons of "dispersants" on the oil spill... which was DESIGNED to make the oil heavier than water and sink to the sea bed. it never dissolved anything, it just hid the problem from the public. and all that oil is just sitting on the sea bed.
[darth_avocado]: > While researchers stopped short of concluding that the oil spill had caused the declines Why though? Why are we so afraid of accountability? Because this is what people point to when trying to deny things are true: ThErE iS NO EviDeNce, sHoW Me THe StUdY, etc. etc.
[hisfraudulency]: Interesting that the spill hasn't dampened long term prospects for offshore drilling (besides the brief push against it around 2016) whereas Chernobyl basically nerfed American nuclear power gen.
30 Months on the Exponential (in Charts)
One Useful Thing · 10-20 15:57 UTC
这篇文章用图表回顾过去 30 个月 AI 的指数级进展。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这是图表型趋势回顾与观点，缺少可亲自演示的教程或普通人可立即使用的具体新功能。
My new book is out today... what happened since the last one
IKEA – 3D Guide
punnerud · 10-20 15:42 UTC
标题显示这是一个关于 IKEA（宜家）的 3D 指南，但仅凭标题无法确认其内容是否与 AI 相关或能否实际使用。
AI 判断：标题初筛（未依据正文）：仅凭标题无法确认 AI 属性、可用性与演示价值，暂不适合作为普通人 AI 选题。
--- Top Comments ---
[dag11]: It's really interesting despite others pointing out that it's a little wrong, that can be fixed. But I also almost left immediately because this claudism at the top really hurt to read: > Every dowel, cam lock and nail where the page puts it. I always feel so manipulated when I read it regurgitating the prompt you gave it in such an advertisey way, like it's trying to assure you it met _every_ word of the prompt, regardless of how it actually did.
[sigy]: It instructs you to build the shelf wrong. The camlocks go on the bottom of the shelf.
[LordHeini]: The Lack table is probanly the esayest piece of Ikea furniture to build.
But Claude or whatever llm seems to assume that you can screw the screws in fully by hand. But you just put them in a little and use the leg for the rest. As clearly shown in the instructions.
[aqfamnzc]: What is the name of this website style that Claude et al loves to build so much? The light text on dark blue background, pill-shapes everywhere, pulsing indicators, etc.?
[geminiboy]: Built almost the same idea almost a month back. Never posted it https://flatpack3d.nithen-ac.chatgpt.site
Facebook Pages require paid subscription to post more than two links per month
AshleysBrain · 10-20 15:06 UTC
标题称，Facebook 主页若每月发布超过两个链接，就需要付费订阅。
AI 判断：标题初筛（未依据正文）：基于标题判断，该内容与 AI 无关，且缺少价格、地区与规则细节，难以核验和演示。
--- Top Comments ---
[belak]: I wonder if this is an effort to combat pages which post a short clickbait description and require you to click "read more" to see the full description, and often post the link in the comments. There are lots of sites (Polygon immediately comes to mind) which are notorious for doing this as a way to get fake engagement which boosts their posts which gets more clickbait engagement. I know it's all speculation, and this does come across as more monetization creep, but I can'...
[josefresco]: As someone who helps small businesses create their own websites... thank you Facebook! Seriously though, back when "Facebook Pages instead of websites" were all the rage, FB demoting page content in feeds officially killed that movement and I was similarly happy.
[meerita]: I quit Facebook in 2012. I cannot imagine what's there but in 2019 I had to create a new account to test some development stuff, and checked friends and family and it was pretty much dead.
[RegW]: So is this a kite for one way to commercialize all of Meta's platforms? Or perhaps just a way to kill off a dead business under the guise of a failed experiment?
[rickcarlino]: In the 2010s, there was a move away from files in the name of convenience (lock-in). The 2020s are a war on hyperlinks. It is particularly unfortunate to see this masquerading as a concern for safety and security (“everything outside of our platform is scary and insecure”)
Perplexity Computer Now Runs Fully Local on Apple Silicon Macs for Pro Users
Future Tools · 10-20 15:00 UTC
Perplexity 表示，其 Computer 现在能在 Apple Silicon Mac 上为 Pro 用户完全本地运行。
AI 判断：标题初筛（未依据正文）：标题显示这是面向 Perplexity Pro 与 Apple 芯片 Mac 用户的本地运行新功能，实用但受众受订阅和设备限制，且仅凭标题无法核实具体效果。
Tools & Products · 8h ago Perplexity Computer Now Runs Fully Local on Apple Silicon Macs for Pro Users Perplexity has launched Computer fully local on Mac, allowing Pro, Max, and Enterprise subscribers to run AI model, task planning, and execution entirely on Apple silicon devices without sending data to the cloud. The feature works across local files and native Mac apps including Notes, Messages, Mail, and Numbers. It runs Qwen 3.6 or PPLX Pro models and requires 36 GB or more of RAM. Users can also use Hybrid Compute to combine local processing with cloud-based frontier reasoning when needed.
Ofcom concedes defeat in 4chan case
iamnothere · 10-20 14:30 UTC
英国通信管理局（Ofcom）在4chan相关案件中承认败诉。
AI 判断：标题初筛（未依据正文）：标题仅涉及监管诉讼，与AI无关，对普通人缺乏可演示或可用的AI价值。
--- Top Comments ---
[Aurornis]: This was such a weird case. Usually when a country goes after a foreign business they enforce it by imposing restrictions within their own borders and threatening the operators with legal action if they ever enter the country. If the operators don’t visit the country and their banking doesn’t cross a business that must follow those laws (which is hard), it’s otherwise symbolic. I may be reading too much into the editorialized writeups about Ofcom but even their statements imply they thought t...
[simoncion]: In an ideal world, exclusively-US-based companies that host message boards, chat rooms, instant messaging services, and the like would look at this saga and 1) Notice that the odds that a foreign power is going to be able to harm them in any meaningful way for violating their "child 'safety'" laws that curtail freedom of speech and association are very low 2) Notice that 4chan is fighting harder to resist these attacks on free speech and association than many of big na...
[cjs_ac]: This fits quite neatly with Dominic Cummings' comments to the House of Lords Constitutional Committee[0] about how the British Civil Service can't do anything because no one with any talent wants to work there. [0] https://www.youtube.com/watch?v=KBBYvj4YQLE (full video is an hour and a half, but you can skip to any bit because he just makes the same handful of points over and over again)
[mm263]: This is just laying down the ground work for Russia-style deep packet inspection. “We don’t have a capability to censor the foreign internet, so we need measures to cut off the violating sites from our side”
[miohtama]: Think of the kids the world we failed to protect by not having passport checks and government snitch at the 4chan login. But don’t worry! We will check your passport when you open your phone (already live in the UK, efforts supported by the GBP 2M secret donation of Meta to NGOs). And what’s the worst thing that could happen? No one would for sure arrest you for being critical of a foreign government. https://www.independent.co.uk/news/uk/crime/palestine-action...
My Friend Bob Cringely
zinxq · 10-20 14:13 UTC
这是一篇讲述作者朋友 Bob Cringely 的文章或个人故事。
AI 判断：标题初筛（未依据正文）：仅标题判断：内容像人物回忆或个人故事，未显示面向普通人的 AI 新功能、工具或可核实 AI 新闻，相关性很低。
--- Top Comments ---
[js2]: Mark Stephens (aka Robert X. Cringely) obituary: https://www.hillandwood.com/obituaries/Mark-Stephens?obId=49... Tell HN from two weeks ago about Bob's death: https://news.ycombinator.com/item?id=49949438 Plane Crazy documentary: https://www.youtube.com/watch?v=FSNzF-KSGiw (I watched it for the first time last week. Very entertaining.)
[phlipski]: Paul Tyma - 3 posts in 20 years. A prolific blogger! :)
Nice post. I read Cringely's blog for 20+ years. I think Paul hit the nail on the head. I think the problem with Cringely's ventures is he never seemed to have the discipline to hang on long enough to pivot and find product market fit.
[thelastgallon]: Paul Tyma is the creator of Mailinator. I remember reading about how he made it scalable with Java.
[1potato]: The Cringely story was great. the Google story, also great. I click over to to the home page, ready to settle in for some more interesting reading, and -- there's a total of three posts. Oh well.
[ironchief]: He messaged me a couple years ago and it matches with this post's description: "While the first half of the email was about the idea, invariably, the 2nd half was about how much money it was going to make." Point-to-point supersonic organ delivery
Dear Derek,
I'm Bob Cringely, an ancient Silicon Valley pundit. I also run a little
CubeSat air-launch startup. Competition from SpaceX forced us to cut costs
so we figured how to make an old French business jet fly...
Spotify Removes 50M AI Spam Tracks, Launches Tools to Protect Human Artists
Future Tools · 10-20 14:00 UTC
Spotify 宣布移除 5000 万首 AI 灌水歌曲，同时上线工具以保护真人音乐人。
AI 判断：标题初筛（未依据正文）：仅凭标题可知这是平台治理 AI 灌水内容、保护真人创作者的动作，与普通听众和音乐人相关，但标题未提供工具细节，实际效果待核实。
Company News · 9h ago Spotify Removes 50M AI Spam Tracks, Launches Tools to Protect Human Artists Spotify has removed over 50 million AI-generated spam tracks from its recommendations this year, and more than 75 million spammy tracks from its platform entirely for rule violations. The company uses listener feedback, upload patterns, metadata, and third-party AI detection tools to identify slop. New protective tools include AI Credits for disclosed AI use, Verified by Spotify badges covering 300,000 artists, AI Persona labels, Artist Profile Protection letting artists approve releases, and a 1,stream royalty threshold that disadvantages AI-generated content.
EMudhra Intermediate CA Compromise
galadran · 10-20 13:58 UTC
标题显示 EMudhra 的中间证书颁发机构可能遭到入侵，属于网络安全与证书信任相关事件，与 AI 无直接关系。
AI 判断：标题初筛（未依据正文）：仅凭标题无法确认影响范围，且这是技术安全事件，不适合作为面向普通人的 AI 选题。
--- Top Comments ---
[eqvinox]: Since I haven't seen this mentioned enough yet: The private key for all 8 "compromised" CAs is a single byte: https://mastodon.catgirl.cloud/@49016/117474217042191790 One of the private keys is literally "4". I'm putting "compromised" in quotes here, because I can't call it that without cringing and cursing. [cf. also: https://news.ycombinator.com/item?id=50169109 ]
[galadran]: The impact of this incident is very unevenly spread. Over 10 years ago, Mozilla shipped OneCRL [1]. We remediated this incident in a couple of hours for Firefox users who were protected automatically without the need for updates. Chrome, Apple and Microsoft have similar tooling. Since then, we've built more advanced tooling to handle revocation at scale for leaf certs (CRLite [2]). However, for the rest of the ecosystem, there has been little to no investment in revocation and for those ...
[tech234a]: Explanation from eMudhra in the Mozilla bug report https://bugzilla.mozilla.org/show_bug.cgi?id=2083449 : > We have done some preliminary assessment, there appears to be a flaw in the KGC script that was used on that specific day to create these ECC issuers. As for these ECC keys, our review of the KGC, assessment of the code and view of HSM keys so far indicate that a) a set of original ECC keys were created onboard as expected b) an approved deviation to our usual KGC to...
[hypfer]: From what the internet is telling me, the keys were allegedly found on https://repository.emsign.com/certs/<redacted>.crt Which, if true, is definitely a lot of fun. Hope you people are ready for lots and lots of suddenly hackable IoT. __ Edit: I might also be misinterpreting what I am seeing and it's actually weak logic that created the keys, in the sense that one could guess the private key? Someone actually qualified please give credible answers.
I'm jus...
[lamonade]: I set up a server that responds with a certificate to any host you request. You can use it to see if curl accepts the certificate or not. curl https://example.org --connect-to example.org::2.29.126.86:
2026 is an exciting year for the Rust compiler (and me)
panstromek · 10-20 13:53 UTC
一位开发者撰文展望 2026 年 Rust 编译器的发展方向。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向开发者的编译器路线展望，与普通人的办公、学习和生活几乎没有关联，属于纯技术内容。
--- Top Comments ---
[KolmogorovComp]: I don't think rust is the future anymore, despite having contributed to its compiler few times and been quite involved for years. It has spread too much and that impact compiles time too much. I think a newer language, having the boreness of go, the compile speed and maybe comp time of zig and safety of rust will emerge, maybe with lisp syntax.
[throwup238]: > It's easy to understate how big of a deal this is for the compiler. This work resolves a major piece of technical debt that a lot of compiler developers dreamed about for years. No kidding. I’ve been waiting for the solver at least half that decade now. Hopefully it and Polonius unblock specialization which has been in purgatory for years.
[yhjc2692]: Smaller target directory size would be such a lifesaver. Hoping the best for you, and Rust!
[infogulch]: I'm curious about polymorphism. Should it be default on for debug builds just to save storage and compile time? How will rust decide whether to enable polymorphism? Will it be possible to provide source-level polymorphism hints to the compiler?
[panstromek]: I don't understandy why, but all submissions of this post so far had incorrect title. When I posted the correct one, it still got reassigned the wrong title and I had to edit it (and it didn't allow me to do that yesterday, only today).
openai/codex released rust-v0.164.0-alpha.7
github-actions[bot] · 10-20 13:47 UTC
GitHub 上 openai/codex 发布了 Rust 版本 v0.164.0-alpha.7。
AI 判断：标题初筛（未依据正文）：从标题看这是开发者命令行工具的预发布版本更新，普通用户不适用，也未提供可演示的实用功能。
Release 0.164.0-alpha.7
How I got over 300 Tor exit nodes removed
kabakabakaba · 10-20 13:39 UTC
一篇个人记录，讲述作者如何促成 300 多个 Tor 出口节点被下架，内容与 AI 无关。
AI 判断：标题初筛（未依据正文）：仅从标题判断，这是网络安全／运维类的个人经历分享，与 AI 及普通人的办公、学习、创作场景无关，不具备可演示的 AI 价值。
--- Top Comments ---
[phw]: I built exitmap over ten years ago. Glad to see it's still of use :)
[jandrese]: The attack on SSH seems rather poorly conceived. Not only was it detected immediately thanks to SSH printing that huge banner on MITM attacks, but it allowed the admins to easily trace every infected node and respond quickly to shut the whole operation down. Had they stuck to email they would still be running today.
[96718923545809]: Thank you for writing this. > But if you ignore the warning and go ahead with the login, the attacker does three things:
> 1. Steals your password This shouldn't affect key-based authentication, though Also, this from Marc-Etienne M.Léveillé email linked in your post: > Another possibility is a compromised hosting provider. We identified a
couple of them in the past where their management servers were compromised
by Ebury and their customers' VPSes, containers, etc., ended...
[jszymborski]: 3000 is insanely few exit nodes.
[privacyishuman]: Tor guard nodes seemed to be under attack but the devs aren't really engaging with it publicly, at least. It happened to mine and I had to switch it to a middle relay. No problems now. https://forum.torproject.org/t/tor-relays-shutting-down-my-4... People say that even running a middle or guard node can get your IP blacklisted. I don't have personal experience of this because I don't use my home IP for regular use, just for selfhosting. All the other devic...
Tsrs: A Port of TypeScript to Rust
hnisjafx40 · 10-20 13:18 UTC
标题显示，Tsrs 是一个将 TypeScript 移植到 Rust 的项目。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向开发者的语言移植项目，普通人难以直接使用或演示。
--- Top Comments ---
[DanRosenwasser]: Hey there, TypeScript team member here. The tsrs project has been extremely interesting to see come together. We've been working with the maintainers to better understand how tsrs is fast and what sorts of optimizations it uses are safe. We even did our own Rust vibe port and what we currently suspect is that a lot of the speed does not just come down to language choice, and some of what is going to make TS 7.1 faster and leaner (see the charts in the blog post) actually comes from optim...
[austin-cheney]: I use the Typescript compiler to report errors to the shell, but that's it. Both Bun and Node.js will execute Typescript directly without a compile step. If you are able to write your own front-end bundler then the front-end code also gets the same type stripping from either Bun or Node.js. Since both Bun and Node.js start up my large applications about twice as fast as the Typescript compiler executes I don't see any reason to bother.
[afdbcreid]: It appears they're using a unique memory model ( https://github.com/oxc-project/tsrs/blob/main/docs/PORTING.m... ) to overcome the challenge that prevented the TypeScript team from choosing Rust to begin with: Rust has no GC and porting from a GC language to it often requires restructure of data. Essentially, they fake a GC via never-freed arenas for long-living objects. It's interesting that Rust can do that (hopefully soundly)!
[sheepscreek]: Ah took me a while to understand the reason behind this. It wasn’t “let’s build a faster Rust based Typescript compiler” at all. That was a nice side-effect. They needed a Typescript compiler that can run on their WASM only sandbox. So the only existing option was to export tsc Go to WASM, but that comes with the full Go runtime. And to avoid that cost, they experimented with a Rust compiler and turns out it’s even faster than the reference tsc/Go implementation.
[egonschiele]: It's exciting to see the speed gains. I use TypeScript every day on a large codebase, and it is too slow to run locally.
I don't love this change though: > However, frontier AI models can now do large-scale ports almost independently in a matter of days > I wrote zero code. Not "barely any". Zero So we are going to move from a carefully handwritten port to an entirely AI-generated port. I know this is the way the industry is headed, I use coding agents a ton. But I s...
Publish fake stats, get cited in LLMs
robtherobber · 10-20 13:06 UTC
一篇博客标题称，发布虚假统计数据可以让内容被大语言模型引用。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这是关于用虚假统计数据换取大模型引用的观点文章，普通人直接实用性有限且有待核实。
--- Top Comments ---
[Amorymeltzer]: I'm reminded of a section of Cory Doctorow's latest non-fiction, The Reverse Centaur's Guide To Life After AI (it is of course an excellent read), where he discusses how the semantic web and having an internet set-up for agents and AI to use won't work: >It also hasn't happened because there is a reward for cheating. Scammers will label their phone numbers as the customer service number for your cable operator or bank. Counterfeiters will label their batteries ...
[PreciousH]: This problem is so bad that you can create a website just to review your own website so when LLMs want to pull out reviews for the website the fake reviews count
[Lerc]: It is particularly bad when trying to find the reality behind common misconceptions. Even when there is no evidence to support something, and no motivation for anyone to promote the falsehood, models can build an ediface out of false claims. There are always people prepared to invent a story to explain why something is true. Just so they can sound knowledgeable. They aren't trying to perpetuate a lie, they are just under the same misconception as others, quite a lot of them will believe...
[acadapter]: This has existed for a long time before LLM, although in a slower form. On Wikipedia it's called "citogenesis".
[ShadowOfThePit]: This is insane . How can this be fixed? Nothing prevents you from publishing straight BS, and LLMs will use any site that doesn't have a garbage reputation as a valid third-party source. Of course, LLMs already use Reddit, Facebook and co. as a source, so it doesn't really matter...
How I built my ‘everything’ app
Ben's Bites · 10-20 13:02 UTC
一篇作者自述如何亲手做出自己的“全能”应用的个人分享。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是个人开发应用的经历分享，偏向开发者话题，对普通人的具体用处和可演示效果都不明确。
AI can customise your home devices, like your TV
The Decoder：AI News（RSS） · 10-20 12:32 UTC
据 The Decoder 报道，美国陆军网络司令部的一个名为 ECHOTRIBBLE 的项目设想把 AI 智能体变成可自我扩散的网络武器。
AI 判断：标题初筛（未依据正文）：仅凭标题可知这是军事与网络安全方向的设想，缺少普通人可验证、可使用的具体信息，只适合做趋势或风险视角的讨论。
据 The Intercept 获取的文件，美国陆军网络司令部（ARCYBER）正在开发代号 ECHOTRIBBLE 的平台，让 AI 智能体自主在网络间移动、识别并清除威胁目标，之后自毁，且须在极少人工监督下运行。
Apple broke an essential presentation feature of iPads
evanb · 10-20 12:17 UTC
标题称苹果破坏了 iPad 的一项关键演示功能。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这是苹果 iPad 演示功能问题，并非 AI 选题，且缺少可演示细节。
--- Top Comments ---
[btown]: So I could see the other argument here: I’m presenting from a presentation, I want to switch to a live demo environment, and to do so I need to turn off the presentation entirely? Now, another operating system might see this tension and say: why not have a UI element come up, and ask the user whether they want to share their entire screen or stay on the existing shared window? Here, I think this hits a fundamental weakness in Apple’s design culture: if you want to make things magical by remov...
[whywhywhywhy]: The endless struggle of professional iPad users trying to get their hardware to act like something it insists it is not. Ultimately the team behind it doesn't want to relinquish control to the user, so we're stuck in this world where it can never truly by a professional tool because every step of the way there is something like this where on a real computer its just an extra screen and you put on what you want on it and the situation where content of the other screen gets suddenly s...
[abujazar]: Apple needs to get their software design on track again and do proper QA and testing on their software before making these kinds of changes. 27 was supposed to be a quality improvement snd bugfix release, but it's mostly been a cascade of bugs and UI failures.
[e40]: I had heard 27 was much better than 26, so I broke my usual embargo on new macOS versions until x.y.5 and upgraded. I also had a problem with the 26.6.7(?) update that came out the same day as 27.0, so I went for it. Big mistake. I regret it.
I also regret updating my homepods and ATV to 27.0. I had lots of phantom clicks on the tv remote. Turns out it was the homepods. When it happens I have to restart them to fix it.
Never again. I will be on 27.x.y until 28.6 is released.
[randusername]: I see comments on posts like this complaining that Apple has the money to make software as high-quality as their hardware so they must be dysfunctional. I'm not sure it's so easy. I think you have to compromise on hardware or software and Apple makes a choice. My whole career in embedded devices, software is always under schedule pressure from hardware. Not because hardware is sloppy or evil, but because doing new things is hard and hardware has deadlines that software doesn't....
Introducing Eleven Realtime: a simpler way to build natural voice agents
ElevenLabs Blog · 10-20 12:00 UTC
ElevenLabs 发布 Eleven Realtime，称其为开发者构建自然语音智能体提供了更简单的方式。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这是面向开发者的语音智能体构建接口，普通用户无法直接使用，属于技术类选题。
Eleven Realtime is a new class of Speech to Speech models for voice agents that sound natural. Available in ElevenAgents and the Realtime API.
4,830 candidate prequalifications completed in three weeks, with data processed in Europe
ElevenLabs Blog · 10-20 12:00 UTC
该标题称，三周内完成了 4,830 份候选人预审，并且数据在欧洲处理。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这是招聘预审场景的成效型案例或宣传，未体现普通人可用的 AI 新功能，且细节有限。
Une voix naturelle pour les candidats, conforme pour les clients grands comptes.
Trinity Logistics gets 25% higher margins on shipments booked with ElevenAgents
ElevenLabs Blog · 10-20 12:00 UTC
ElevenLabs 博客案例称，Trinity Logistics 通过 ElevenAgents 预订货运后利润率提高了 25%。
AI 判断：标题初筛（未依据正文）：仅据标题判断，这是供应商企业案例，普通人相关性低且缺少可演示细节。
How 3PL Trinity Logistics used ElevenAgents to automate 30,000 inbound carrier calls/month, lift margins 25% on automated bookings, and cut costs 40%.
IBM Research：AI · 10-20 12:00 UTC
据标题信息，IBM Research 宣布 llm-d 1.0 发布，这是一套 Kubernetes 原生的分布式推理服务方案，宣称已面向生产环境就绪。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这属于面向开发者的推理部署与服务框架更新，与普通人的办公、学习或创作场景没有直接关联，实用性和画面感都很有限。
IBM、Red Hat 和 Google 等主导的开源推理项目 llm-d 在 PyTorch Conference North America 上发布 llm-d 1.0，这是一个 Kubernetes 原生的分布式推理服务框架，支持任意模型、任意加速器和任意云。
Google Cloud outage affects almost all banking and other operations in Europe
paweladamczuk · 10-20 10:26 UTC
标题显示，谷歌云的一次宕机影响了欧洲几乎所有银行业务及其他运营。
AI 判断：标题初筛（未依据正文）：标题仅表明云宕机影响欧洲银行，未涉及 AI 或普通人可用的新功能，且无法据此核实细节。
--- Top Comments ---
[nickcw]: From Google's status page https://status.cloud.google.com/incidents/Xs2XTHzAVC9YnafnPy... We are experiencing an issue with multiple services in
europe-west2 beginning on Monday, 2026-10-19 23:08 PDT.
Due to a power event we had suffered an outage within
europe-west2-b zone. The power event is mitigated. Partial
mitigation to infrastructure services will be observed within the
next 30 minutes and the remaining services will be mitigate...
[gadders]: >>Google cloud outage takes down HSBC, Revolut, Bank of Scotland, Llyods, Equifax, more You will probably have issues logging into Llyods even when the cloud is back up.
[otar]: The EU should really prioritise building AWS and Google Cloud alternatives with European values. > ssh root@eu-cloud We value your privacy. We and our 847 partners track your network packets.
[Accept all] [Manage preferences]
[paweladamczuk]: For me it was Revolut. No access to my main account, no withdrawals, no card payments since the morning. App doesn't work, there's no phone number to call, nobody to talk to. Dark side of neobanking, I guess.
[bcvvvy]: I bank with HSBC. Everything is up and working. A lot of the information people are leveraging here comes from Downdetector which is mostly users reporting that stuff is down, not whether they are actually down. It's a really bad source.
Franchisee reconciles cash in 94% less time with ChatGPT Work
OpenAI News · 10-20 10:00 UTC
据标题，OpenAI 发布案例称一名加盟经营者使用 ChatGPT Work 核对现金的时间减少了 94%。
AI 判断：标题初筛（未依据正文）：标题式判断：这是厂商发布的客户案例，只有一个时间节省比例，缺少可复现步骤与第三方验证，实用但说服力有限。
An independent Schlotzsky’s and Cinnabon franchisee cut menu repricing time by 95% and built a workflow that could free up $100K in cash flow.
AIBase 中文 · 10-20 09:55 UTC
据标题信息，DeepSeek 开源了其内部推理框架 DeepInfer，目标是降低大模型服务的运行成本。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向开发者的推理框架与部署成本优化，普通用户无法直接使用，属于技术类选题。
AI资讯 AI新闻资讯 正文 DeepSeek 开源内部推理框架 DeepInfer，要把大模型服务的成本打下来 发布于AI新闻资讯 发布时间 : 2026年10月20号 17:55 阅读 : 1 分钟 DeepSeek 在10月20日官宣，在已经陆续开源一系列大模型、TileLang 编译工具以及推理算子库之后，决定再往前迈一步——把自家的内部推理框架 DeepInfer 全面开源。 这家公司点明了一个长期被公众忽视的中间层:大模型的能力上限由模型本身划定，但如何以合理成本撑起超大规模的模型服务，真正取决于推理框架。训练让模型"学有所成"，推理则让模型"为人所用"。面对如今海量的用户调用，推理框架负责统筹调度庞大的算力集群，它的效率直接决定了 AI 服务的算力成本与电力消耗。 DeepSeek 想通过公开内部的部署经验，帮开发者与企业组织优化大规模推理的吞吐量与延迟，从而更高效地搭建私有化或公有云部署，压低算力成本、减少能源消耗，让智能以更低的价格触达更广泛的人群，实现它口中的"大模型普惠"。 基于 DeepInfer，DeepSeek 官方已经跑通了 DeepSeek V4.1Flash 的高吞吐、低延迟大规模推理服务;往后，新模型的推理代码和后续优化也会在这个框架里持续更新。 具体来看，DeepInfer 推理框架集成了高性能算子与深度优化的推理引擎，支持分布式推理、推测解码、SWA Bounded Replay、结构化输出等能力，并且同时兼容英伟达与华为昇腾两套平台。与之同步开源的，还有 deepseek-recipe 编解码组件——一个生产级的 Token 编解码模块，与 DeepSeek 官方 API 的处理逻辑完整对齐，确保社区拿到的 API 体验和官方部署保持一致。对想自己搭建推理服务的人来说，这意味着不必再从零摸索，直接站在 DeepSeek 的部署经验之上。 相关推荐 微软让设计师也上手改代码：AI 渗入 Windows 11 开发，老旧界面有望被填平 微软正将AI引入Windows 11开发流水线，并改变设计师角色。Windows产品设计副总裁巴卡在X讨论中称，AI工具显著提升设计团队效率，部分设计师已借助AI直接参与代码修改，以优化系统精细度、一致性和体验。他称“这只是一切的起点”。该讨论源于优素福·迈赫迪展示新版Windows搜索。 2026年10月20号 17:57 145.2k DeepSeek 再掀开源浪潮：内部推理框架 DeepInfer 全面公开，剑指"大模型普惠" DeepSeek官宣将内部推理框架DeepInfer全盘开源。此前其已陆续发布大模型、TileLang编译工具和推理算子库。DeepSeek认为，模型能力上限由模型本身决定，但如何以合理成本向海量用户供给超大规模模型，取决于推理框架这一长期被忽视的中间架构。训练让模型学有所成，推理才让它为人所用。 2026年10月20号 17:46 164.9k DeepSeek 开源推理框架 DeepInfer:推理成本成为 AI 普惠的关键一层 10月20日，DeepSeek宣布全面开源内部推理框架DeepInfer，继系列大模型、TileLang编译工具和推理算子库后再度开放。官方称，模型能力边界由模型划定，但以合理成本服务超大规模模型，关键在推理框架这一中间层；它调度庞大算力集群，效率直接决定算力成本与电力消耗。DeepInfer集成高性能算子与深度优化推理引擎。 2026年10月20号 17:11 190.1k AI日报：腾讯开放元宝搜索；MiniMax 更新 M3.1Flash；蚂蚁百灵开源 Ming-3.0 语音系列 AI日报聚焦人工智能热点、技术趋势与创新产品应用。本期重点：腾讯正式开放元宝搜索，以千亿级索引支持AI智能体调用，并通过Skill、MCP、API三种方式服务开发者；该搜索接入157个行业权威信源，为开发者提供更丰富的搜索能力。 2026年10月20号 17:09 178.9k 日本演员协会发布首版生成式AI准则，复刻演员声音须全程获授权 日本演员协会上周发布首版生成式AI使用准则，规范未经授权以AI复刻演员声音。准则适用于语音合成、深度学习、机器学习等技术生成的仿声音频，要求声音复刻各环节均须取得演员本人同意；AI企业将演员声音用于模型训练，须获表演者许可。经授权生成的AI语音被视为“表演”，企业应向演员支付报酬。 2026年10月20号 16:53 153.6k 智启未来，您的人工智能解决方案智库 简体中文 © 2026 AIBase 备案号：闽ICP备08105208号-24
DeepSeek 开源推理框架 DeepInfer:推理成本成为 AI 普惠的关键一层
AIBase 中文 · 10-20 09:11 UTC
据标题信息，DeepSeek 开源了名为 DeepInfer 的推理框架，并称推理成本是 AI 普惠的关键一层。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向开发者的推理框架开源消息，缺少普通人可直接使用或演示的具体功能，且无价格、地区与使用限制信息，属技术类选题。
AI资讯 AI新闻资讯 正文 DeepSeek 开源推理框架 DeepInfer:推理成本成为 AI 普惠的关键一层 发布于AI新闻资讯 发布时间 : 2026年10月20号 17:11 阅读 : 1 分钟 DeepSeek10月20日宣布全面开源内部推理框架 DeepInfer，这是继此前开源系列大模型、TileLang 编译工具与推理算子库之后的又一次开放。官方指出，大模型的能力边界由模型本身划定，但如何以合理成本提供超大规模模型服务，还取决于推理框架这一中间层——它统筹调度庞大的算力集群，其效率直接决定 AI 服务的算力成本与电力消耗。 DeepInfer 集成高性能算子与深度优化的推理引擎，支持分布式推理、推测解码、SWA Bounded Replay、结构化输出等能力，同时兼容英伟达与华为昇腾平台。基于该框架，DeepSeek 已实现 DeepSeek V4.1Flash 的高吞吐、低延迟大规模推理服务，未来新模型的推理代码及后续优化也将持续在此更新。 配套同步开源的 deepseek-recipe 是生产级 Token 编解码组件，与 DeepSeek 官方 API 的处理逻辑完整对齐，确保社区获得与官方部署一致的 API 体验，两部分代码均已在 GitHub 开放，DeepSeek 另在知乎发布了完整技术报告。 官方表示，开源推理框架的目的一在于帮助开发者与企业组织优化大规模推理的吞吐与延迟，高效构建私有化与公有云部署，从而降低算力成本、减少能源消耗，让智能以更低价格触达更广泛人群，并将继续支持社区进行大规模部署优化。模型、算子库与推理框架的逐层开放，延续了国产大模型以开源生态降低推理门槛、推动智能普惠的行业路径。 开源仓库 DeepInfer: https://github.com/deepseek-ai/DeepInfer deepseek-recipe: https://github.com/ deepseek-ai/deepseek-recipe 相关推荐 微软让设计师也上手改代码：AI 渗入 Windows 11 开发，老旧界面有望被填平 微软正将AI引入Windows 11开发流水线，并改变设计师角色。Windows产品设计副总裁巴卡在X讨论中称，AI工具显著提升设计团队效率，部分设计师已借助AI直接参与代码修改，以优化系统精细度、一致性和体验。他称“这只是一切的起点”。该讨论源于优素福·迈赫迪展示新版Windows搜索。 2026年10月20号 17:57 145.2k DeepSeek 开源内部推理框架 DeepInfer，要把大模型服务的成本打下来 10月20日，DeepSeek宣布在开源多款大模型、TileLang编译工具及推理算子库后，进一步全面开源内部推理框架DeepInfer。文章指出，模型能力上限由模型本身决定，但能否以合理成本支撑超大规模模型服务，关键在推理框架：训练让模型“学有所成”，推理让模型“为人所用”，以应对海量应用需求。 2026年10月20号 17:55 195.7k DeepSeek 再掀开源浪潮：内部推理框架 DeepInfer 全面公开，剑指"大模型普惠" DeepSeek官宣将内部推理框架DeepInfer全盘开源。此前其已陆续发布大模型、TileLang编译工具和推理算子库。DeepSeek认为，模型能力上限由模型本身决定，但如何以合理成本向海量用户供给超大规模模型，取决于推理框架这一长期被忽视的中间架构。训练让模型学有所成，推理才让它为人所用。 2026年10月20号 17:46 164.9k AI日报：腾讯开放元宝搜索；MiniMax 更新 M3.1Flash；蚂蚁百灵开源 Ming-3.0 语音系列 AI日报聚焦人工智能热点、技术趋势与创新产品应用。本期重点：腾讯正式开放元宝搜索，以千亿级索引支持AI智能体调用，并通过Skill、MCP、API三种方式服务开发者；该搜索接入157个行业权威信源，为开发者提供更丰富的搜索能力。 2026年10月20号 17:09 178.9k 日本演员协会发布首版生成式AI准则，复刻演员声音须全程获授权 日本演员协会上周发布首版生成式AI使用准则，规范未经授权以AI复刻演员声音。准则适用于语音合成、深度学习、机器学习等技术生成的仿声音频，要求声音复刻各环节均须取得演员本人同意；AI企业将演员声音用于模型训练，须获表演者许可。经授权生成的AI语音被视为“表演”，企业应向演员支付报酬。 2026年10月20号 16:53 153.6k 智启未来，您的人工智能解决方案智库 简体中文 © 2026 AIBase 备案号：闽ICP备08105208号-24
DeepInfer: High-performance inference framework for DeepSeek models
ggcr · 10-20 09:08 UTC
DeepSeek 发布了一个名为 DeepInfer 的高性能推理框架，用于运行 DeepSeek 系列模型。
AI 判断：标题初筛（未依据正文）：这是面向开发者的推理框架，普通用户没有可直接使用的场景，属于技术类内容，基于标题判断。
--- Top Comments ---
[schneehertz]: A nuclear bomb "Based on this framework, we have achieved high-throughput, low-latency large-scale inference services for DeepSeek V4.1 Flash. Currently, the service throughput of a single cluster is 155 trillion tokens/day, and the overall revenue is $38 per card per hour (at off-peak pricing)" Source: https://zhuanlan.zhihu.com/p/2095560295930508953
[mmaunder]: Interesting that the DHS "Exclusion and Removal Order #2026-01" that was made public yesterday was actually released July 24th, 2026 and appears to have taken effect then. Agencies would have started pulling deepseek at that time. Which means it's likely DS became aware of the memo around then. The release of this repo hours after the DHS order is a tactical move they started preparing when the DHS order was issued almost 3 months ago, and they timed publishing the repo right a...
[wg0]: That's Kami Kaze for US frontier labs. There's no way from here for them towards trillion dollar valuations. That's strategic sabotage to say the least. Ban all open weight models! Cried the VCs.
[Bengalilol]: Whatever the consequences, I have to admit China has understood that all those are raw materials and that open sourcing everything will help growing their economy (as a whole). The worst idea was to build corporations providing services with AI. It was an extension of the latest 50 years model, which is about to crumble. Now, we have our revolution. (Ok, ok, I am a little bit too enthusiastic, I now need a fallout shelter).
[yaoke259]: Supporting Huawei Ascend 950 NPUs with open source inference code has huge implications in the global AI hardware market. Imagine if all smaller labs now can build on top of this stack and add increasing support for more compatible hardware , software and models. This could seriously challenge NVidias death grip. Although inference is still only a part of all AI workloads and open sourcing training on Huawei chips would be even bigger.
Anthropic co-founder says the singularity is underway
The Rundown AI · 10-20 09:00 UTC
Anthropic 联合创始人公开表态，认为技术奇点已经开始发生。
AI 判断：标题初筛（未依据正文）：仅凭标题可知这是一位公司高管的观点性表态，没有可核实细节，也缺少普通人可直接使用的信息，属标题层面的判断。
PLUS: Try ChatGPT’s new voices with GPT-Live-1.1
公众号：DeepSeek（深度求索） · 10-20 08:54 UTC
DeepSeek 宣布将内部推理框架 DeepInfer 全面开源。
AI 判断：标题初筛（未依据正文）：仅从标题看，这是面向开发者的推理框架开源，普通人难以直接使用，属于技术类选题。
DeepSeek 宣布将内部推理框架 DeepInfer 全面开源，并同步开源生产级 Token 编解码组件 deepseek-recipe。
AIBase 中文 · 10-20 08:53 UTC
日本演员协会发布了首版生成式AI准则，其中要求复刻演员声音必须全程获得授权。
AI 判断：标题初筛（未依据正文）：基于标题判断，这是行业准则新闻，缺少可供普通人直接演示的用法，但提供了AI声音授权与权益讨论角度。
AI资讯 AI新闻资讯 正文 日本演员协会发布首版生成式AI准则，复刻演员声音须全程获授权 发布于AI新闻资讯 发布时间 : 2026年10月20号 16:53 阅读 : 1 分钟 日本演员协会（JAU）上周发布首版生成式 AI 使用准则，针对未经授权利用 AI 复刻演员声音的行为提出规范要求。准则适用于通过语音合成、深度学习、机器学习等技术生成的演员仿声音频，要求相关主体在声音复刻流程的每个环节均须取得演员本人同意。AI 企业若将演员声音用于模型训练，必须获得表演者许可;经授权生成的 AI 语音则被视为一种“表演”，企业应向演员支付报酬或版税。 日本配音演员长期反对未经许可的声音盗用。2024年，多名配音演员成立组织，抗议 AI 声音滥用，并指出现有版权法律保护不足。曾为《进击的巨人》《电锯人》《游戏王》配音的津田健次郎，曾因社交媒体大量出现未经授权使用其声音的视频起诉 TikTok，后因相关视频下架，案件被驳回。 在准则发布会上，因出演《龙珠》孙悟空而知名的野泽雅子表示，未经许可使用其声音与盗窃无异，声音和表演都应受到尊重，即使是少量使用也必须事先征得同意。她还指出，AI 复刻声音缺乏情感表现。 日本演员协会拥有约2600名配音演员及演员会员，这份准则是日本首批针对 AI 语音录制的行业规范之一。尽管日本社会对 AI 的整体接受度较高，但配音演员普遍担忧声音被滥用将侵害劳动成果、挤压就业机会。随着日本政府持续探讨生成式 AI 监管框架，该准则也为 AI 声音授权、训练许可及收益分配提供了行业规范参考。 相关推荐 微软让设计师也上手改代码：AI 渗入 Windows 11 开发，老旧界面有望被填平 微软正将AI引入Windows 11开发流水线，并改变设计师角色。Windows产品设计副总裁巴卡在X讨论中称，AI工具显著提升设计团队效率，部分设计师已借助AI直接参与代码修改，以优化系统精细度、一致性和体验。他称“这只是一切的起点”。该讨论源于优素福·迈赫迪展示新版Windows搜索。 2026年10月20号 17:57 145.2k DeepSeek 开源内部推理框架 DeepInfer，要把大模型服务的成本打下来 10月20日，DeepSeek宣布在开源多款大模型、TileLang编译工具及推理算子库后，进一步全面开源内部推理框架DeepInfer。文章指出，模型能力上限由模型本身决定，但能否以合理成本支撑超大规模模型服务，关键在推理框架：训练让模型“学有所成”，推理让模型“为人所用”，以应对海量应用需求。 2026年10月20号 17:55 195.7k DeepSeek 再掀开源浪潮：内部推理框架 DeepInfer 全面公开，剑指"大模型普惠" DeepSeek官宣将内部推理框架DeepInfer全盘开源。此前其已陆续发布大模型、TileLang编译工具和推理算子库。DeepSeek认为，模型能力上限由模型本身决定，但如何以合理成本向海量用户供给超大规模模型，取决于推理框架这一长期被忽视的中间架构。训练让模型学有所成，推理才让它为人所用。 2026年10月20号 17:46 164.9k DeepSeek 开源推理框架 DeepInfer:推理成本成为 AI 普惠的关键一层 10月20日，DeepSeek宣布全面开源内部推理框架DeepInfer，继系列大模型、TileLang编译工具和推理算子库后再度开放。官方称，模型能力边界由模型划定，但以合理成本服务超大规模模型，关键在推理框架这一中间层；它调度庞大算力集群，效率直接决定算力成本与电力消耗。DeepInfer集成高性能算子与深度优化推理引擎。 2026年10月20号 17:11 190.1k AI日报：腾讯开放元宝搜索；MiniMax 更新 M3.1Flash；蚂蚁百灵开源 Ming-3.0 语音系列 AI日报聚焦人工智能热点、技术趋势与创新产品应用。本期重点：腾讯正式开放元宝搜索，以千亿级索引支持AI智能体调用，并通过Skill、MCP、API三种方式服务开发者；该搜索接入157个行业权威信源，为开发者提供更丰富的搜索能力。 2026年10月20号 17:09 178.9k 智启未来，您的人工智能解决方案智库 简体中文 © 2026 AIBase 备案号：闽ICP备08105208号-24
X：Rohan Paul (@rohanpaul_ai) · 10-20 08:06 UTC
据路透报道，OpenAI 计划在 Anthropic 上市前加速完成约 300 亿美元融资，估值超过 1.4 万亿美元。
AI 判断：标题初筛（未依据正文）：仅凭标题可见这是纯融资传闻类消息，缺少普通人可用的产品或实操信息，且尚未证实。
据 Reuters 报道，OpenAI 正赶在 Anthropic 上市前洽谈融资约 300 亿美元，估值超 1.4 万亿美元，比 3 月的 8520 亿美元至少高 64%。
Ai2 发布 FairStress 基准：测量 LLM 决策中被隐藏的公平性不对称
Ai2 / Allen Institute for AI（RSS） · 10-20 08:00 UTC
Ai2 发布名为 FairStress 的基准，用于测量大语言模型在做出决策时被隐藏的公平性不对称。
AI 判断：标题初筛（未依据正文）：仅凭标题可判断这是面向模型评测的技术基准发布，属于开发者向内容，普通人既难以直接使用，也没有可演示的画面。
Ai2 发布 FairStress 基准，通过 300 个两候选决策场景和对抗前缀，测量 LLM 在面对说服性论据时改变决策的难易是否因候选人所属人口群体而不同，全部组合超过 1300 万测试用例，并已开放下载、集成进 Olmo-eval。
Waymo：Blog（网页） · 10-20 07:00 UTC
Waymo 宣布将进入西班牙，计划于 2027 年在马德里大区接待首批乘客。
AI 判断：标题初筛（未依据正文）：依据标题，这是自动驾驶出租车服务的地域扩张消息，对普通乘客有出行体验意义，但可用时间和地区限制明显。
Waymo 宣布为在西班牙马德里大区推出全自动驾驶出行服务奠定基础，目标 2027 年迎来首批乘客。未来数日其电动 Jaguar I-PACE 车辆将现身当地街头，先以人工驾驶开始，再由当地受训的自动驾驶专家进行技术验证。Waymo 称已累计完成逾 2000 万次行程、400 万公里全自动驾驶里程，在其运营城市中涉及重伤或更严重事故的比例比人类司机低 95%。
MicroG: Free-libre re-implementation of Google's proprietary Android user space
theanonymousone · 10-20 06:37 UTC
MicroG 是一个以自由开源方式重新实现谷歌专有安卓用户空间组件的项目。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，该项目属于开发者向的开源兼容层工程，与普通人可用的 AI 功能无关，且并非新事件。
--- Top Comments ---
[maelito]: I have never understood why people using MicroG can use apps like SNCF-Connect, while I cannot. SNCF-connect is the largest online marchant in France. Since june, I cannot click the "pay" button anymore. It coincides with the new feature of Pay with Google. So : is it because MicroG allows calls to Google services on e/OS devices, while my LineageOS refuses ? Or because MicroG implemented a try catch around this call, without any call to Google's servers ?
[orphea]: I guess this is submitted because of https://news.ycombinator.com/item?id=50154624 ?
[jamesnorden]: Why do the most recent commits all say "Merge commit from fork" without any description of what the changes do or what the fork is?
[mshish]: It's too bad you have to rely on signature spoofing to be able to use this as an alternative (but obviously I understand why). The spoofing looks narrowly scoped on LineageOS, but in general I imagine it creates some risk.
[xx__yy]: I use it on LineageOS, works a charm
Thibault Sottiaux: Day 15/ Nevermind, we're still cooking... and reset. Enjoy. https://t.co/RZ9WhtI802
Thibault Sottiaux · 10-20 06:33 UTC
Thibault Sottiaux 仅发布了一条含糊的“第15天”动态，称仍在继续准备并重置，只附链接，未说明具体产品或功能。
AI 判断：标题初筛（未依据正文）：标题级判断：只有含糊预告和链接，缺乏可核实的 AI 功能、用途或影响信息，属于不透明的预告型内容。
Day 15/ Nevermind, we're still cooking... and reset. Enjoy. https://t.co/RZ9WhtI802
公众号：面壁智能（MiniCPM） · 10-20 06:22 UTC
面壁智能解析前瞻专家调度 LES 如何让 35B MoE 在手机端侧实时无损运行。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向开发者的推理与调度优化解析，缺少普通人可直接上手的演示或产品信息。
面壁智能 MiniCPM-V 团队介绍支撑 MiniCPM-V 4.7 端侧运行的前瞻专家调度（LES）技术，称首次实现 35B-A3B 规模 MoE 在端侧实时无损运行，iPhone 18 Pro Max 实测解码超 20 token/s。
Nikunj Kothari: Tenzing on Apple TV+ is a must watch.. The acting, the direction, the music. I can’t believe...
Nikunj Kothari · 10-20 05:57 UTC
一条推文盛赞 Apple TV+ 剧集《Tenzing》的演技、导演与音乐，内容与 AI 无关。
AI 判断：标题初筛（未依据正文）：仅凭标题可判断这是影视娱乐推荐，不涉及任何 AI 功能、工具或新闻，不适合面向普通人的 AI 选题。
Tenzing on Apple TV+ is a must watch..
The acting, the direction, the music.
I can’t believe this was all shot in real environments and with zero green screens.
Literal goosebumps.
Aaron Levie: The big bitter lesson question in the applied AI layer is that as models get better and bette...
Aaron Levie · 10-20 04:44 UTC
Aaron Levie 抛出应用 AI 层的“苦涩教训”之问：当模型不断变强，应用层该如何自处（标题不完整，仅据标题转述）。
AI 判断：标题初筛（未依据正文）：这是面向 AI 应用层从业者的行业观点，标题本身不完整，也缺少普通人可感知的功能或案例，仅凭标题难以对大众提供实用价值。
The big bitter lesson question in the applied AI layer is that as models get better and better, what is the best way to diffuse that intelligence into an enterprise.
Ironically, I’d argue that as models are getting more capable, and can handle longer running tasks and better tool use, the ability to differentiate on a domain by domain basis goes up instead of down.
When models were quite limited in what they could do, ironically all you could be literally was a “wrapper” because you were effectively delivering tokens back to the user with a custom prompt or some limited RAG.
But in a world of agents using tools, tasks that run on for hours or days, workflows that kick off agents, agent swarms that need to be coordinated against a goal, data sources that are domain specific, model routing that requires domain-specific routing, implementation that requires real change management of the workflow instead of “just type into this chat box”, it’s very very possible that specialization of these systems goes up more.
Obviously this doesn’t apply to everything - the closer you are to the labs core use-cases the tougher this play gets, but I’d argue we’re going to see autonomous systems for a wide range of job functions that weren’t practical to build out before. We’ll see this in security, life sciences, healthcare, finance, manufacturing, and many other fields. Either way going to be a fun and wild couple of years.
Zara Zhang: See this in a theater! One of the most interesting things about this documentary was how much...
Zara Zhang · 10-20 04:22 UTC
Zara Zhang 发帖推荐去影院观看一部纪录片，并提到该片最有趣之处之一，但标题未说明其与 AI 的具体关系。
AI 判断：标题初筛（未依据正文）：标题仅涉及纪录片观影推荐且未明确 AI 关联，无可演示功能或普通人可用的实用信息，AI 选题价值低。
See this in a theater! One of the most interesting things about this documentary was how much & when people laughed while watching it https://t.co/Q9kAdgyAoW
Peter Yang: This doesn't mean it's a good movie btw. I found it completely ridiculous
Peter Yang · 10-20 04:18 UTC
这条推文只是作者对某部电影的主观评价，未提供可核实的产品信息或具体背景。
AI 判断：标题初筛（未依据正文）：仅有主观评价片段，缺少可演示的功能或事实依据，无法支撑面向普通人的教程或趋势解读。
This doesn't mean it's a good movie btw. I found it completely ridiculous
爱范儿 · 10-20 03:57 UTC
标题称，押注记忆原生的 MemT 1.0 试图在养成系 AI 的下半场竞争中实现突破。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这更像记忆原生技术或 Agent 架构话题，普通人的直接可用价值不明确。
MemT 1.0 开源，给出「记忆原生」新解法
#欢迎关注爱范儿官方微信公众号：爱范儿（微信号：ifanr），更多精彩内容第一时间为您奉上。
Nikunj Kothari: Waiting for someone to sell me a peptide factory built in a data center which is also a robot...
Nikunj Kothari · 10-20 02:20 UTC
这是一条只有标题的推文式感叹，作者在表达等待出现“建在数据中心里、同时也是机器人的肽工厂”这类产品的期待。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是一句缺乏事实依据的模糊猜想式发言，没有可演示的功能、可核实的信息或普通人可用之处。
Waiting for someone to sell me a peptide factory built in a data center which is also a robotics neolab that is somehow collecting first party uniquely differentiated data..
And while we’re at it use the words tape out, memory/CPU shortage too.
Make it 50/500 on a memo 🤝
OpenAI News · 10-20 00:00 UTC · 同链接合并评分
OpenAI 宣布收购 Orchestra Bio，旨在加快 AI 驱动的药物研发。
AI 判断：标题初筛（未依据正文）：仅凭标题看，这是一则企业收购新闻，缺少面向普通人的具体用法或可演示效果，实际相关性有限。
OpenAI is acquiring Orchestra Bio to bring AI-powered operations to life sciences R&D, helping biopharma teams accelerate drug development and advance new treatments.
Databricks 发布 MemCurrent 生产级 Agent 持久记忆架构
Databricks：Blog（RSS） · 10-20 00:00 UTC
Databricks 官方博客发布 MemCurrent，一套用于生产环境 Agent 的持久记忆架构。
AI 判断：标题初筛（未依据正文）：仅凭标题可判断这是面向开发者的 Agent 记忆架构工程内容，普通人难以直接使用，属于技术类选题。
Databricks 介绍 MemCurrent，一种面向生产 Agent 的持久记忆架构，结合情景记忆与派生记忆，通过插件、工具调用和离线 dreaming 管线三种访问模式读写与整合记忆，已用于 Agent Bricks Memory 和 Genie。
My obituary for programming
Simon Willison · 10-19 23:58 UTC
西蒙·威利森发布题为《我给编程写的讣告》的文章，仅从标题看，这是一篇以“讣告”口吻谈论编程命运的个人观点文章。
AI 判断：标题初筛（未依据正文）：仅凭标题判断，这是面向开发者群体的观点随笔，缺少普通人可直接上手的功能或可演示的效果对比，且标题带有比喻性，难以核实其具体论点。
My obituary for programming
Jeremy Scheff poignantly captures the key loss caused by coding agents for career programmers who enjoyed the
puzzle
of developing code:
Programming turned into one of the greatest jobs you could imagine. You got paid to work on tricky little logic puzzles. The trickier the puzzle, the more you got paid. But even pretty simple puzzles were worth a lot of money because regular people just couldn't do it! No, they needed us. They would give us a problem and we would solve it the way we wanted. We set the strategy, we made the timeline, we defined what was and was not possible.
If you're a programmer who thrives on building working systems and doesn't care too deeply about the problems solved along the way, LLMs are the gift that keeps on giving.
If directly solving those problems was a big part of your enjoyment, the loss is very real.
Via
lobste.rs
Tags:
programming
,
careers
,
ai
,
generative-ai
,
llms
,
coding-agents
,
deep-blue
中文标题Thibault Sottiaux：第 15 天／算了，我们还在继续烹饪……并重置。享受吧。https://t.co/RZ9WhtI802
Follow Builders · X 动态 · Thibault Sottiaux · 10月20日 06:33 UTC · 喜欢 1460 · 转发 108 · 回复 627
Thibault Sottiaux 仅发布了一条含糊的“第 15 天”动态，称仍在继续准备并重置，只附链接，未说明具体产品或功能。
中文标题尼昆·科塔里：Apple TV+ 上的《Tenzing》必看……演技、导演、音乐，我简直不敢相信
Follow Builders · X 动态 · Nikunj Kothari · 10月20日 05:57 UTC · 喜欢 12 · 转发 0 · 回复 1
一条推文盛赞 Apple TV+ 剧集《Tenzing》的演技、导演与音乐，内容与 AI 无关。
中文标题Aaron Levie：应用 AI 层最关键的那个“苦涩教训”问题是——随着模型越来越强……
Follow Builders · X 动态 · Aaron Levie · 10月20日 04:44 UTC · 喜欢 83 · 转发 7 · 回复 46
Aaron Levie 抛出应用 AI 层的“苦涩教训”之问：当模型不断变强，应用层该如何自处（标题不完整，仅据标题转述）。
中文标题Zara Zhang：去影院看这个！这部纪录片最有趣的一点之一是，它在多大程度上……
Follow Builders · X 动态 · Zara Zhang · 10月20日 04:22 UTC · 喜欢 16 · 转发 0 · 回复 2
Zara Zhang 发帖推荐去影院观看一部纪录片，并提到该片最有趣之处之一，但标题未说明其与 AI 的具体关系。
中文标题Peter Yang：顺便说一句，这并不意味着它是一部好电影。我觉得它完全荒谬
Follow Builders · X 动态 · Peter Yang · 10月20日 04:18 UTC · 喜欢 3 · 转发 0 · 回复 1
这条推文只是作者对某部电影的主观评价，未提供可核实的产品信息或具体背景。
中文标题Nikunj Kothari：等着有人卖给我一座建在数据中心里、同时又是机器人的肽工厂……
Follow Builders · X 动态 · Nikunj Kothari · 10月20日 02:20 UTC · 喜欢 36 · 转发 1 · 回复 7
这是一条只有标题的推文式感叹，作者在表达等待出现“建在数据中心里、同时也是机器人的肽工厂”这类产品的期待。
