ed interactive group

We're ed interactive group. We're about growing B2B different, better, smarter. Different in experience. Better at creating value. Smarter at sharing insight.

[AYi @AYi_AInotes]

网景创始人 、硅谷顶级投资人Marc Andreessen 在访谈里坦言，他看到了科技史上最可怕的一道认知鸿沟，

传统巨头以为 AI 只是个降本增效的插件，而新一代 AI 原生团队，正在用完全非人的组织效率掀翻整个商业版图。

当年写出网景浏览器、见证了整个互联网三十年兴衰的 Marc Andreessen，坐在 Ben Horowitz 面前直接发出了感慨，
如果 1994 年我们做 Netscape 的时候就有今天这种 AI 工具，那简直不可想象。

现在绝大多数传统巨头和现有企业，依然在用改良软件的旧思维去看待大模型。

在他们的组织逻辑里，AI 只是一个高级功能，用来给现有业务做个插件、在钉钉里加个智能助手、或者把客服团队砍掉 20% 省点成本。

他们以为自己在拥抱变革，其实只是在给老旧的蒸汽机打补丁。

但真正恐怖的事情，正在那些刚成立的 AI 原生团队里发生。

那些创始人对 AI 能力的理解，跟大厂完全不在同一个平行宇宙。

他们每周都在看着底层技术飞速进化，而且他们压根不会把自己定义成一家软件公司、硬件公司、甚至是一家 AI 公司。

以前创办一家公司，你的组织架构是按人类的协同瓶颈搭出来的：
招产品经理写文档、招工程师写代码、招法务改合同、招运营做投放。
因为人类的脑力带宽有限，所以必须靠庞大的部门和管理层级来支撑业务。

而真正的 AI 原生物种，是用机器智能作为最底层的骨架。

两个人加上一整套自动化 Agent，可以直接把从市场洞察、软件工程、全栈设计到商业销售的链路全部跑通。

他们不是在效率上快了 30%，他们是在组织形态上彻底跨越了一个物种。

大厂的盲区，从来不是买不起算力，而是他们无法理解组织本身才是最大的负债。

当现有企业还在开会讨论如何把 AI 塞进现有的业务流程时，新一代的野蛮人正在用完全超越行业界限的方式，把传统的业务流程连根拔起。

未来拉开人与人、公司与公司之间差距的，根本不是你用了哪个模型，而是你到底是在用新工具维持旧世界的体面，还是用新世界的底层逻辑，重新定义了自己能做的事情。

[Financelot @FinanceLancelot]

AI Token spending has completely collapsed the last two months.

There has been zero growth in the AI sector, that's why pre-IPO shareholders in Anthropic and OpenAI are panicking to go public.

Brad Gerstner is a pre-IPO investor, that's why he's been on CNBC telling everyone to buy AI.

The bubble popped in July 2026.

[Adam Livingston @AdamBLiv]

Want to have your mind blown?

In 1971, the median household income was $9,030, or 220 ounces of gold.

In 2025, it was $87,460, or just 25 ounces of gold.

Your paycheck went up 10x.

Measured in gold, it fell 88%.

To earn what the median household earned in 1971, you’d need about $890,000 per year today.

That’s what fiat debasement looks like.

[micro1 @micro1_ai]

Answer 5 questions and find out how much your company data is worth.

Companies are earning $100K–$2M+ by licensing anonymized operational data to micro1 for AI training.

See how much your company’s data could be worth

[Alex Prompter @alex_prompter]

This Claude prompt wipes your name, phone number and address off the internet.

"Adopt the role of a privacy investigator who's helped hundreds of people disappear from data brokers, people-search sites and old accounts.
I want my personal information off the internet. Work through it with me one step at a time, then give me:

→ every place my name, phone, email and address show up, with links
→ the opt-out page for each one, worst first
→ the old accounts I should close, and how to delete each one for good
→ a GDPR or CCPA deletion letter for every site with no opt-out button
Don't submit, send or delete anything until I approve it.

My name, city, email and old usernames: [YOUR DETAILS]"

[Financelot @FinanceLancelot]

Uh oh... Jensen Huang just jumped ship, abandoning OpenAI and Anthropic; telling every nation they need their own sovereign AI.

Is it over for Sam Altman? $NVDA

[Blacklock's Reporter @mindingottawa]

Proposed regulations would permit cellphone jamming by the Ontario Provincial Police, Sureté du Québec, Royal Newfoundland Constabulary, and municipal police departments in Barrie, Calgary, Durham Region, Edmonton, Halton Region, Hamilton, Kingston, Lethbridge, Medicine Hat, Niagara Region, Ottawa, Peel Region, Regina, Saskatoon, Sault Ste. Marie, Sudbury, Toronto, Thunder Bay, Vancouver, Victoria, Windsor, Winnipeg and York Region.

No deadline for enforcement of the new rules was set.

[Mika @mika_systems]

this is f*cking insane, Claude Code agents turned a reading list into a lab you can run in your browser

Boris Cherny (creator of Claude Code) gave 12 agents 64 research papers and one assignment: build a browser lab where readers can rerun the experiments

pick a fluid simulation, change the conditions and watch how the result changes

switch methods while keeping the same scene. inspect the equations, code and measurements beside the animation

give each group a specific job:

→ readers extract equations, assumptions and experimental settings
→ builders turn the methods into runnable code
→ reviewers compare outputs against the papers
→ interface agents build the controls for exploring each experiment

a simulation can look convincing while producing the wrong answer

reviewers check whether it preserves mass, stays stable and matches the paper's measurements before accepting the implementation

every experiment has to show:

> the claim being tested
> the conditions used
> the difference from the reference result
> the details still needed to verify it

the lab covered 11 methods across 28 reference scenes. seven matched the selected reference results within the stated tolerances; four stayed marked for further review

give Claude Code your paper collection, then ask:

"Turn these papers into an interactive browser lab

1. Extract implementable methods, equations and experimental settings. Attach page references and flag missing details

2. Assign independent methods to separate agents. Use shared inputs and measurement tools so the results can be compared

3. Have reviewers check each implementation against the paper. Record discrepancies and keep unverified results clearly marked

4. Build controls for changing parameters, comparing methods and opening the code behind each result"

start with one paper and reproduce one reference experiment before adding more agents

include the inputs, dependencies and checks so the next reader can rerun it from a clean checkout

watch the agents at work below

[Kyrox @kyr0stack]

Anthropic engineer:

"90% of our engineers were already running self-improving loops

Now everyone is building agentic graphs"

"Prompting is basically over"

In just 10 minutes, she builds her whole Claude Code system live from an empty terminal

Prompts → Agents → Loops → Graphs

skip the loop and the same mistake ships again next week

skip the graph and you're still the one wiring every agent together by hand

prompting was the old workflow, graph engineering is the new one

watch it today

then save the full guide below before everyone catches up ↓

[前端哥Liam @jinglian]

REA（Reverse Engineer Anything）逆向一切

三天从 3 万 Star 涨到 12 万+，一天新增 2.5 万以上。MIT 开源。

大家都在说它，"一切软件都可逆"，

我看到更好玩的是：拆硬件的通信协议。

家里的空调、热水器、灯，基本都配了一个难用的厂商 App。App 背后，其实就是一套设备通信协议。

REA 这几天的更新正好补齐了这条链路：
· adb 连模拟器或真机，抓到官方 App
· 分析 HAR 或 mitmproxy 抓包，把每个请求定位到 App 源码
· 用 Binwalk/Unblob 拆固件，再交给 Ghidra 分析

然后把证据交给 Claude Code / Codex，还原协议，自己写控制端。

X 上已经有人用同样的思路，逆向了空调，做成 Mac 状态栏应用，又花 15 分钟移植到 AR 眼镜上，看一眼就能控制。

以前是厂商 App 决定你能怎么用自己的设备，以后你可以自己写。

[AYi @AYi_AInotes]

Scale 创始人，现在在 Meta 带超级智能实验室的 Alexandr Wang，说了句很颠覆的话，他说现在是人类历史上建十亿美元公司最好的时机，理由跟多数人想的相反，他说瓶颈早就不是模型不够强了，

他的判断是，今天 AI 真正的瓶颈，是世界接受和用上它的速度太慢，技术已经摆在那，社会还在后面追，他甚至给了个更狠的假设，就算模型从今天起完全停止进步，经济和世界的运转方式，照样会被改写几十年。

打个老比方，电刚发明出来的时候，发电机已经有了，可工厂，家庭，城市要花几十年才真正把电用起来，AI 现在就站在那个位置，谁帮世界把这道坎跨过去，谁就站在新公司出生的地方，他管这叫文明级别的机会。

最反直觉的是竞争规则变了，十年前他创办 Scale，小公司打巨头是大卫打歌利亚，只能靠聪明，找缝隙，现在小团队配上智能体，变成了机甲歌利亚，巨头反而是传统歌利亚，因为大公司的流程，分工，汇报线，全是按没有 AI 的世界设计的，转身慢，小团队从第一天就按 AI 的方式搭，人少，反而能正面硬刚。

说穿了，机甲歌利亚的内核是一种新组织，少数人，加一大群智能体，大公司真正的包袱从来不是技术，是组织惯性，这也是给每个做组织，带团队的人的信号，未来的竞争力，不在你有多少人，在你这群人能指挥多少智能体。

他后半段讲到 Meta 在做的个人超级智能，让几十亿人每个人都有一个懂自己，为自己定制的智能，本质是把一个人的能动性放大到以前不敢想，他最后抛了个问题，如果有一天一切都变得很容易，你会去做什么，这个问题，比任何创业建议都值得停下来想想。

当然也留个清醒，他在 Meta 负责超级智能，看好 AI 本就符合他的立场，渠道，品牌，数据，监管依然是巨头的护城河，窗口也不会一直开着，但对普通人最实在的就一句，别再等下一代模型了，现有的已经够用，你和机会之间，差的不是更强的工具，是把它用进一件具体的事里。

[Rohan Paul @rohanpaul_ai]

New Amazon paper shows that an LLM can act as a research world model, predicting whether a training change will help before anyone spends GPU time on it.

AI research agents can propose experiments far faster than teams can afford to run them. Choosing what gets GPU time means guessing outcomes in advance.

They used an LLM as a research world model that predicts an experiment's gain before it runs. They tested it on 2,653 real experiment records from 9 setups, from pretraining to inference.

Past records from the same setup raised average ranking correlation with actual results from 0.506 to 0.774 across 5 setups. On the OLMo3-100M setup, adding records at low reasoning effort scored 0.892, while max effort without records reached 0.648.

If you run research agents, log every experiment, including failures, and feed those records to whatever model picks the next run.

[Aakash Gupta @aakashgupta]

The guy who invented Amazon Prime disappeared for six years. He came back this week with $250 million and a factory that turns lines of code into tiny working machines. Type different code, a different machine comes out the other end.

That sounds fake. Here's why it took six years.

A chip fab is the most precise factory humans have ever built, and it makes exactly one kind of thing. Flat silicon that processes information. The moment you want moving parts, multiple materials, or real 3D structure, you leave the fab for a world where getting one new device into production takes years and tens of millions of dollars. Engineers call it one product, one process. It's the reason micro-machines never got their Moore's Law.

Holden's Matter Compiler kills the per-product process entirely. Features down to a few microns, no new tooling, and the system measures and corrects every step while it builds, then feeds what it learned into the next design.

His first product is the tell on how sharp the strategy is.

PrimeSwitch is a 150-amp relay for 800-volt power in AI data centers. It opens in roughly 50 microseconds, about 1,000x faster than a normal contactor. So the first thing to come out of a universal factory is aimed at the one market where demand is basically infinite and a faster switch means less fried hardware.

Neal Stephenson wrote the Matter Compiler as fiction in 1995. Holden just shipped the first real product out of one, and it's a light switch for AI data centers.

[Jason Smith - 上官杰文 @ShangguanJiewen]

China isn't holding back.

If the USA is determined to play "Trade-War Tantrum" China is just going to move on, and leave the USA in a museum of 20th century artifacts.

The ONLY way the US advances is WITH China.

[The Bitcoin Historian @pete_rizzo_]

JUST IN: JACK DORSEY'S BITCHAT APP THAT ALLOWS YOU TO SEND #BITCOIN WITHOUT INTERNET WAS JUST BANNED IN INDIA

AN APP SO POWERFUL, GOVERNMENTS AROUND THE WORLD ARE TRYING TO BAN IT

A MONEY SO POWERFUL, GOVERNMENTS AROUND THE WORLD ARE TRYING TO STOP IT

ONE BILLIONAIRE STANDING UP FOR FREEDOM

WHAT AN ABSOLUTE LEGEND 🔥

[Tom Zaragoza @tomzaragoza]

Connect Google Search Console to ChatGPT

with the https://t.co/MaP8m570bz MCP

Try it for free!

↓

[Tavus @tavus]

Introducing Griffin, the first model to pass the video Turing test.

48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video.

It’s the first Human Interaction Model (HIM).

[Tom Zaragoza @tomzaragoza]

Looking for a Google Search Console MCP?

Try https://t.co/MaP8m570bz for free! ↓

[Kacee Allen @KaceeRAllen]

A federal court told Honda, Toyota, Volkswagen, and GM they can legally record and intercept your private texts and call logs the second you plug your phone into the car.

The infotainment system downloads everything. You can’t see it. You can’t delete it. Police can.

The judge said that’s not a privacy violation unless you prove it already damaged your reputation or your business.

You paid tens of thousands of dollars for a rolling wiretap. This is the surveillance state on four wheels.

[Dickie Bush @dickiebush]

I am begging you to start writing by hand to preserve your ability to think.