每日科技推特精选 - 2026年4月1日

2026年4月1日科技每日简报

Today's top tech conversations are led by @chhddavid, whose post about 'so you're telling me Claude Co...' garnered the highest engagement. Key themes trending across the top stories include https, claude, intelligence, models, apple. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. chhddavid (Group Score: 518.7 | Individual: 35.5)

Cluster: 21 tweets | Engagement: 36 (Avg: 26) | Type: Tech

so you're telling me Claude Code Opus 4.6 can now...

build a full company
launch it for me
do email marketing
self-build features

without any human in the loop??

IT'S SO OVER... https://t.co/sB0z3H4WcH\n\nQT @shipper_now: Today, we've ended vibe coding as a whole.

I just watched my Mac build a complete business in 183 seconds.

This is just absurd. https://t.co/63YBTIvKyn

See 20 related tweets

@chhddavid: you don’t understand how big this is...

you can now build real apps with Claude Code Opus 4.6 (not ...

@chhddavid: Day 242 of building in public

FINALLY, our account @shipper_now started to blow up and we've gone p...

@chhddavid: FINALLY!! Everyone gets their tony stark moment.

> introducing Claude as your AI business builde...

@chddaniel: Anthropic CEO:

“I think we might be six to 12 months away from when Claude doing most or maybe all ...

@chhddavid: AUTONOMOUS AI VIBE CODING BUILDING IS HERE

> ai co-founder > launches apps for me > 100+ c...

2. QuixiAI (Group Score: 452.6 | Individual: 45.0)

Cluster: 21 tweets | Engagement: 11156 (Avg: 1056) | Type: Tech

RT @Fried_rice: Claude code source code has been leaked via a map file in their npm registry!

Code: https://t.co/jBiMoOzt8G https://t.co/rYo5hbvEj8

See 20 related tweets

@RoundtableSpace: CLAUDE CODE SOURCE CODE JUST LEAKED VIA THEIR NPM REGISTRY.

EVERY FILE. EVERY FOLDER. ALL OF IT.

Y...

@WesRoth: Earlier today, the complete, un-obfuscated TypeScript source code for Anthropic’s official Claude Co...
@rohanpaul_ai: The 12-Hour Claude Code Saga:
Today around midday, security researcher Chaofan Shou spotted the ...
@edzitron: https://t.co/wqYTF7X9JM\n\nQT @Fried_rice: Claude code source code has been leaked via a map file in...
@irl_danB: the “developer” who leaked the source map was almost certainly claude code

this was a self exfiltra...

3. adonis_singh (Group Score: 228.6 | Individual: 45.1)

Cluster: 13 tweets | Engagement: 2293 (Avg: 406) | Type: Tech

"i went through all 600,000 lines of code"

😂\n\nQT @AlexFinn: Unreal. The entire Claude Code source code just leaked

It reveals EVERY secret Anthropic has in store for Claude

I went through all 600,000 lines of code

Here's EVERYTHING juicy detail you need to know about how Claude Code is built and what is coming next: https://t.co/WAsDKkBeQ1

See 12 related tweets

@tunguz: TL;DR - we're about to have open source Claude Code in Rust\n\nQT @altryne: If you, like me, just wo...
@aakashgupta: Claude Code’s source code was leaked. Here’s what you can learn from it.\n\nQT @mal_shaik: https://t...
@MLStreetTalk: RT @DavidKPiano: Ironically this is probably the first time that actual humans are carefully & t...
@SIGKITTEN: I made this leaked claude-code source build so we can play with it before they take it down. enjoy h...
@NielsRogge: Hang it in the Louvre https://t.co/k8E3K5H4IH\n\nQT @Fried_rice: Claude code source code has been le...

4. scaling01 (Group Score: 224.2 | Individual: 46.0)

Cluster: 10 tweets | Engagement: 1450 (Avg: 303) | Type: Tech

A few take-aways from the Claude Code Leak:

Anthropic is actively using Capybara (Mythos) for development
they are already at Capybara v8
Capybara still has issues with over-commenting and false-claims
Capybara has 1M context and fast mode
Numbat is another interesting code name tagged with "@[MODEL LAUNCH]: Remove this section when we launch numbat."
Fennec seems to be Opus 4.6\n\nQT @Fried_rice: Claude code source code has been leaked via a map file in their npm registry!

Code: https://t.co/jBiMoOzt8G https://t.co/rYo5hbvEj8

See 9 related tweets

@Meer_AIIT: 🚨Breaking: Anthropic accidentally leaks Claude Code source code via source map file in npm package\n...
@scaling01: they have kairos, dreaming, ultrathink, ultraplan, ultrareview and complete integration with github ...
@cryptopunk7213: alright i've combed through the entire anthropic leak, honestly can't believe this shit is public. i...
@AndrewCurran_: RT @scaling01: A few take-aways from the Claude Code Leak:
Anthropic is actively using Capybara (M...
@mark_k: Claude Code source code leaked:\n\nQT @Fried_rice: Claude code source code has been leaked via a map...

5. Meer_AIIT (Group Score: 183.9 | Individual: 33.7)

Cluster: 7 tweets | Engagement: 6 (Avg: 39) | Type: Tech

🚨12 months ago OpenAI made $1B in a quarter.

now they make $2B in a month.

they just raised $122 billion at an$ 852B valuation and buried the most important numbers inside the announcement:

900M weekly active users

6x more web traffic than the second largest AI app

APIs processing 15 billion tokens per minute

Codex users up 5x in 3 months

but the revenue curve is what should actually stop you:

$1B/year →$ 1B/quarter → $2B/month

they’re growing 4x faster than Google and Meta did at this stage.

the flywheel they’re building: More compute = smarter models Smarter models = better products Better products = more revenue More revenue = more compute

they’re also turning ChatGPT into a full superapp, one surface for browsing, coding, agents, everything.\n\nQT @OpenAI: Today, we closed our latest funding round with $122 billion in committed capital at an$ 852B post-money valuation.

The fastest way to expand AI’s benefits is to put useful intelligence in people’s hands early and let access compound globally.

This funding gives us resources to lead at scale. https://t.co/sY7YNUPSYO

See 6 related tweets

@cryptopunk7213: very impressive from openai, massive raise and new ways for retail to invest
announced a $122B ra...
@wallstengine: OpenAI is now generating $2B in revenue per month.

The company says enterprise now makes up 40% of ...

@business: OpenAI has completed a deal to raise $122 billion from investors at an$ 852 billion valuation, marki...
@Forbes: OpenAI Valuation Reaches $852 Billion After Massive Funding Round https://t.co/TcvjQodp8i https://t....
@AndrewCurran_: OpenAl raised $122 billion, at a valuation of$ 852 billion.

They also posted about the OpenAI Supe...

6. seraleev (Group Score: 166.7 | Individual: 33.8)

Cluster: 7 tweets | Engagement: 31 (Avg: 57) | Type: Tech

On one hand, I understand that this is a major hit for companies building apps with vibe coding. They will lose one of their key user acquisition channels.

But on the other hand, Apple is reducing the number of repetitive, template-based apps that overload the App Review team.

What’s interesting is that Apple didn’t even need to change its guidelines. All vibe-coded apps fall under Guideline 2.5.2 (download, install, or execute code which introduces or changes features or functionality of the app).

I expect Apple to keep tightening things. They are clearly focused on higher-quality apps.\n\nQT @seraleev: Apple just removed another vibe coding app: Anything AI

Reason: violation of Guideline 2.5.2

Apps should be self-contained in their bundles, and may not read or write data outside the designated container area, nor may they download, install, or execute code which introduces or changes features or functionality of the app, including other apps.

See 6 related tweets

@seraleev: Officially: vibe coding on iOS apps is dead https://t.co/Th4PUgBKmM\n\nQT @seraleev: Apple just remo...
@seraleev: https://t.co/WRuuxWSF3b\n\nQT @seraleev: Apple just removed another vibe coding app: Anything AI

Re...

@seraleev: Meanwhile, the Replit team… https://t.co/Zoym7YCM7P\n\nQT @seraleev: Apple just removed another vibe...
@amasad: RT @dee_bosa: Apple's crackdown on vibe coding apps puts it on the wrong side of history

vibe codin...

@seraleev: Vibe coding effect

the number of monthly app submissions to the App Store had risen by 56% by Decem...

7. Shipper_now (Group Score: 164.9 | Individual: 29.7)

Cluster: 7 tweets | Engagement: 8 (Avg: 7) | Type: Tech

Introducing Claude Opus 4.6 to Figma.

Turn your Figma design into a full-scale app.

0 API tokens
0 confusing cursor
0 bankruptcy on lovable

Just paste a Figma URL into Shipper and Claude Opus 4.6 will turn it into a launchable app.

Everything 1:1 as in Figma... https://t.co/rSy1NYPBiA

See 6 related tweets

@Shipper_now: Today, Claude Opus 4.6 meets Figma.

Shipper turns any Figma design into a launch-ready app...

→ Tu...

@Shipper_now: Claude 🤝 Figma

Introducing the Figma connector in Shipper. From today on, CC Opus 4.6 can scan your...

@chddaniel: Introducing the Shipper Figma connector.

From today on, you can turn Figma designs into real apps. ...

@chddaniel: This is honestly scary.

What happens when you give Claude Opus 4.6 access to Figma?

→ Turn designs...

@chddaniel: Claude Opus 4.6 meets Figma in @shipper_now!

Shipper turns any Figma design into an app:

→ Turn de...

8. ivanfioravanti (Group Score: 157.0 | Individual: 37.9)

Cluster: 5 tweets | Engagement: 218 (Avg: 72) | Type: Tech

242 toks/s on M5 Max 🚀 I need to run evals on this magic! https://t.co/dioSmZRc2D\n\nQT @PrismML: Today, we are emerging from stealth and launching PrismML, an AI lab with Caltech origins that is centered on building the most concentrated form of intelligence.

At PrismML, we believe that the next major leaps in AI will be driven by order-of-magnitude improvements in intelligence density, not just sheer parameter count.

Our first proof point is the 1-bit Bonsai 8B, a 1-bit weight model that fits into 1.15 GBs of memory and delivers over 10x the intelligence density of its full-precision counterparts. It is 14x smaller, 8x faster, and 5x more energy efficient on edge hardware while remaining competitive with other models in its parameter-class. We are open-sourcing the model under Apache 2.0 license, along with Bonsai 4B and 1.7B models.

When advanced models become small, fast, and efficient enough to run locally, the design space for AI changes immediately. We believe in a future of on-device agents, real-time robotics, offline intelligence and entirely new products that were previously impossible.

We are excited to share our vision with you and keep working in the future to push the frontier of intelligence to the edge.

See 4 related tweets

@eliebakouch: very interesting to see a new company pushing the frontier of "intelligence per bits"! https://t.co/...
@ClementDelangue: RT @PrismML: Today, we are emerging from stealth and launching PrismML, an AI lab with Caltech origi...
@_xjdr: Really excited about what @HessianFree and co are building at Prism. It's amazing someone got 1bit ...
@ferologics: HOLY BASED\n\nQT @PrismML: Today, we are emerging from stealth and launching PrismML, an AI lab with...

9. Gorden_Sun (Group Score: 154.3 | Individual: 28.8)

Cluster: 8 tweets | Engagement: 233 (Avg: 57) | Type: Tech

Claude Code源码泄露我让Claude Code分析了，基本确定是Claude Code源码，而且是比较新的版本，因为支持claude-opus-4-6。下载地址：https://t.co/7x3H6ZCoyy https://t.co/msIa27LHcV

See 7 related tweets

@lifesinger: 受 Claude Code 泄漏的源码启发我们立刻做出了下面这个 Skill 可一键发微信公众号文章

我不知道有没有用但值得一试\n\nQT @liu10102525: https://t.co...

@wshuyi: 我想测试一下 Claude Code 中的 Codex 官方 Plugin 工作效果，于是让 Claude Code 编写了一段代码，让 Codex 审核。我没提具体要求，结果 Claude 说自己一...
@Gorden_Sun: 几个小发现： 1、Anthropic内部员工有单独的订阅等级； 2、毫无悬念采集的数据非常多； 3、明天愚人节会上线Buddy彩蛋，上线后会一直可用，一共18种，稀有度不同看你能抽中哪个了； http...
@GitHub_Daily: 用 Claude Code 编码，有时候开几个窗口同时跑，切来切去，谁在干什么、干到哪了，全靠脑子记，一乱就容易翻车。

于是找到 Claude Agent Teams UI 这个工具，给 Claud...

@Gorden_Sun: 速度也太快了。让Claude Code分析自己的源码，来重写Claude Agent SDK，以后Agent基建可以用这个了。\n\nQT @idoubicc: 让 Claude Code 分析了一遍...

10. FirstSquawk (Group Score: 140.1 | Individual: 27.3)

Cluster: 8 tweets | Engagement: 531 (Avg: 129) | Type: Tech

IRAN'S IRGC: ISSUED A WARNING AGAINST 18 AMERICAN TECHNOLOGY COMPANIES, AMONG THEM MICROSOFT, APPLE, GOOGLE, INTEL, AND BOEING, STATING THAT THEY WILL BE CONSIDERED LEGITIMATE TARGETS IN RESPONSE TO TERRORIST OPERATIONS CARRIED OUT BY THE US AND ISRAEL.

See 7 related tweets

@FirstSquawk: IN THE STATEMENT, THE IRGC SAID THAT AMERICAN ICT AND AI FIRMS PLAY A CENTRAL ROLE IN DESIGNING AND ...
@FirstSquawk: Iran has warned that it may attack major U.S. companies’ operations in the Middle East if the confli...
@AndrewCurran_: Iran's IRGC released a statement today warning that it will target the regional facilities of major ...
@FirstSquawk: THE ISLAMIC REVOLUTIONARY GUARD CORPS (IRGC) WARNED AMERICAN COMPANIES ACTIVE IN THE FIELDS OF INFOR...
@zerohedge: RT @Breaking911: 🚨 BREAKING: The IRGC has reportedly threatened to target the Middle East operations...

11. e2b (Group Score: 128.6 | Individual: 28.2)

Cluster: 7 tweets | Engagement: 21 (Avg: 9) | Type: Tech

Big news: we're on the 2026 Enterprise Tech 30!

We’re proud to be included alongside our customers and partners - and excited to keep building the future together.

The ET30 is voted on by 90+ of the top investors, recognizing private companies shaping the future of enterprise technology.

Thank you @Wing_VC and @EricNewcomer for including us on the list. Now back to building.

https://t.co/kvluUaxBvU #ET30

@getbasis, @braintrustdata, @browserbasehq, @ExaAILabs, @MadeByGamma, @granaborr, @linearapp, @OpenRouterAI, @tryparallel, @RilletHQ, @ServalAI, @turbopuffer, @WisprFlow, @AbnormalSec, @canva, @clay, @DecagonAI, @elevenlabsio, @FireworksAI_HQ, @glean, @Gong_io, @harvey, @WeAreLegora, @lovable, @n8n_io, @EvidenceOpen, @SierraPlatform, @supabase, @anduriltech, @AnthropicAI, @AppliedInt, @ClickHouseDB, @cursor_ai, @databricks, @OpenAI, @tryramp, @stripe, @SpaceX, and others

See 6 related tweets

@llama_index: LlamaIndex is proud to be named to the 2026 Enterprise Tech 30, #3 in the Early Stage category.

The...

@e2b: RT @mlejva: Big news - @e2b is one of the 2026 Enterprise Tech 30 firms.

Very proud to be one of th...

@e2b: RT @dobrac8: E2B made the 2026 Enterprise Tech 30 -- Early Stage category. The later stages include ...
@databricks: The 2026 Enterprise Tech 30 list is out, and Databricks has been named #2 in the Giga Stage!

Over 9...

@e2b: RT @badphilosopher: E2B just made the 2026 Enterprise Tech 30.

The list includes Anthropic, OpenAI,...

12. ycombinator (Group Score: 123.0 | Individual: 44.9)

Cluster: 5 tweets | Engagement: 1166 (Avg: 79) | Type: Tech

RT @ollama: Ollama is now updated to run the fastest on Apple silicon, powered by MLX, Apple's machine learning framework.

This change unlocks much faster performance to accelerate demanding work on macOS:

Personal assistants like OpenClaw
Coding agents like Claude Code, OpenCode, or Codex

See 4 related tweets

@awnihannun: You can now run LMs with Ollama + MLX!

I've been waiting for this moment since MLX was first open s...

@Prince_Canuma: You can now run Ollama using MLX as a backend 🚀\n\nQT @ollama: Ollama is now updated to run the fast...
@ivanfioravanti: Ollama has got MLX finally! You did it @ollama 🚀 I have to deep dive on this ASAP!\n\nQT @ollama: Ol...
@Shashikant86: This is the new I was waiting for! @ollama finally delivered the MLX integration!

Look forward to ...

13. EHuanglu (Group Score: 119.9 | Individual: 29.0)

Cluster: 5 tweets | Engagement: 98 (Avg: 136) | Type: Tech

Google just dropped Veo 3.1 Lite

it generates motion graphics in seconds, you can define art direction and text style https://t.co/fIpEdWWMeg\n\nQT @OfficialLoganK: Video’s here to stay - introducing Veo 3.1 Lite, our most cost efficient video generation model to date, and on April 7th we are also reducing the price for Veo 3.1 Fast : ) https://t.co/s7OMZaMPho

See 4 related tweets

@WesRoth: Google released the API specifications and pricing structure for Veo 3.1 Lite for developers.

Veo 3...

@FlowbyGoogle: Veo 3.1 Lite is now available in Flow. It’s perfect for fast prompt iterations and storyboarding bef...
@rseroter: RT @googleaidevs: Introducing Veo 3.1 Lite, now available via the Gemini API and @GoogleAIStudio. Ve...
@Techmeme: Google launches Veo 3.1 Lite, costing <50% of Veo 3.1 Fast and meant for "high-volume video appli...

14. Yuchenj_UW (Group Score: 111.2 | Individual: 41.7)

Cluster: 3 tweets | Engagement: 5892 (Avg: 1494) | Type: Tech

> Anthropic leaked Claude Code source code > someone forked it > 32.6k stars, 44.3k forks > got scared of getting sued > convert the whole codebase from TypeScript to Python with Codex

AI is quietly erasing copyright. https://t.co/75zN9kNhaE

See 2 related tweets

@TheAhmadOsman: The # of DMCAs Anthropic is about to send is going to be crazzzy

I personally will spin up 3 Hermes...

@arvidkahl: Either panic-porting a leaked TypeScript app to Python to avoid being sued for copyright infringemen...

15. Austen (Group Score: 110.5 | Individual: 44.7)

Cluster: 4 tweets | Engagement: 1847 (Avg: 253) | Type: Tech

I’ll simply say this:

Everyone I know working at a frontier AI model company is either suffering from extreme AI psychosis or genuinely believes we’re single digit months away from liftoff.\n\nQT @Hadley: I’ve heard from 2 people in the last 2 days that internally Anthropic expects to have AGI in 6-12 months. That’s faster than Dario has stated publicly. Plan your business and personal finances appropriately.

See 3 related tweets

@chatgpt21: So someone heard someone who knew someone at Anthropic expects to have AGI in 6–12 months.

I’m not ...

@scaling01: this really depends on the definition

but once AI automates research you can't really argue anymor...

@StefanFSchubert: Completely devoid of meaning https://t.co/qR5HeeQhPa\n\nQT @Hadley: I’ve heard from 2 people in the ...

16. wallstengine (Group Score: 98.6 | Individual: 60.7)

Cluster: 3 tweets | Engagement: 1177 (Avg: 133) | Type: Tech

WARREN BUFFETT SAYS IF THERE IS A BIG MARKET DECLINE, BERKSHIRE WILL DEPLOY CASH

"I am still involved in making investments at Berkshire, we have north of $350B in cash on hand; Things don't look substantially cheaper, not in it to make 5-6% which is what indices are down" https://t.co/liVQZ2MJ6U

See 2 related tweets

@Cointelegraph: ⚡️ INSIGHT: Warren Buffett sees limited opportunities in current market, waiting for a major dip to ...
@unusual_whales: Warren Buffett says if there is a big market decline, Berkshire will deploy cash, per CNBC...

17. Prince_Canuma (Group Score: 96.7 | Individual: 34.2)

Cluster: 3 tweets | Engagement: 65 (Avg: 91) | Type: Tech

Congratulations to @liquidai on this awesome release!

LFM2.5-350M comes with day-0 MLX support. 🚀\n\nQT @liquidai: Today, we release LFM2.5-350M. Agentic loops at 350M parameters.

A 350M model trained for reliable data extraction and tool use, where models at this scale typically struggle.

<500MB when quantized, built for environments where compute, memory, and latency are constrained.

🧵 https://t.co/zZPKzcCwH9

See 2 related tweets

@ivanfioravanti: sci-fi! You rock!\n\nQT @liquidai: Today, we release LFM2.5-350M. Agentic loops at 350M parameters. ...
@_lewtun: RT @liquidai: Today, we release LFM2.5-350M. Agentic loops at 350M parameters.

A 350M model trained...

18. seraleev (Group Score: 91.6 | Individual: 33.5)

Cluster: 3 tweets | Engagement: 34 (Avg: 57) | Type: Tech

Are you ready to trust your ads to an AI agent?

I’m not.

I’ve already seen Google Ads burn my budget with no results. Making money is hard. Spending it is easy\n\nQT @RyanZambrano8: Introducing: the first AI agent that runs your paid ads  The future of paid advertising is autonomous.  RT+ comment “Pixero” and we’ll make an ad for your brand for FREE. https://t.co/ETmcOCII0o

See 2 related tweets

@EHuanglu: this AI agent can run product ad campaigns now

it can analyze your website, then automatically crea...

@shiri_shh: This AI Agent just killed a bunch of ad agencies

Pixero scrapes your site, builds the strategy, mak...

19. tunguz (Group Score: 91.5 | Individual: 33.0)

Cluster: 3 tweets | Engagement: 100 (Avg: 68) | Type: Tech

LOL Python always wins\n\nQT @GergelyOrosz: This is either brilliant or scary:

Anthropic accidentally leaked the TS source code of Claude Code (which is closed source). Repos sharing the source are taken down with DMCA.

BUT this repo rewrote the code using Python, and so it violates no copyright & cannot be taken down! https://t.co/uSrCDgGCAZ

See 2 related tweets

@Rasmic: Will i get in trouble if I do a video covering this?\n\nQT @GergelyOrosz: This is either brilliant o...
@chrisalbon: So many things are based on an implicit assumption that someone could do something but they won’t ...

20. unusual_whales (Group Score: 87.5 | Individual: 34.6)

Cluster: 3 tweets | Engagement: 624 (Avg: 2780) | Type: Tech

Anthropic accidentally leaked Claude Code's full source code via a public npm package, same day as AI security startup launches with $80M in funding, per reports.\n\nQT @andreamichi: depthfirst has raised an$ 80M Series B at a $580M valuation.

Attackers are using AI to break into systems faster than ever before. depthfirst is on a mission to stop this.

RT + Comment “depthfirst” and I’ll send you a FREE vibe coding security agent. https://t.co/EOpQ7Nj9Lv

See 2 related tweets

@shiri_shh: BIGGEST AI LEAK OF 2026 JUST DROPPED

Anthropic “super safe” AI company accidentally shipped its ent...

@shiri_shh: Holy shit, talk about perfect timing.

Today, Anthropic accidentally leaked the entire source code o...