科技推特日报 - 2026年3月31日

2026年3月31日科技每日简报

Today's top tech conversations are led by @HamelHusain, whose post about 'RT @claudeai: Computer use is ...' garnered the highest engagement. Key themes trending across the top stories include https, codex, claude, review, plugin. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. HamelHusain (Group Score: 573.9 | Individual: 57.3)

Cluster: 18 tweets | Engagement: 5726 (Avg: 280) | Type: Tech

RT @claudeai: Computer use is now in Claude Code.

Claude can open your apps, click through your UI, and test what it built, right from the CLI.

Now in research preview on Pro and Max plans. https://t.co/s2FDQaDmr1

See 17 related tweets

@rohanpaul_ai: Anthropic's release speed is insane

Computer use is now in Claude Code.

i.e. In one prompt, Claud...

@tunguz: Claude just solved GUI.\n\nQT @claudeai: Computer use is now in Claude Code.

Claude can open your a...

@minchoi: Claude just killed many software 😬 https://t.co/ICtay9QxiL\n\nQT @claudeai: Computer use is now in C...
@Yuchenj_UW: Claude Code is turning into a super app.

There isn’t a single day this team isn’t shipping. The pac...

@kimmonismus: Another day, another Claude release:

Computer use now in Claude code\n\nQT @claudeai: Computer use...

2. shiri_shh (Group Score: 201.1 | Individual: 34.9)

Cluster: 6 tweets | Engagement: 627 (Avg: 797) | Type: Tech

babe wake up.

the world’s first AI marketer just launched.

https://t.co/72iXbBFmr1\n\nQT @SeijinJung: Introducing Helena: the world's first autonomous AI marketer.

Businesses spend 4,000 hours on marketing…before their first $1M in revenue.

We built Helena to solve this. Helena can:

➤ Track competitor ads & create TikTok slideshows, UGC, static ads - all while you sleep ➤ Analyze performance across GA4, Search Console, paid/organic social for daily insights ➤ Research trends to draft GEO optimized blogs directly on WordPress, Framer, Webflow

...and more

Helena has her own memory, scheduled tasks, 100+ custom marketing tools and native integrations.

No dev. No CLI. No n8n. No API keys needed.

Helena doesn't replace CMOs, and every marketer who's demoed it has asked us for early access.

Want to hire her? Check the next thread ⬇️

See 5 related tweets

@godofprompt: The "AI marketing tools" category spent five years building tools that made human marketers faster. ...
@alexcooldev: I’m building an AI like this, but this guys already built it before me and it went viral. 💪\n\nQT @S...
@alex_prompter: I've watched founders spend $120k and 18 months trying to get marketing to actually work.

Three age...

@heyshrutimishra: how me and Helena "collaborate" on the work I used to hire people for https://t.co/8xlZ4ZDmoW\n\nQT ...
@svpino: Marketing (and selling a product) is harder than building it.

This is exactly the type of work that...

3. mervenoyann (Group Score: 192.1 | Individual: 42.3)

Cluster: 5 tweets | Engagement: 588 (Avg: 173) | Type: Tech

RT @Alibaba_Qwen: 🚀 Qwen3.5-Omni is here! Scaling up to a native omni-modal AGI.

Meet the next generation of Qwen, designed for native text, image, audio, and video understanding, with major advances in both intelligence and real-time interaction.

A standout feature: 'Audio-Visual Vibe Coding'. Describe your vision to the camera, and Qwen3.5-Omni-Plus instantly builds a functional website or game for you.

Offline Highlights: 🎬 Script-Level Captioning: Generate detailed video scripts with timestamps, scene cuts & speaker mapping. 🏆 SOTA Performance: Outperform Gemini-3.1 Pro in audio and matches its audio-visual understanding. 🧠 Massive Capacity: Natively handle up to 10h of audio or 400s of 720p video, trained on 100M+ hours of data. 🌍 Global Reach: Recognize 113 languages (speech) & speaks 36.

Real-time Features: 🎙️ Fine-Grained Voice Control: Adjust emotion, pace, and volume in real-time. 🔍 Built-in Web Search & complex function calling. 👤 Voice Cloning: Customize your AI's voice from a short sample, with engineering rollout coming soon. 💬 Human-like Conversation: Smart turn-taking that understands real intent and ignores noise.

The Qwen3.5-Omni family includes Plus, Flash, and Light variants.

Try it out: Blog: https://t.co/yuSAz3DuO8 Realtime Interaction: click the VoiceChat/VideoChat button (bottom-right): https://t.co/nnAW9ZfRet HF-Demo: https://t.co/rLsqejKgCG HF-VoiceOnline-Demo: https://t.co/LIGtmITeSw API-Offline: https://t.co/lNE7fH5YUt API-Realtime: https://t.co/9A3lopXGwV

See 4 related tweets

@testingcatalog: Qwen released the Qwen3.5-Omni model family, designed for native understanding of text, images, audi...
@ivanfioravanti: RT @Ali_TongyiLab: 1/10 🚀 Qwen3.5-Omni is here! Scaling up to a native omni-modal AGI. Meet the next...
@MatthewBerman: Qwen cooked, but will they run out of steam soon?\n\nQT @Alibaba_Qwen: 🚀 Qwen3.5-Omni is here! Scali...
@kimmonismus: Alibaba's Qwen3.5-Omni just dropped with script-level captioning, audio-visual vibe coding, and real...

4. etnshow (Group Score: 174.5 | Individual: 35.7)

Cluster: 6 tweets | Engagement: 69 (Avg: 31) | Type: Tech

BREAKING: Starcloud has raised a $170M Series A at a$ 1.1B valuation to build data centres in space.

CEO and founder @PhilipJohnston says: "The AI revolution is colliding with the physical limits of our terrestrial energy grid".

"We are quickly running out of places to build new energy projects for data centers on Earth. By moving AI compute to space, we unlock access to unlimited solar power and completely remove the energy bottleneck".

@Starcloud_'s Series A, which closed 17 months after its demo day presentation, was led by @benchmark and @eqt.

See 5 related tweets

@ycombinator: RT @PhilipJohnston: I am super excited to share that @Starcloud_ has raised a $170M Series A at a$ 1...
@shiri_shh: Why is no one talking about DATA CENTRE in Space?

Starcloud just raised $170 million in Series A hi...

@wallstengine: Space data center startup Starcloud raised a $170M Series A led by Benchmark and EQT at a$ 1.1B valu...
@Techmeme: Space data center startup Starcloud raised a $170M Series A led by Benchmark and EQT at a$ 1.1B valu...
@garrytan: RT @StockSavvyShay: Starcloud (fastest unicorn in Y Combinator history) just raised a $170M Series A...

5. badlogicgames (Group Score: 142.8 | Individual: 33.0)

Cluster: 5 tweets | Engagement: 297 (Avg: 128) | Type: Tech

ok, this is hilarious\n\nQT @romainhuet: We’ve seen Claude Code users bring in Codex for code review and use GPT-5.4 for more complex tasks, so we thought: why not make that easier?

Today we’re open sourcing a plugin for it! You can call Codex from Claude Code with your ChatGPT subscription.

We love an open ecosystem!

See 4 related tweets

@TheRealAdamG: You can use Codex everywhere, including inside Claude Code now. https://t.co/w8aiG6YlkH\n\nQT @dkund...
@TheRealAdamG: https://t.co/t9IttkmaSN\n\nQT @dkundel: I built a new plugin! You can now trigger Codex from Claude ...
@omarsar0: This is really cool! I was already calling Codex inside of Claude Code, but this plugin should make ...
@reach_vb: btw the codex for claude code plugin is open sauce - go play around with it, tinker and make it fit ...

6. Techmeme (Group Score: 120.8 | Individual: 26.5)

Cluster: 6 tweets | Engagement: 14 (Avg: 13) | Type: Tech

OpenAI's ChatGPT app store now has 300+ app integrations six months after its launch, but faces sluggish adoption due to limited functionality for many apps (@natlungfy / Bloomberg)

https://t.co/pnmjCM8AHL

https://t.co/cS471sPHLB

📥 Send tips! https://t.co/wlNZvXuhJs

See 5 related tweets

@kimmonismus: Seriously, what happened to the ChatGPT App Store? Is anyone using it?\n\nQT @business: OpenAI said ...
@aakashgupta: This is OpenAI's second failed app store in two years.

The GPT Store launched January 2024 with 3 m...

@business: OpenAI said its ChatGPT app store is central to its strategy, but a number of issues have led to a l...
@markgurman: OpenAI’s ChatGPT App Store Took Aim at Apple, But Results Lag So Far — Six months in, ChatGPT’s thir...
@business: RT @natlungfy: NEW: OpenAI said its ChatGPT app store is central to its strategy, but a number of is...

7. btibor91 (Group Score: 114.5 | Individual: 29.2)

Cluster: 4 tweets | Engagement: 88 (Avg: 412) | Type: Tech

This looks useful (Claudex?)

Using both models almost daily - very different "personalities", like asking two different developers to review/plan how to fix things\n\nQT @dkundel: I built a new plugin! You can now trigger Codex from Claude Code!

Use the Codex plugin for Claude Code to delegate tasks to Codex or have Codex review your changes using your ChatGPT subscription.

Start by installing the plugin: https://t.co/u6gBpArwBc https://t.co/HyEdMPWees

See 3 related tweets

@nummanali: Orchestrating Codex from Claude Code is not a bad idea!

OpenAI dev team have officially released a...

@ajambrosino: RT @dkundel: I built a new plugin! You can now trigger Codex from Claude Code!

Use the Codex plugin...

@TheRealAdamG: Codex 🤝 Everywhere\n\nQT @dkundel: I built a new plugin! You can now trigger Codex from Claude Code!...

8. MorePerfectUS (Group Score: 114.4 | Individual: 27.2)

Cluster: 5 tweets | Engagement: 1031 (Avg: 728) | Type: Tech

65% of Americans now oppose the building of AI data centers in their community, with the top concerns being electricity costs and water use, according a to Quinnipiac poll.

See 4 related tweets

@Techmeme: Quinnipiac poll: 55% of Americans say AI will do more harm than good in their day-to-day lives, and ...
@secureainow: POLL: 65% of Americans oppose building AI data centers in their communities. Among opponents, concer...
@Miles_Brundage: RT @metzgov: Some interesting stuff on American views on AI:
Just 27% have never used it
55% th...
@ednewtonrex: RT @PatrickSvitek: In new Quinnipiac poll, Americans oppose building AI data centers in their commun...

9. rohanpaul_ai (Group Score: 114.3 | Individual: 32.6)

Cluster: 6 tweets | Engagement: 45 (Avg: 76) | Type: Tech

Mistral just raised $830M to build Nvidia-powered AI data centers in Europe and turn European demand for sovereign AI into its own infrastructure business.

The bet is that many companies and governments no longer want to rent core AI capacity only from US cloud groups, because the model, the servers, and the control layer all sit outside Europe.

Over half of Mistral’s revenues come from Europe

Mistral was valued at nearly €12bn last year in a €1.7bn equity financing led by chip equipment maker ASML. It was on track to surpass $1bn in annual recurring revenue by the end of this year

Mistral’s pitch is a full-stack setup, which means selling the model, the software around it, and the computing layer together so customers can run custom systems without handing the whole stack to Microsoft, Amazon, or Google.

The financing is debt, not fresh equity, which tells you Mistral thinks these centers can produce predictable cash flow fast enough to repay lenders from real usage rather than distant hope.

The sharp question is whether Europe’s hunger for AI autonomy stays strong enough to fill all this new capacity before cheaper outside clouds close the gap on price and convenience.

See 5 related tweets

@Techmeme: Paris-based AI startup Mistral raised $830M in its debut debt financing to build Nvidia-powered data...
@Cointelegraph: 🇪🇺 LATEST: Mistral AI raises $830M in debt financing to build NVIDIA powered data centers across Eur...
@testingcatalog: Mistral AI raised $830 million in debt to operate its new data center near Paris, powered by NVIDIA ...
@TechCrunch: Mistral AI raises $830M in debt to set up a data center near Paris https://t.co/1ACqx3clQT...
@CNBC: Mistral secures $830 million in debt financing to fund AI data center https://t.co/CWIp07ysZ7...

10. reach_vb (Group Score: 103.4 | Individual: 39.0)

Cluster: 5 tweets | Engagement: 3393 (Avg: 574) | Type: Tech

Starting today you can use Codex in Claude Code 👀

/plugin marketplace add openai/codex-plugin-cc

Try it out today with: /codex:review for a normal read-only Codex review /codex:adversarial-review for a steerable challenge review /codex:rescue to let codex rescue your code

Enjoy Codex-ing!\n\nQT @reach_vb: https://t.co/II9ILcpPUX

See 4 related tweets

@Techmeme: OpenAI introduces a Codex plugin for Claude Code, letting users invoke Codex from inside Claude Code...
@ns123abc: Sam wouldn’t even hold Dario’s hand…

How desperate is openai that they’re now offering codex in cl...

@manthanguptaa: This is a real dario and sam hand holding moment https://t.co/THo6THub72\n\nQT @reach_vb: Starting t...
@Shashikant86: "If you can't beat them, join them" Whatever, Codex is exceptionally amazing for Python 🐍 and Rust! ...

11. brunoborges (Group Score: 96.8 | Individual: 21.8)

Cluster: 5 tweets | Engagement: 4 (Avg: 7) | Type: Tech

I hope schools are teaching kids to just sit down with TVs and watch stuff.\n\nQT @jaltma: I hope schools are teaching kids to just sit down with codex / claude code and make stuff.

See 4 related tweets

@brunoborges: I hope schools are teaching kids to just sit down with forks and knives and eat stuff.\n\nQT @jaltma...
@brunoborges: I hope schools are teaching kids to just sit down with jukeboxes and listen stuff.\n\nQT @jaltma: I ...
@brunoborges: I hope schools are teaching kids to just sit down with calculators and math stuff.\n\nQT @jaltma: I ...
@nmasc_: RT @jaltma: I hope schools are teaching kids to just sit down with codex / claude code and make stuf...

12. MarioNawfal (Group Score: 94.2 | Individual: 30.2)

Cluster: 4 tweets | Engagement: 1646 (Avg: 839) | Type: Tech

🚨🇨🇳 China launched its first fully automated humanoid robot production line with a capacity of 10k robots per year

Each robot takes 30 minutes to build and goes through 77 inspection stages

Robots building robots in half an hour. What could go wrong https://t.co/IE2yGn5Etp

See 3 related tweets

@kimmonismus: China announces its first automated manufacturing line capable of producing 10.000 humanoid robots p...
@WesRoth: China launched its first fully automated production line dedicated to manufacturing humanoid robots ...
@BrianRoemmele: The first known fully automated humanoid‑robot production line. Annual capacity of >10,000 robots...

13. adamlyttleapps (Group Score: 78.9 | Individual: 44.5)

Cluster: 2 tweets | Engagement: 1448 (Avg: 216) | Type: Tech

Designing App Store screenshots takes too long

So I made a Claude Skill that does it for me

Instead of spending hours in Figma… now it's done in 15 minutes on autopilot

Just add your app screenshots and it does the rest

The final product can even be copied straight into App Store Connect

It's the ASO App Store Screenshots skill and it's available on github here:

https://t.co/jfsgfxmQ68\n\nQT @adamlyttleapps: I'm creating Claude Skill that does it all in 15 mins... https://t.co/1DyDNykhBW

See 1 related tweets

@adamlyttleapps: Lesson: Every little project release needs a 15-30 sec launch video

The format is:

Pain point The...

14. aakashgupta (Group Score: 76.2 | Individual: 32.3)

Cluster: 3 tweets | Engagement: 126 (Avg: 246) | Type: Tech

In 2019, OpenAI was a nonprofit valued at $1 billion with fewer than 100 employees. GPT-2 existed and they were afraid to release it because it could write a convincing paragraph.

Seven years later, OpenAI is valued at $730 billion. NVIDIA went from a gaming GPU company to the most valuable company on Earth at$ 4.37 trillion. The total market cap of AI companies crossed $41 trillion. Anthropic, which didn't exist in 2019, just raised at$ 170 billion.

In 2019, "AI" was a computer science elective. In 2026, it's a $300 billion annual infrastructure buildout. Remote work was a negotiation line item in your offer letter. TikTok had just launched in the US. The entire internet ran on a set of assumptions that have been systematically dismantled in 84 months.

2019 you would open this app and think it was science fiction.\n\nQT @KristinRaworth: I've never sen anything more accurate https://t.co/a1E0KAUpxy

See 2 related tweets

@cryptopunk7213: openai's biggest investor is now launching products powered by Claude and competing AI labs. microso...
@Reuters: 🔊 ChatGPT's maker, Open AI, has 900 million users and an $840 billion valuation, yet its financial f...

15. rohanpaul_ai (Group Score: 75.4 | Individual: 49.7)

Cluster: 2 tweets | Engagement: 384 (Avg: 76) | Type: Tech

Marc Andreessen explains how AI has triggered a three-way standoff among Product-Managerss, engineers, and designers.

Each thinks AI lets them do the other two jobs, and the irony is that all may soon find AI can manage them too.

https://t.co/G7HIJ8QVXi\n\nQT @rohanpaul_ai: "you are the product manager, the agents are your engineers, and your job is to keep all of them running at all times"

The mental model of “AI helps me code faster” may already be obselete.

The more disruptive model is that the engineer starts to resemble a product manager for a temporary workforce of fast, literal, tireless contributors who are also error-prone and need constant steering.

This does not make engineering easier.

It makes good engineering look less like solitary craftsmanship and more like running a noisy distributed system with humans still responsible for architecture, taste, and truth.

See 1 related tweets

@rohanpaul_ai: RT @rohanpaul_ai: "you are the product manager, the agents are your engineers, and your job is to ke...

16. blackboxai (Group Score: 74.4 | Individual: 30.0)

Cluster: 3 tweets | Engagement: 73 (Avg: 137) | Type: Tech

100% no need to pick only one

we support claude code and codex cli to collaborate on each task or to compete and have a judge layer pick the best execution

try the Blackbox CLI here https://t.co/ocvvUgn9mQ

here is a brief demo https://t.co/A2Iht1CznR\n\nQT @rezoundous: After using Codex and Claude separately for a few weeks, I find that it's not Codex vs Claude.

It's Codex AND Claude.

See 2 related tweets

@RobRizk1: RT @blackboxai: 100% no need to pick only one

we support claude code and codex cli to collaborate o...

@RobRizk1: RT @blackboxai: @imsroch @steipete no need to pick only one

we support claude code and codex cli to...

17. omarsar0 (Group Score: 73.6 | Individual: 39.3)

Cluster: 2 tweets | Engagement: 361 (Avg: 235) | Type: Tech

NEW research from CMU.

(bookmark this one)

The biggest unlock in coding agents is understanding strategies for how to run them asynchronously.

Simply giving a single agent more iterations helps, but does not scale well.

And multi-agent research shows that coordination > compute.

A new paper from CMU proves this with a practical multi-agent system.

CAID (Centralized Asynchronous Isolated Delegation) borrows proven human SWE practices: a manager builds a dependency graph, delegates tasks to engineer agents who work in isolated git worktrees, execute concurrently, self-verify with tests, and integrate via git merge.

CAID improves accuracy over single-agent baselines by 26.7% absolute on paper reproduction tasks (PaperBench) and 14.3% on the Python library development tasks (Commit0).

The key insight is that isolation plus explicit integration beats both single-agent scaling and naive multi-agent approaches.

For long-horizon software engineering tasks, multi-agent coordination using git-native primitives should be the default strategy, not a fallback.

Paper: https://t.co/cRAbG7SrR5

Learn to build effective AI agents in our academy: https://t.co/1e8RZKs4uX

See 1 related tweets

@dair_ai: Effective strategies for asynchronous software engineering agents.\n\nQT @omarsar0: NEW research fro...

18. alexocheema (Group Score: 70.1 | Individual: 44.3)

Cluster: 2 tweets | Engagement: 6420 (Avg: 319) | Type: Tech

Big moment coming for local AI.

@nvidia x @apple https://t.co/MuHV6skiWQ

See 1 related tweets

@exolabs: Something small is coming. https://t.co/B84zh7RzdL\n\nQT @alexocheema: Big moment coming for local A...

19. Lovable (Group Score: 67.0 | Individual: 34.0)

Cluster: 2 tweets | Engagement: 69 (Avg: 169) | Type: Tech

We're hosting a Design Night in Stockholm on April 15th, bringing together design leaders to discuss the future of design.

Speakers include Mehmet Aydın Baytaş (Attio), Marina Kondratenko (ElevenLabs), Jan Six (GitHub, Token Studio), and Dan Hollick (Author of "Making Software"). They'll cover topics from dark mode maximalism to building brand at speed and design systems in the age of AI.

The event is for brand designers, product designers, and design engineers. There will be talks, Q&A, and time for mingling. Registration is required and subject to approval.

See 1 related tweets

@felixhhaas: We’re putting a lot of strong opinions in one room.

Stockholm, April 15 - Register now\n\nQT @Lovab...

20. danshipper (Group Score: 66.7 | Individual: 41.4)

Cluster: 2 tweets | Engagement: 312 (Avg: 49) | Type: Tech

RT @ivanhzhao: The loudest story about AI is a lonely one. One person with an army of chatbots. Other humans are friction. That gets the future wrong. The best things aren’t built alone. In a moment of change, we want to remind the world (and ourselves) what Notion stands for: — Think Together

See 1 related tweets

@ankrgyl: this warms my heart\n\nQT @ivanhzhao: The loudest story about AI is a lonely one. One person with an...