科技推文精选 - 2026年4月21日

2026年4月21日科技每日简报

Today's top tech conversations are led by @scaling01, whose post about 'RT @Kimi_Moonshot: Meet Kimi K...' garnered the highest engagement. Key themes trending across the top stories include https, vercel, security, apple, agent. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. scaling01 (Group Score: 908.5 | Individual: 60.7)

Cluster: 26 tweets | Engagement: 2860 (Avg: 201) | Type: Tech

RT @Kimi_Moonshot: Meet Kimi K2.6: Advancing Open-Source Coding

🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench Multilingual (76.7), BrowseComp (83.2), Toolathlon (50.0), Charxiv w/ python(86.7), Math Vision w/ python (93.2)

What's new: 🔹Long-horizon coding - 4,000+ tool calls, over 12 hours of continuous execution, with generalization across languages (Rust, Go, Python) and tasks (frontend, devops, perf optimization). 🔹Motion-rich frontend - Videos in hero sections, WebGL shaders, GSAP + Framer Motion, Three.js 3D. 🔹Agent Swarms, elevated - 300 parallel sub-agents × 4,000 steps per run (up from K2.5's 100 / 1,500). One prompt, 100+ files. 🔹Proactive Agents - K2.6 model powers OpenClaw, Hermes Agent, etc for 24/7 autonomous ops.
🔹Claw Groups (research preview) - bring your own agents, command your friends', bots & humans in the loop.

K2.6 is now live on https://t.co/YutVbwktG0 in chat mode and agent mode. For production-grade coding, pair K2.6 with Kimi Code: https://t.co/uvoSJKyGCY

🔗 API: https://t.co/EOZkbOwCN4 🔗 Tech blog: https://t.co/9wWvgIQSS3 🔗 Weights & code: https://t.co/Be0hjs2RTP

See 25 related tweets

@cryptopunk7213: dear god lol - new kimi model is a fucking beast.

GPT 5.4 level coding, 76% cheaper than opus 4.7 ...

@NousResearch: Kimi K2.6 is now available in Hermes Agent.

Simply run hermes update and use hermes model to se...

@huggingface: new open coding sota 🤗\n\nQT @Kimi_Moonshot: Meet Kimi K2.6: Advancing Open-Source Coding

🔹Open-sou...

@TheAhmadOsman: Been running Kimi K2.6 locally all morning

I can say with full confidence that this is the current ...

@Rasmic: I want the opensource models to win so bad...

let's go kimi !!!\n\nQT @Kimi_Moonshot: Meet Kimi K2....

2. Techmeme (Group Score: 220.6 | Individual: 35.8)

Cluster: 9 tweets | Engagement: 307 (Avg: 9) | Type: Tech

Sources: Google has created strike team to improve its coding models; Sergey Brin told DeepMind staffers that they must aggressively pivot to catch up on agents (@erinkwoo / The Information)

(Visit Techmeme dot com for the link and full context!)

See 8 related tweets

@ns123abc: BREAKING: Google DeepMind has assembled a strike team because Anthropic is mogging them on coding

...

@deredleritt3r: "The end goal is AI takeoff, or AI that can improve itself."

Google is back in the game.\n\nQT @Yuc...

@Yuchenj_UW: Google DeepMind formed a strike team to improve its coding models, with Sergey Brin directly involve...
@cryptopunk7213: google has assembled an emergency 'strike team' to improve their coding AI models after falling behi...
@theinformation: Sergey Brin is personally overseeing a new Google strike team to accelerate the path toward AGI:

@e...

3. yacineMTB (Group Score: 158.3 | Individual: 46.0)

Cluster: 5 tweets | Engagement: 1917 (Avg: 565) | Type: Tech

the true limit to scaling labour\n\nQT @Meta_Engineers: Today we're announcing LevelUp: a free, four-week training program that takes people with no prior experience and prepares them to work as fiber technicians on data center construction sites across the US.

We built this program with CBRE because the fiber technician field, and the broader construction industry, is facing a nationwide shortage at a time when data center demand is higher than ever.

How it works: 🔧 Classroom instruction, hands-on labs + team activities covering transferable technical skills

🎓 Graduates have the opportunity to work at Meta's US construction sites through our contractor network

🤝 Open to everyone from recent high school grads to mid-career professionals

Since 2010, Meta's data center projects have supported 30,000+ skilled trade jobs during construction + 5,000+ permanent operational roles. LevelUp is about building the pipeline to keep that going.

Learn more: https://t.co/9XluD5IHbz

See 4 related tweets

@aakashgupta: Meta is now running a free trade school because the bottleneck on AI infrastructure stopped being ch...
@_xjdr: i love to see this and we need a lot more of it.\n\nQT @Meta_Engineers: Today we're announcing Level...
@inafried: RT @Meta_Engineers: Today we're announcing LevelUp: a free, four-week training program that takes pe...
@StockSavvyShay: $META just partnered with CBRE on a new multiyear “LevelUp” program to recruit and train thousands o...

4. rohanpaul_ai (Group Score: 139.8 | Individual: 30.7)

Cluster: 7 tweets | Engagement: 12 (Avg: 52) | Type: Tech

OpenAI just gave Codex a screen-grounded memory system that can remember what you were working on without you restating it.

Before this, Codex mostly knew the current thread, so prompts like “fix this” or “sync that draft” often broke unless you rebuilt the missing context yourself.

Chronicle watches recent screen context, turns selected frames, OCR text, timing, and file paths into local memories, and helps Codex map your prompt to the right file, doc, tool, or recurring workflow instead of guessing.

It is a research preview for Pro users on macOS, needs Screen Recording and Accessibility permissions, uses rate limits fast, raises prompt injection risk, keeps screenshot files temporarily on device for up to 6 hours, and stores generated memories as unencrypted markdown on device, with no launch in the EU, UK, or Switzerland.\n\nQT @OpenAIDevs: Last week, we released a preview of memories in Codex.

Today, we’re expanding the experiment with Chronicle, which improves memories using recent screen context.

Now, Codex can help with what you’ve been working on without you restating context. https://t.co/b3p8I5eXOy

See 6 related tweets

@TheRealAdamG: https://t.co/fRdxzc4ZSZ

The Codex team continues to cook:

"Chronicle augments Codex memories with...

@testingcatalog: OpenAI introduced the Chronicle feature for Codex, which runs as a background agent and builds memor...
@testingcatalog: Chronicle feature will become available in Personalization settings on Codex and will require screen...
@Cointelegraph: ⚡️ LATEST: OpenAI expanded its Codex memory preview with Chronicle, allowing the system to use recen...
@TFTC21: OpenAI just launched Chronicle, an expansion of Codex memories that uses background agents and scree...

5. matteocollina (Group Score: 136.3 | Individual: 46.3)

Cluster: 4 tweets | Engagement: 1996 (Avg: 272) | Type: Tech

RT @rauchg: Here's my update to the broader community about the ongoing incident investigation. I want to give you the rundown of the situation directly.

A Vercel employee got compromised via the breach of an AI platform customer called https://t.co/7PY6gGtzgI that he was using. The details are being fully investigated.

Through a series of maneuvers that escalated from our colleague’s compromised Vercel Google Workspace account, the attacker got further access to Vercel environments.

Vercel stores all customer environment variables fully encrypted at rest. We have numerous defense-in-depth mechanisms to protect core systems and customer data. We do have a capability however to designate environment variables as “non-sensitive”. Unfortunately, the attacker got further access through their enumeration.

We believe the attacking group to be highly sophisticated and, I strongly suspect, significantly accelerated by AI. They moved with surprising velocity and in-depth understanding of Vercel.

At the moment, we believe the number of customers with security impact to be quite limited. We’ve reached out with utmost priority to the ones we have concerns about. All of our focus right now is on investigation, communication to customers, enhancement of security measures, and sanitization of our environments. We’ve deployed extensive protection measures and monitoring. We’ve analyzed our supply chain, ensuring Next.js, Turbopack, and our many open source projects remain safe for our community.

The recommendation for all Vercel customers is to follow the Security Bulletin closely (https://t.co/BLVnic9fJC). My advice to everyone is to follow the best practices of security response: secret rotation, monitoring access to your Vercel environments and linked services, and ensuring the proper use of the sensitive env variables feature.

In response to this, and to aid in the improvement of all of our customers’ security postures, we’ve already rolled out new capabilities in the dashboard, including an overview page of environment variables, and a better user interface for sensitive env var creation and management. As always, I’m totally open to your feedback.

We’re working with elite cybersecurity firms, industry peers, and law enforcement. We’ve reached out to Context to assist in understanding the full scale of the incident, in an effort to protect other organizations and the broader internet. I also want to thank the Google Mandiant team for their active engagement and assistance.

It’s my mission to turn this attack into the most formidable security response imaginable. It’s always been a top priority for me. Vercel employs some of the most dedicated security researchers and security-minded engineers in the world. I commit to keeping you updated and rolling out extensive improvements and defenses so you, our customers and community, can have the peace of mind that Vercel always has your back.

See 3 related tweets

@itsumeshk: when we launched Runable 2.0, Chinese hackers spun up hundreds of fake accounts and tried to DDoS us...
@elvissun: vercel is still safer than running "claude --dangerously-skip-permissions" on your VPS\n\nQT @rauchg...
@ibuildthecloud: Security is getting tough.\n\nQT @rauchg: Here's my update to the broader community about the ongoin...

6. Vtrivedy10 (Group Score: 126.5 | Individual: 32.1)

Cluster: 5 tweets | Engagement: 52 (Avg: 43) | Type: Tech

we want to help you navigate all the stuff that comes with shipping an agent to production - so we wrote a guide!

this guide packages learnings across our + customer deployments, building open source agent infra/frameworks/examples, debugging infra errors, and building great agents that help you improve over time\n\nQT @LangChain: Things don’t always go to plan when bringing agents into production. ‘deepagents deploy’ is purpose-built around the challenges teams face when deploying.

✅ Every infrastructural consideration is mapped to a purpose-built runtime capability. No need to rebuild components from scratch.

✅ Open models, open harness, open memory. Your team is in control.

Here’s our conceptual guide by @sydneyrunkle + @Vtrivedy10 on the biggest issues faced in production, the runtime solutions we built around them, and how you can go from a toml file to a LangSmith Deployment in minutes.

https://t.co/wsOCSvc3VX

See 4 related tweets

@LangChain: Things don’t always go to plan when bringing agents into production. ‘deepagents deploy’ is purpose-...
@sydneyrunkle: developing an agent is a harness problem. deploying an agent is a runtime problem.

here's a guide t...

@hwchase17: RT @Vtrivedy10: we want to help you navigate all the stuff that comes with shipping an agent to prod...
@hwchase17: RT @sydneyrunkle: developing an agent is a harness problem. deploying an agent is a runtime problem....

7. MatthewBerman (Group Score: 118.5 | Individual: 31.8)

Cluster: 6 tweets | Engagement: 140 (Avg: 186) | Type: Tech

Tim Cook is a scaling god, he’s not a new idea guy.

No one else could have built up Apple into the juggernaut it is today.

Someone else needs to bring Apple into the AI era.\n\nQT @markgurman: BREAKING: Tim Cook steps down. Ternus to CEO.

See 5 related tweets

@MyLordBebo: 📱 TIM COOK IS OUT!

John Ternus, Apple’s senior vice president of hardware engineering, will take ov...

@Reuters: JUST IN: Apple named ‌insider John Ternus as its next CEO, tasking the long-time hardware chief to s...
@alexocheema: Ternus as CEO, Srouji as Chief Hardware Officer.

The people who turned Apple Silicon into the leadi...

@Cointelegraph: 🔥 NEW: Apple is reportedly replacing CEO Tim Cook with Senior VP of Hardware Engineering John Ternus...
@testingcatalog: Apple has a new CEO 👀

Tim Cook is stepping down from his role. Wondering if AI strategy will chang...

8. gdgtify (Group Score: 115.0 | Individual: 52.3)

Cluster: 3 tweets | Engagement: 410 (Avg: 25) | Type: Tech

RT @Alibaba_Qwen: 🚀 Introducing Qwen3.6-Max-Preview, an early preview of our next flagship model

Highlights： ⚡️ Improved agentic coding capability over Qwen3.6-Plus 📖 Stronger world knowledge and instruction following 🌍 Improved real-world agent and knowledge reliability performance

Smarter, sharper, still evolving. More Qwen3.6 models to come. Stay tuned!

🔗👇 Blog: https://t.co/6hDQJhmkjM Qwen Studio: https://t.co/Fe2X1IrW6r API: https://t.co/xWPs39LBIm

See 2 related tweets

@TeksEdge: 🆕 Qwen3.6-Max-Preview is here! Alibaba’s early preview of their next flagship model is here for test...
@kimmonismus: Qwen-3.6 Max preview shows some impressive results. However, i prefer their benchmarks against opus ...

9. tbpn (Group Score: 113.3 | Individual: 34.8)

Cluster: 4 tweets | Engagement: 568 (Avg: 161) | Type: Tech

Three months ago on TBPN, @markgurman explained why John Ternus is the "only guy" that could become the next CEO of Apple:

"All signs are turning towards Ternus."

"Ternus is 50. Everyone else on the Apple executive team is late 50s through their mid 60s. Turning 66 this year, in the case of Tim Cook. You're Apple's board, you like continuity, you like an insider, you like people who know what they're doing, and who have been there for a while, they know where the bodies are buried."

"At 50, he's the only one. Let's say Tim Cook hangs out for another three to five years. You're not going to appoint another CEO who's 65 or 70 years old. He's the only guy."

"Apple gets the vast majority of its revenue from hardware. He's the hardware guy. Have they screwed up any hardware since he's been in charge? No. He's a steady hand who knows what he's doing. He's really the only choice."

"A few months ago, he took full control of the Apple Watch engineering team that was co-run with the old COO until he retired."

"When they started soft-firing the Head of AI [John Giannandrea], they took the robotic stuff away from him and gave that to Ternus. And then the real news that I broke last week is, even though on paper Tim Cook is running the Apple design teams, he's not. It's Ternus. He took that over at the end of last year."

See 3 related tweets

@aakashgupta: The succession everyone called for years just happened, and it's the most revealing personnel decisi...
@nytimes: Breaking News: Tim Cook is stepping down as Apple’s chief executive and will be replaced by John Ter...
@SawyerMerritt: Tim Cook: “It has been the greatest privilege of my life to be the CEO of Apple and to have been tru...

10. amasad (Group Score: 112.0 | Individual: 37.8)

Cluster: 4 tweets | Engagement: 362 (Avg: 125) | Type: Tech

Go to an existing Replit project
“Make a mobile app”
Publish to App Store\n\nQT @zhenthebuilder: Now you can add a mobile app to the website you’ve been building on Replit.

Ship it to the App Store from your existing project.

Just open your app and ask the Agent. https://t.co/a3Y67uM2Ve

See 3 related tweets

@Replit: You can now add a mobile app, slide deck, or animation to any web app you have built prior to Replit...
@amasad: Creating a mobile app version of a live web app with users and data in 30 seconds (sped up). https:/...
@amasad: RT @amasad: 1. Go to an existing Replit project

“Make a mobile app”
Publish to App Store...

11. startupideaspod (Group Score: 109.6 | Individual: 32.2)

Cluster: 4 tweets | Engagement: 8 (Avg: 79) | Type: Tech

Most people obsess over their agent's setup.

customizing
tweaking
optimizing

when the point is what it gets done.

"Customizing is not the skill."

The Key:

Agent handles the background work
Get 20-30% more done
More reps → better signal → better returns

The means: setup.

The ends: output.

Don't be an artist about it\n\nQT @gregisenberg: how to set up hermes agent step by step. built-in memory, 40+ tools, works on your phone, and what to think of hermes vs openclaw:

hermes is a personal AI agent that runs in your terminal. think of it like open claw but with built-in memory, 40+ tools out of the box, and 90% cheaper token costs. you install it with one command.
the 3 problems with open claw that hermes solves: no memory (you keep repeating yourself), constant gateway restarts, and zero visibility into what you're spending on tokens.
hermes remembers everything. every completed task gets saved to memory. it searches through past logs to find solutions. over time it literally gets smarter at your specific workflows.
connect it to open router. you see exact costs per model per task. free models rotate weekly. one founder went from $130 every five days on open claw to$ 10 on hermes. same output.
it comes preloaded with skills. apple notes, imessage, find my, browser, web search, image generation, cron jobs. no hunting for plugins.
connect it to obsidian so it reads your entire vault. connect it to gstack for your dev environment. create custom skills for your specific workflows.
the biggest money saver: have it write code once for recurring tasks. then it runs without burning tokens every time. stop paying an LLM to do the same scrape or report daily.
run it on android via telegram. name your agents. talk to them like coworkers. in this episode imran shows you how to set this up.
you can run it bare metal, in docker, or serverless on modal. pick your risk level.

i begged @imranye to come on @startupideaspod and walk through the full installation live. he made it impossibly clear.

if you've heard of Hermes Agent and want the clearest explanation of how to get set up like a pro

let me know what you want me to cover on the next ep

this is the best personal agent setup video on the internet right now.

watch

See 3 related tweets

@startupideaspod: RT @gregisenberg: how to set up hermes agent step by step. built-in memory, 40+ tools, works on your...
@aakashgupta: The chart is the story. Flat line for seven months. Then vertical.

Hermes Agent is Nous Research's ...

@RoundtableSpace: HERMES AGENT IS NOW GAINING MORE INTEREST THAN OPENCLAW

> open source > self-improving > ...

12. MarioNawfal (Group Score: 109.5 | Individual: 62.2)

Cluster: 3 tweets | Engagement: 6661 (Avg: 587) | Type: Tech

Wait? What did I just read? Grok 4.3 just broke the game?

@xAI dropped the nuke: @Grok can now spit out full Microsoft Word docs, PDFs, PowerPoints, and Excel spreadsheets on command.

No more "sorry I can’t generate files" excuses, this AI is actually useful AF

@elonmusk https://t.co/vdpiHL44Nc\n\nQT @MarioNawfal: Forget about any other AI, the signs are clear: Grok 4.3 Beta is already #1

Creates clean slides, graphs, images & dashboards
Pulls real-time web data + images
Runs smooth multi-step workflows
Analyzes videos like a beast

From idea → full presentation or dashboard in one chat.

Grok is finally becoming that everyday superpower.

And the Grok Computer drops soon

This is the one.

Source: @grok, @elonmusk

See 2 related tweets

@MarioNawfal: RT @MarioNawfal: Wait? What did I just read? Grok 4.3 just broke the game?

@xAI dropped the nuke: @...

@XFreeze: Grok Imagine's improvement rate is insanely fast.... it's now blurring the line between reality and ...

13. chrysb (Group Score: 107.6 | Individual: 38.3)

Cluster: 3 tweets | Engagement: 56 (Avg: 60) | Type: Tech

word on the street is a16z is looking to invest in legit teams as early as possible. pre-traction or even pre-product

if that's you, def apply. @andrewchen is goated!\n\nQT @andrewchen: ok - dropping big dates/news for a16z speedrun:

starting TODAY, founders can apply for the 2026 program that runs July 27 to Oct 11 in SF

here's the link: https://t.co/opkX4cIaCF

we will be investing up to $1M and funding 70+ companies over the next few weeks
But there's also $5M in credits/tokens/etc from AWS/GCP/Open AI/Azure/NVIDIA/Deel/Stripe/etc. You'll also work with our amazing operating team (GTM, talent, brand, people, and more), and join our community of elite founders
we offer a Global Founders Program for international founders, to help with visas, banking support, relo recommendations
yes you can be solo (but better if you're further along, and have built a team). No you don't have to have an idea yet. Yes you have to know how to build (even if you're not technical)
Also, in other news: speedun is officially moving full-time to SF. (prev it alternated SF/LA) this is for all the obv reasons
we've continued to have an insane lineup of speakers, including the founders of Carta / DoorDash / Twilio / Figma / Zynga / Airtable / Twitch / and of course, lunch/dinners with Marc/Ben alongside a16z team - and much more
the deadline for applying is May 17!

See 2 related tweets

@pmarca: It's time to build!\n\nQT @andrewchen: ok - dropping big dates/news for a16z speedrun:
starting T...
@a16z: RT @andrewchen: ok - dropping big dates/news for a16z speedrun:
starting TODAY, founders can appl...

14. jasonlk (Group Score: 103.9 | Individual: 35.1)

Cluster: 3 tweets | Engagement: 5 (Avg: 38) | Type: Tech

50% of SaaStr AI 2026 (May 12-14) attendees are Founder, CEO, or C-Suite.

Who's showing up:

↳ Founder: 16.8% ↳ CEO/President: 16.7% ↳ C-Suite: 16.0% ↳ VP+: 63.1% of the room ↳ Director+: 72.8%

This is leadership coming to to get ahead, and stay ahead, in AI + B2B.

By function, four things stand out:

1/ Engineering/Tech is 12% of the room. Five years ago that was 3-4%. AI pulled the builders into the conversation with the buyers.

2/ Marketing (10%) is slightly ahead of Sales (9.6%). AI is reshaping marketing workflows faster than sales workflows right now.

3/ Customer Success at 1.5% is the surprise. Given how much CS is being rewired by AI agents, this number should be 3-4x higher. Are CS leaders ... staying home during disruption?

4/ Investors will land at ~500.

SaaStr AI 2026 is where B2B leaders learn to keep ... winning in the Age of AI.

That's what May 12-14 is actually about.

See you then in SF Bay!\n\nQT @jasonlk: It's time to learn to Build it. Ship it. Vibe it. Get it into production. For real. We'll make you an agentic expert.

Together with @Replit at 2026 https://t.co/PBv85S5Cyt May 12-14 we'll teach you:

-How to Build Your Own AI VP Marketing

How to Build Your Own AI VP Customer Success
How to Ship AI-Powered Sales & Marketing Tools in 30 Min
How to Turn a Mockup into a Working Prototype
How to Go From Prompt to Product in 30 Min
How to Build Your Own AI-Powered MVP

No code required. Just bring your laptop.

We'll give you the prompt.

https://t.co/PBv85S5Cyt 2026. May 12-14 in SF Bay!!

See 2 related tweets

@jasonlk: Stop feel behind on AI. We'll get you ahead. In just 48 hours

SaaStr AI Annual 2026 May 12-14 in SF...

@jasonlk: Build It. Ship. Vibe It. Release It.

SaaStr AI Annual 2026. May 12-14 in SF Bay https://t.co/QC...

15. gdgtify (Group Score: 102.7 | Individual: 18.6)

Cluster: 7 tweets | Engagement: 98 (Avg: 25) | Type: Tech

RT @OfficialLoganK: Excited to share that Google AI subscriptions (Pro and Ultra) now work with @GoogleAIStudio!!

Come vibe code and use the playground with higher rate limits. Available right now! https://t.co/Rf1KlA0iwn

See 6 related tweets

@AndrewCurran_: I'm very happy about this, long rumored and finally arrived. I tried it and It's already synced with...
@scaling01: bullish\n\nQT @OfficialLoganK: Excited to share that Google AI subscriptions (Pro and Ultra) now wor...
@seconds_0: LETS GO\n\nQT @OfficialLoganK: Excited to share that Google AI subscriptions (Pro and Ultra) now wor...
@GoogleAIStudio: start vibe coding with your Google AI plan

AI Pro and Ultra subscribers now get higher usage limits...

@WesRoth: RT @WesRoth: Google is in the middle of a staged rollout integrating Google AI subscriptions (Pro an...

16. rseroter (Group Score: 100.9 | Individual: 36.4)

Cluster: 3 tweets | Engagement: 152 (Avg: 48) | Type: Tech

RT @levie: What gets missed with AI productivity gains is that by and large, most roles will continue to be as sophisticated as the tools allow.

This is why also thinking through “today’s jobs will be replaced with AI” is a fallacy. Everyone thinks the market is static, but it’s not.

As a result of everyone having access to the same technology which augments our work, then users of the tools will increasingly raise their level of output to the point where the prior definition of the job is no longer relevant. Thus, those that understand their particular field and grow in their skills will continue to be differentiated vs. others.

If you can do far more, then you start to tackle bigger and harder problems. If you do that, then the expertise still is required to get the job done fully.

The engineer with AI is going to be far more productive and capable with AI than the non-engineer trying to build the same piece of software. Building a lightweight app is no longer the definition of getting by in software development. Reviewing a contract will no longer be the definition of a paralegal. Splicing a video won’t be the definition of a video editor. Providing basic financial research won’t be the job of the financial analyst in the future.

Simply put, AI will naturally cause most roles to actually grow in complexity rather than reduce in complexity, because we can do far more with the tools.

See 2 related tweets

@levie: What gets missed with AI productivity gains is that by and large, most roles will continue to be as ...
@altryne: 1) Aaron is absolutely right

this is why most AI Engineers can barely sleep and have FOMAT
don...

17. chddaniel (Group Score: 92.7 | Individual: 29.9)

Cluster: 4 tweets | Engagement: 28 (Avg: 40) | Type: Tech

Today we're introducing Vibe Coding 2.0

I just watched my Mac build a complete business in 182 seconds...

This is just unbelieving. https://t.co/MqsyeA6XXn

See 3 related tweets

@chddaniel: so you're telling me Claude Code w/ Opus 4.7 can now...
build an app
launch it
do email marke...
@chhddavid: this is like actually terrifying...\n\nQT @shipper_now: BREAKING: Today, we pulled the plug on vibe ...
@Shipper_now: Join the revolution ⛴️\n\nQT @shipper_now: BREAKING: Today, we pulled the plug on vibe coding.

I ju...

18. zerohedge (Group Score: 91.7 | Individual: 31.5)

Cluster: 4 tweets | Engagement: 1137 (Avg: 935) | Type: Tech

*ADOBE TO TEAM UP WITH MICROSOFT, ANTHROPIC, OPENAI, NVIDIA: WSJ

*ADOBE AI AGENT TO HELP BOOST CUSTOMER ENGAGEMENT, SALES: WSJ

better to make an agent that shorts ADBE stock

See 3 related tweets

@rohanpaul_ai: WSJ: Adobe just launched CX Enterprise, a new AI agent platform to stop its business software from b...
@wallstengine: WSJ: Adobe $ADBE is launching CX Enterprise, an AI agent platform for businesses, and partnering wit...
@negligible_cap: *ADOBE UNVEILS AGENTS FOR BUSINESSES CALLED CX ENTERPRISE: WSJ

$ADBE is releasing their AI Agent pl...

19. shiri_shh (Group Score: 88.2 | Individual: 27.4)

Cluster: 5 tweets | Engagement: 300 (Avg: 514) | Type: Tech

Vercel got hit yesterday… and Lovable is the NEXT one on fire RIGHT NOW.

Any free user can read your full codebase, prod creds, AI chat histories, and live customer records if you built before Nov 2025.

precisely why Anthropic is holding Claude Mythos back from the public.

Their new model is scary good at hacking and finding zero-days.\n\nQT @weezerOSINT: Lovable has a mass data breach affecting every project created before november 2025.

I made a lovable account today and was able to access another users source code, database credentials, AI chat histories, and customer data are all readable by any free account.

nvidia, microsoft, uber, and spotify employees all have accounts. the bug was reported 48 days ago. its not fixed. They marked it as duplicate and left it open.

See 4 related tweets

@shiri_shh: RT @shiri_shh: Vercel got hit yesterday… and Lovable is the NEXT one on fire RIGHT NOW.

Any free us...

@runable_hq: The vibe coding space needs to take security seriously

Seeing reports of a potential platform-level...

@Hesamation: How billion-dollar companies are handling the security of their users in the peak of AI-assisted cyb...
@inafried: RT @weezerOSINT: Lovable has a mass data breach affecting every project created before november 2025...

20. Mayhem4Markets (Group Score: 88.1 | Individual: 25.7)

Cluster: 4 tweets | Engagement: 14 (Avg: 61) | Type: Tech

Kimi K2.6 just dropped.

It's an open-source, native multimodal agentic AI model built for serious work:

🧠 Long-horizon coding across Rust, Go & Python 🎨 Prompts → production-ready UIs with polish 🤖 Competes with much larger frontier models

At 1.1T parameters, Kimi K2.6 is way too big for me (and most others) to run locally.

I would love to see a Flash version that's 150-250B parameters.

See 3 related tweets

@testingcatalog: Moonshot AI is rolling out Kimi K2.6 on Kimi Chat and APIs. All models got upgraded.
Kimi K2.6 I...
@testingcatalog: Kimi K2.6 from @Kimi_Moonshot is a new open-source SOTA on HLE with tools, SWE Bench Pro, and other ...
@pingToven: RT @OpenRouter: Kimi K2.6 is live on OpenRouter!

@Kimi_Moonshot's new model is a long-horizon codin...