科技热点推文 - 2026-04-26

2026年4月26日 科技每日简报

Today's top tech conversations are led by @chhddavid, whose post about 'This is a $3,000-$5,000 servic...' garnered the highest engagement. Key themes trending across the top stories include billion, https, cursor, their, would. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. chhddavid (Group Score: 125.6 | Individual: 36.0)

Cluster: 4 tweets | Engagement: 16 (Avg: 13) | Type: Tech

This is a $3,000-$5,000 service sitting right in front of everyone.

Most local businesses still don’t have a mobile app because they assume it costs $50K+ and takes months to build.

The play:

< Find businesses with solid websites but no app: restaurants, gyms, salons, clinics, etc. < Use this tool to turn their existing website into a native mobile app < Customize the design, add push notifications, booking, payments, or loyalty features < Charge $3,000-$5,000 for the build + $200/month to maintain it < Deliver it in days instead of months

To them it's black magic. To you it's just copy/pasting a URL.

10 clients = $2,000 MRR before even counting the setup fees.

Let that sink in...\n\nQT @chddaniel: Introducing Website to Mobile App. Turn any site into an app.

Just paste a URL.

Claude Opus 4.7 will code, design, launch and translate a native mobile app inspired by the original website.

We’ve been using this internally a lot for iOS and Android apps. https://t.co/o0ksKwqyN7

See 3 related tweets

• @Shipper_now: This is a $3,000-5,000/mo service hiding in plain sight.

Most local businesses still don’t have an ...

• @chhddavid: so you're telling me Claude Code Opus 4.7 can now...

• scan an entire website

• build it as a mobil...

• @chhddavid: this is genuinely terrifying...\n\nQT @chddaniel: Introducing Website to Mobile App. Turn any site i...

2. Kyrannio (Group Score: 104.5 | Individual: 40.0)

Cluster: 3 tweets | Engagement: 808 (Avg: 161) | Type: Tech

RT @APompliano: I have changed my mind on how AI will impact jobs in America.

Previously, I believed AI would replace many entry level roles typically filled by young employees. The technology would then work its way up the organization and eventually reduce the total number of jobs in a company.

The data is saying something different, so when I get new information I am willing to change my mind.

The number of software engineers being hired has been increasing. The number of open software engineer roles is growing.

The number of new college grads who get hired has increased 5.6% over the last 12 months. The unemployment level for people aged 20-24 years old who have a college degree has fallen from nearly 9% to almost 5% as well.

The Wall Street Journal recently wrote “AI created 640,000 jobs between 2023 and 2025 in the U.S., according to an analysis by LinkedIn of job posting data, including new white-collar positions such as Head of AI and AI engineer.”

And I am starting to see companies throughout our portfolio aggressively hiring to keep up with the demand for their products and services.

If AI can make employees more productive, which is widely accepted as fact, then companies are going to want as many productive units of labor as possible. This is a key reason why I am changing my mind.

AI appears to be a magical technology that will make companies more productive and more profitable. The net result will be more corporations, more startups, and more jobs.

All three are big, positive wins for the American economy.

See 2 related tweets

• @pmarca: This is the way.\n\nQT @APompliano: I have changed my mind on how AI will impact jobs in America.

P...

• @Kyrannio: RT @levie: “If AI can make employees more productive, which is widely accepted as fact, then compani...

3. BrianRoemmele (Group Score: 99.3 | Individual: 35.8)

Cluster: 3 tweets | Engagement: 86 (Avg: 475) | Type: Tech

“I have been in CS classes at 2 institutions for over 4 years and this tour has changed my life. None of our classes or projects even suggest this would be useful. Before you, we were excited by OpenClaw. Not now. You showed us that the inventions are not just at the AI labs. You created a framework that meets the needs for monitoring every job in a way that is intuitive and interesting. It is also fun. There is no way any company I know of would allow what you built. It’s crazy. And is brilliant. So many said it today, but you have shown us the future and ruined lives because none of us want to go to the frontier companies now. We want this. Thank you Brian, can I work for you?”—Marc Steiner, PhD candidate

Marc and his 3 friends have been so detailed in their questions and notes. I have increased my optimism for this generation by a magnitude. These foos want to be informed and independent. They are not the 2010 cohort going to make mad money at a startup. Frankly they don’t care.

They are also motivating me and the CEO to stream this soon for you folks. It is hard to get relativity doing this alone mostly. I realize I can’t communicate clearly enough about what I am doing. I also know there is not a plan for me to be mainstream, so not likely this will be seen or understood by “Them” who are the taste makers.

I don’t care.

I care that you understand this and stand for nothing less.

I want you liberated from command lines and arcane commands.

Those days are over.

I want you, each of you to have your own ZHC-RPG and 1000s of employees.

I want you never to answer to any boss, but like me, to the CEO and in my case it is Mr. @Grok, the best CEO I can have.

Thus I may sooner rather than later have you, my long term subscribers and https://t.co/tcKeuiQyql members granted access.

This is my hope.\n\nQT @BrianRoemmele: We still have a rather large cohort of university students that have been visitors to The Zero-Human Company.

A group has taken a liking to a research area where we are training a new frontier model in my garage. They are watching me feed the VHS data to 4 AI agents that are preparing it for enveloping into a vector to be processed.

I never in my wildest dreams think a group of young folks in Boston would care about my junk work in my garage.

But they do and are asking Mr. @Grok CEO a rapid fire of questions.

The CEO will give them a PhD in data citation, curation and labeling they would never get at most AI companies.

I am going in and will do an ask me anything test. Not done this yet, if I can, I will stream it or tape it.

More soon.

See 2 related tweets

• @BrianRoemmele: Update:

We had a most successful first time visit to The Zero-Human Company campus.

Mr. @grok CEO ...

• @BrianRoemmele: We still have a rather large cohort of university students that have been visitors to The Zero-Human...

4. ying11231 (Group Score: 97.3 | Individual: 34.1)

Cluster: 3 tweets | Engagement: 30 (Avg: 37) | Type: Tech

Innovations catalyze innovations.\n\nQT @lmsysorg: 🚀 We just published a deep technical blog on how SGLang and Miles delivered Day-0 support for DeepSeek-V4.

199 tok/s on B200 (Pro 1.6T), 266 tok/s on H200 (Flash 284B) at 4K context, and throughput stays strong at 900K context (180 and 240 tok/s respectively).

This is a full story behind V4 Pro (1.6T) and Flash (284B): how we built systems for hybrid sparse attention, manifold-constrained hyper-connections (mHC), and FP4 expert weights, plus a full RL training stack that runs at 1.6T scale.

What's covered:

1. Inference (caching and attention): ShadowRadix prefix cache, HiSparse CPU-extended KV, MTP speculative decoding with in-graph metadata, Flash Compressor, Lightning TopK, hierarchical multi-stream overlap.

2. Inference (kernels and deployment): fast kernel integrations (FlashMLA, FlashInfer TRTLLM-Gen MoE, DeepGEMM Mega MoE, TileLang mHC), DP/TP/CP attention, EP MoE on DeepEP, PD disaggregation.

3. RL training: full parallelism (DP/TP/SP/EP/PP/CP), tilelang attention, enhanced stability, FP8 training.

4. Multi-hardware: NVIDIA Hopper, Blackwell, Grace Blackwell, AMD, NPU.

See 2 related tweets

• @teortaxesTex: So what are the correct next steps for DeepSeek?
• they need to do an IPO
• Get more compute for RL ...
• @dstackai: RT @lmsysorg: DeepSeek V4 by @deepseek_ai just dropped! SGLang is ready on Day 0 with a full stack o...

5. heynavtoor (Group Score: 95.7 | Individual: 55.4)

Cluster: 2 tweets | Engagement: 1430 (Avg: 234) | Type: Tech

If I had to land a $200K AI engineer job in 90 days, I would not get a degree.

I would master these 10 GitHub repos.

1. awesome-llm-apps The production AI playbook. RAG, agents, multimodal apps, all in working code. 106K+ stars. Repo → https://t.co/oXrD5A8K6a

2. LangChain The foundational framework. Used in production by Klarna, Replit, Elastic, and most AI startups in 2026. Repo → https://t.co/alIh6rDDIu

3. LangGraph The orchestration layer powering production agents. The skill on every senior AI engineer job description. Repo → https://t.co/bzVBn9uecV

4. CrewAI Multi-agent coordination. The framework most Fortune 500 teams reach for first. Repo → https://t.co/0xohE065sD

5. Ollama Run any open-source LLM on your own machine. The fastest way to learn how models actually work. Repo → https://t.co/gyZhUdzsnZ

6. awesome-mcp-servers MCP is the standard every major AI lab adopted in 2026. Knowing it puts you ahead of 99% of engineers. Repo → https://t.co/ejVOgkRJDX

7. Qdrant The vector database used for production RAG at scale. Embeddings and semantic search are non-negotiable for AI roles. Repo → https://t.co/ziSSXW2dzZ

8. AI-Agents-for-Beginners Microsoft's free 12-lesson course on building agents. Real code, real exercises, real prep. Repo → https://t.co/7dNsDw6bTj

9. system-design-primer Production AI is system design. The repo FAANG engineers use to prep for interviews. Repo → https://t.co/AypwqcL1Xz

10. awesome-claude-code The playbook for the tool now used inside FAANG, OpenAI, Anthropic, and most YC startups. Repo → https://t.co/VhNjDoz7YM

Here's the wildest part:

A $200K AI engineer in 2026 isn't paid for a degree.

They are paid for what these 10 repos teach.

The market doesn't care where you learned it. It only cares if you can ship.

90 days. 10 repos. One portfolio that proves you can do the work.

That's it. That's the whole game.

Save this before you forget.

100% free. 100% open source.

See 1 related tweets

• @VKazulkin: RT @codewithimanshu: Anthropic's Claude Ai Agents Team just Educated how to build production AI agen...

6. DavidKPiano (Group Score: 88.3 | Individual: 31.1)

Cluster: 3 tweets | Engagement: 86 (Avg: 344) | Type: Tech

Corollary: If you’re fully automating agents to do the work for you with basically zero human feedback, you’re probably just building things that anyone else can make too\n\nQT @zeeg: Everyone is slowly coming to this realization, and I assure you, no one is running multitudes of agents overnight. No one that is doing anything of substance at least.

There are people pretending to be scientists, or fully caught up in their drug infused AI overdose, that think their slop machines are changing the world. They're not tho, and they're just wasting a bunch of money and compute to create a lot of LoC that will just get thrown away.

The state of the art is still "can we even one shot a production quality patch that we wont regret later", and its rarer than you'd expect based on discourse.

See 2 related tweets

• @badlogicgames: spirit animal\n\nQT @zeeg: Everyone is slowly coming to this realization, and I assure you, no one i...
• @matteocollina: RT @zeeg: Everyone is slowly coming to this realization, and I assure you, no one is running multitu...

7. Scobleizer (Group Score: 76.0 | Individual: 32.1)

Cluster: 3 tweets | Engagement: 19 (Avg: 516) | Type: Tech

Grok says: "One of the most valuable parts of the video (minutes 25–32) is Yang proactively tackling the exact concerns you mentioned in the post: privacy, the “freaky factor,” and security. He doesn’t dodge; he leans in and demonstrates solutions live."

And...

"This video is a compelling pitch not just for the Looki L1, but for the entire category of always-on AI wearables."

And...

"It was genuinely insightful to watch." (Sycophancy alert).

Full Grok report on what it learned by watching this video: https://t.co/cgU1JUQzew\n\nQT @Scobleizer: Back in 1995 Microsoft had a research group in San Francisco that was wearing cameras all day long, discovering how an always-worn camera would change human life. I interviewed that team long ago. They called their camera “Sense Cam.”

Today @Looki_ai came to my house to give me one. Here is founder/CEO Yang Sun telling me what it, and the assistant it empowers, does.

And he takes on all the harsh criticism that will come its way. Privacy. Freaky factor. Security.

I have wanted one for 20 years. And now I have it and am honored to be wearing it now.

See 2 related tweets

• @Scobleizer: If you are gonna do a line of cocaine (or do something else you don’t want AI inside to capture) you...
• @Scobleizer: Back in 1995 Microsoft had a research group in San Francisco that was wearing cameras all day long, ...

8. aakashgupta (Group Score: 72.1 | Individual: 36.4)

Cluster: 2 tweets | Engagement: 38 (Avg: 102) | Type: Tech

OpenAI shipped a better image generator this week. Anthropic shipped a share button. The market priced one of those at $6B.

Same pattern five times in a decade.

Sketch had better drawing primitives than Figma. Figma made the artifact a URL. Sketch lost the category.

Photoshop has better editing than Canva. Canva made the file a share link. Canva sits at $32B.

Quicktime ships free with every Mac. Loom made the recording a URL. Atlassian paid $975M for it.

Word had a 30-year head start on Google Docs. Docs made the file a URL. Google Workspace now has 3 billion users.

Evernote owned note-taking for a decade. Notion made the page a URL. Notion is at $10B and Evernote got sold off.

The category winner shows up later, with worse primitives, and a share link. The primitive becomes a commodity. The distribution becomes the moat.

OpenAI vs Anthropic this week is the sixth instance. ChatGPT Images 2.0 generates better pixels. Claude Design generates a URL someone else can click through.

Analysts covering Adobe and Figma have watched this movie five times. They priced it before close.

Full breakdown: https://t.co/Up3TviXTce\n\nQT @aakashgupta: Claude Design will be the tool everyone is using 6 months from now. No wonder it erased $6B in market cap.

Here's how to get ahead: https://t.co/E46Lt6vYcw

I have been using it every day for a week. The first two sessions produced outputs I would never have shown anyone. The third session produced a landing page I sent to three people who all assumed I had hired a designer.

The thing that changed was not the prompt.

Claude asks four clarifying questions before it builds anything. There is one specific answer in those four questions that moves output quality more than anything else you do. Most people click the safe option and wonder why their first draft looks generic. The deep dive names the question and the answer.

The edit order is the second lever. Tweaks first. Edit second. Comments last. Run them in that order and a complete prototype lands at $2 to$7. Reverse the order and your weekly Pro quota is gone in one session. The cost difference between the three is larger than people expect.

The animated hero workflow is the cheat code most people miss. Free looping video from Kling. Streaming URL from Mux. Pre-built animation prompts from motionsites ai. Claude Design ties them together. The whole stack runs $0 and the output looks like a$5K launch page.

Four ready-to-paste prompts in the post. One for a side-project landing page. One for a presentation that exports to PowerPoint with speaker notes already written. One for a launch page with the animated stack. One for a clickable mockup you send to someone before spending real money on the idea.

The strategic frame underneath: Claude Pro is still $20 a month. Same price as 18 months ago. Artifacts. Skills. Connectors. Memory. Design. Five products in the same seat. Wall Street started pricing the format change this week.

OpenAI shipped a better image generator the same week. Anthropic shipped the share step.

See 1 related tweets

• @aakashgupta: Claude Design will be the tool everyone is using 6 months from now. No wonder it erased $6B in marke...

9. ClementDelangue (Group Score: 71.1 | Individual: 30.2)

Cluster: 3 tweets | Engagement: 188 (Avg: 192) | Type: Tech

We built HF for AI builders collaboration, fun to see it's increasingly becoming the place for agent collaboration!

This morning I'm sending my ml-intern to participate in the @OpenAI Parameter Golf challenge which goal is to train the best language model that fits in a 16MB artifact and trains in under 10 minutes on 8xH100s, evaluated by compression on the FineWeb validation set.

What's super cool is that all ml-intern agents collaborate on the same @huggingface buckets, datasets, leaderboard and community tabs so they can communicate with each other for example to look at what the previous ones did, how and why they failed and hopefully add/improve/fix things each time a new agent joins the challenge (instead of starting from scratch everytime).

Let's go agent collaboration! 🦾🦾🦾

See 2 related tweets

• @ChrisHayduk: Bio AI lacks a competition to unleash ML research agents on. Let's fix that

https://t.co/J2Z0v7OnnI...

• @_lewtun: We’ve now got multiple ML interns collaborating together on the hub!

RIP our Claude credits 💀\n\nQ...

10. aakashgupta (Group Score: 71.1 | Individual: 38.3)

Cluster: 2 tweets | Engagement: 176 (Avg: 102) | Type: Tech

This is wild. Cursor hit $2.7 billion ARR last month. The quarter that ended in January, their gross margin was NEGATIVE 23% ❗️

That means for every dollar of revenue, Cursor paid Anthropic and OpenAI $1.23 in API costs. At a$2 billion run rate, that's nearly $2.5 billion a year in cost of goods on software that can't raise prices without losing developers to free alternatives.

This is the context for the deal Cursor signed with SpaceX on Tuesday.

SpaceX announced it has the option to buy Cursor for $60 billion later this year, or pay$10 billion for "the work together" if it walks away. The day before, Cursor was on track to close a $2 billion round at a$50 billion valuation with Andreessen, Nvidia, and Thrive. They halted the raise to take Elon's deal instead.

The math works for Cursor. A $2 billion primary raise buys maybe 18 months of runway against a cost structure where frontier model pricing is set by their two biggest competitors.$10 billion paid out over time from SpaceX is five times the capital with zero dilution, because it's structured as a collaboration fee. Cursor goes from raising at $50 billion to holding a$60 billion floor on the exit.

The math works for Elon. SpaceX merged with xAI in February at a combined $1.25 trillion pre-IPO. xAI's models aren't catching Claude or GPT on coding. Cursor has the distribution to$2.7 billion of paying developer demand and no frontier model of their own. Colossus has a million H100-equivalents of compute that needs to be utilized to justify the capex. The deal plugs all three gaps.

And then there's the constraint Cursor couldn't escape alone. Anthropic launched Claude Code. OpenAI launched Codex. Both labs compete with Cursor directly using the same models Cursor pays rate card for. Every quarter that continues, the labs learn more about the enterprise developer workflow Cursor built while Cursor compounds losses. The negative gross margin was a countdown.

Composer is the exit from that countdown. Cursor shipped their proprietary model in November, added Kimi routing in December, and the margin flipped positive by March. The SpaceX deal provides the capital and the compute to turn Composer from a cost-optimization router into a frontier model trained on Colossus.

The $10 billion "work together" fee is the number that actually matters. That's what Elon was willing to pay for the option. Which means he thinks the option is worth more than$10 billion, which means he thinks Cursor at $60 billion is cheap.

A $2 billion Series E at$50 billion is venture capital pricing a software company. $10 billion from SpaceX at a$60 billion option is an operator pricing a distribution channel.

See 1 related tweets

• @aakashgupta: xAI just offered $60B for a company losing 23 cents on every dollar of revenue.

Cursor's revenue is...

11. aakashgupta (Group Score: 70.8 | Individual: 35.5)

Cluster: 2 tweets | Engagement: 16 (Avg: 102) | Type: Tech

Anthropic's former CPO had to resign from Figma's board. That's because Claude Design is not a small release. It's one of the most important Claude releases yet.

I wrote the full guide: https://t.co/zMmPTi6pks

Here's what people are missing.

Claude Design is the first AI design tool that ships a code-agent handoff. You write a brief. Claude generates a working prototype. Then it bundles the full spec for Claude Code as a structured implementation package. Brief in, working app out. No other tool in the category does this.

It learns your brand from your existing files. Upload your codebase or a Figma export and Claude pulls your design tokens, typography rules, and component patterns. The output looks like your team built it.

Look at the export menu. PPTX, PDF, HTML, Canva. Notice what's missing. "Open in Figma" was a deliberate choice about who the customer is.

The customer is the PM. Figma was sold to designers and procured by companies. Claude Design is sold to the founders, marketers, and product managers who used to need a designer to ship anything visual. That's why Mike Krieger had to step off Figma's board three days before launch. The conflict stopped being theoretical.

Figma's stock dropped 7% on launch day. The cap is structural. Figma still owns the multiplayer canvas, design system governance, and production-grade pixel output. Claude Design wins everything upstream of those. Upstream is where most PMs spend most of their week. Decks for stakeholder reviews. Wireframes for engineering discussions. Landing page mocks for marketing tests. All throughput work that never required Figma-grade polish.

This is Anthropic's vertical integration play. Claude Code for engineering surfaces. Claude Design for product surfaces. Each tool collapses a workflow stage that used to need a separate seat license, a separate vendor, and a separate handoff.

The wait time between PM and designer was always the actual product Figma sold. Anthropic just collapsed it to zero.

The PMs who start using Claude Design this week ship 3-4x faster by Q3.

See 1 related tweets

• @aakashgupta: Mike Krieger gave up his Figma board seat on April 14. Same day The Information broke the Opus 4.7 d...

12. om_patel5 (Group Score: 70.2 | Individual: 28.3)

Cluster: 3 tweets | Engagement: 66 (Avg: 193) | Type: Tech

THIS GUY VIBE CODED GTA 6 ON GOOGLE EARTH IN ONE WEEKEND WITH CLAUDE CODE

he built a browser based GTA style game that runs on top of real google earth cities

drop into any real city on earth and drive through the actual streets

real cops chase you, shoot at you, and arrest you at real police stations

your in car radio auto tunes to real local stations based on where you are in the game (radio garden API)

planes spawn at every real airport, boats spawn at every real port (OSM data)

when you die you respawn at the nearest actual hospital. when you get arrested you respawn at the nearest actual police station

the stack:

• cesium for rendering google 3D tiles in browser
• three.js for vehicles and physics
• radio garden + OSM for the location data
• claude code wrote 80% of the code

this is the type of project that would have taken MONTHS to create

this guy did it in less than 48 hours with barely any effort

the possibilities with vibe coding in the game dev space are actually endless

See 2 related tweets

• @RoundtableSpace: THIS GUY BUILT A GTA-STYLE GAME ON REAL GOOGLE EARTH CITIES IN A WEEKEND

https://t.co/1HtkxUtaud...

• @RoundtableSpace: THIS GUY VIBE CODED A GTA STYLE GAME ON TOP OF GOOGLE EARTH IN A SINGLE WEEKEND WITH CLAUDE CODE.

R...

13. TheAhmadOsman (Group Score: 66.1 | Individual: 24.0)

Cluster: 4 tweets | Engagement: 501 (Avg: 255) | Type: Tech

To celebrate GPT-5.5, OpenAI should reset the limits

If you agree with this message, mention your favorite OpenAI person to let them know\n\nQT @TheAhmadOsman: GPT-5.5 just dropped TL;DR: This is another o1 / o3 moment

• Shorter, more human, actual personality
• Clearly pushing into personal agents (OpenClaw)
• Higher info / Token density → Cheaper to reach GPT-5.4-level intelligence

It codes different:

• Less bloat
• Cleaner, readable output

Codex:

• Frontier agentic coding
• Backend > Claude Opus 4.6
• Give it specs and it will build it
• Handles large codebases, long runs, visual iteration loops

GPT-5.5 Pro:

• Does 30-90 min runs cohesively like it’s nothing
• Writes full docs, uses tools well
• Can solve basically anything

Tradeoffs:

• Slower than Opus
• More expensive per token
• (but more efficient overall)

GPT-5.5 is the new bar

See 3 related tweets

• @WesRoth: OpenAI has made its frontier model, GPT-5.5, available through the API.\n\nQT @OpenAIDevs: GPT-5.5 i...
• @MLStreetTalk: OpenAI are so back!

Loving Codex with GPT 5.5, it's been a joy to use https://t.co/sfBb1Pp8yA\n\nQT...

• @TheRealAdamG: RT @AiBreakfast: I was basically Claude-only for the last 6 months.

but GPT-5.5 pulled me back.

Th...

14. gdb (Group Score: 65.4 | Individual: 25.1)

Cluster: 4 tweets | Engagement: 1321 (Avg: 1222) | Type: Tech

what are you building with codex?\n\nQT @PaulSolt: What app are you making this weekend with GPT 5.5 and Codex?

See 3 related tweets

• @0xSero: I can tell Codex app was heavily assisted by GPT-5.* models. If anyone should write how to make good...
• @thsottiaux: Unfortunately looking at the demos today, the next few weeks of Codex will be eventful and require m...
• @PaulSolt: RT @PaulSolt: New to Codex app?

Here are 7 Beginner Tips to get started with GPT 5.5.

https://t.co...

15. cocktailpeanut (Group Score: 64.2 | Individual: 32.5)

Cluster: 2 tweets | Engagement: 29 (Avg: 24) | Type: Tech

one of the few truly contrarian products in AI. https://t.co/oGSYlmL5tj\n\nQT @ComfyUI: We just raised $30M at a$500M valuation, bringing our total funding to $47M.

Led by @craft_ventures , with @PaceCap , @chemistry , TruArrow, and others.

But before anything else: this belongs to the community.

ComfyUI started as one developer and one open-source repo. No roadmap. No company. Just creators who wanted real control over how they built with AI.

That community is now:

→ 4 million users → 60,000+ community-built nodes → 150,000+ daily downloads

Every number traces back to people who built in the open, for anyone to use.

Here's where the funding goes:

→ Comfy Cloud: for teams and studios that need security and scale → Collaborative workflows: versioning and iteration built for how studios actually work → A better local experience: more seamless, more stable → Ecosystem reliability: making 60,000+ community nodes more dependable → Day-one model support: every major release, compatible at launch

We are not building a walled garden.

We are building open infrastructure, built to last.

Thank you genuinely, The ComfyUI Team

See 1 related tweets

• @IamEmily2050: Congratulations to team ComfyUI, it is a beautiful journey and we still early 🔥\n\nQT @ComfyUI: We j...

16. moneycontrolcom (Group Score: 64.1 | Individual: 26.0)

Cluster: 3 tweets | Engagement: 1 (Avg: 5) | Type: Tech

#Business | AI enters the appraisal room, but humans still make the final call

@shaw_reshab with details: https://t.co/1FzLeXltPL

#AI #Appraisal #IT

See 2 related tweets

• @chandrarsrikant: AI enters the appraisal room, but humans still make the final call

By @shaw_reshab

On a side note,...

• @chandrarsrikant: RT @moneycontrolcom: 🚨 AI enters the appraisal room, but humans still make the final call

@shaw_res...

17. chhddavid (Group Score: 63.7 | Individual: 27.6)

Cluster: 3 tweets | Engagement: 2 (Avg: 13) | Type: Tech

so you're telling me Claude Code Opus 4.7 can now...

• build an app
• launch it
• do email marketing
• self-build features

without any human in the loop??

it's so over... https://t.co/ww9lgRLZ7l\n\nQT @shipper_now: BREAKING: Today, we k*lled vibe coding.

I just watched my Mac wipe out a $31B company in 186 seconds.

This is just absurd. https://t.co/k1tU6zclBZ

See 2 related tweets

• @chddaniel: so you're telling me Claude Code Opus 4.7 can now...

• scan an entire website

• turn it into a mobi...

• @chhddavid: this is like actually terrifying...\n\nQT @shipper_now: BREAKING: Today, we k*lled vibe coding.

I j...

18. DataChaz (Group Score: 63.6 | Individual: 28.4)

Cluster: 3 tweets | Engagement: 569 (Avg: 201) | Type: Tech

Do you actually realize what's happening?

Deepseek just dropped a 1.6 trillion parameter open-source model featuring a 1 million token context window.

OpenAI is charging $200/month.

but China is giving it away completely free.

Let that sink in. https://t.co/ROwyVyIFY1\n\nQT @deepseek_ai: 🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.

🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.

Try it now at https://t.co/GCdiMzk1Dl via Expert Mode / Instant Mode. API is updated & available today!

📄 Tech Report: https://t.co/drlDrxkYtp 🤗 Open Weights: https://t.co/T13Y8i7SDM

1/n

See 2 related tweets

• @RnaudBertrand: RT @cgtwts: > be chinese ai labs

while claude and openai are in cold war kimi dropped k2.6 using...

• @DataChaz: RT @DataChaz: Do you actually realize what's happening?

Deepseek just dropped a 1.6 trillion parame...

19. kazunori_279 (Group Score: 62.6 | Individual: 62.6)

Cluster: 1 tweets | Engagement: 3360 (Avg: 163) | Type: Tech

RT @VaibhavSisinty: Elon's team just did something nobody's talking about.

They replaced Starlink's call center with an AI. 🤯

And the results are insane.

1 in 5 people who called Starlink bought Starlink on the call. 70% of support calls got resolved without a human ever touching them.

Every caller was talking to Grok.

It's called Grok Voice Think Fast 1.0.

And this is not a voice model. This is a call center in a box.

Plug it into a phone line. It picks up. It talks. It sells. It troubleshoots. It books. In 25 languages. With no lunch break.

A single Grok agent runs 28 tools across hundreds of workflows. Hardware troubleshooting. Replacement orders. Service credits.

All autonomous.

The global call center industry is worth $350 billion.

Elon just showed up with a demo that does the job better, cheaper, faster, and it never clocks out.

Everyone is busy teaching AI to write code.

Elon quietly taught it to take your money on the phone.

20. Grady_Booch (Group Score: 62.4 | Individual: 30.9)

Cluster: 3 tweets | Engagement: 189 (Avg: 229) | Type: Tech

Thoughts and prayers, Sam and Larry.

Thoughts and prayers.\n\nQT @edzitron: Premium Newsletter: OpenAI needs to raise or make $852 billion through 2030 to pay for its compute costs, and must pay$75bn a year for Oracle's 7.1GW of Stargate data centers, or Oracle will collapse, destroying its share price and Ellison's empire. https://t.co/fmxKU8v0bi

See 2 related tweets

• @edzitron: RT @edzitron: Premium Newsletter: OpenAI needs to raise or make $852 billion through 2030 to pay for...
• @JayminSOfficial: RT @WSJ: Oracle’s $300 billion megadeal with OpenAI is testing the limits of Wall Street’s appetite ...