热门科技推文 — 2026年7月29日

今日科技动态中，人工智能的开放性与基础设施成为讨论焦点。月之暗面发布了参数规模达2.8万亿的Kimi K3模型；与此同时，行业领袖围绕开放权重展开讨论，英伟达则支持安全倡议及安全超级智能公司（Safe Superintelligence）。语音克隆初创企业Fish Audio宣布完成5200万美元种子轮融资，Meta与贝莱德合作建设一座投资140亿美元的得克萨斯州数据中心。此外，智能体工具正在加速应用开发、编程、研究和机器人技术的发展，但围绕人工智能公司为获取训练数据而收购并销毁珍稀书籍的争议也日益升温。

1. ShriramKMurthi (Group Score: 588.8 | Individual: 65.8)

Cluster: 14 tweets | Engagement: 5728 (Avg: 252) | Type: Tech

RT @charliedbecker: I'm a second-generation bookseller. My family runs Houston's largest used & rare bookstore and I'm building an AI tool for used bookstores. We got hit by exactly these orders, including a single order for 70 obscure books that made us pause online sales entirely. So I dug in. I don't think this connects the way people want it to.

Every ingredient of the viral story is true. ISBNdb really does advertise sourcing books for AI companies. Anthropic really did destructively scan millions of books though they were mostly acquired through things like library deaccessions, not bookstore inventory. And booksellers really are getting bizarre bulk orders. But "rare" here isn't what you're picturing. It's The Insider's Guide to Metro Denver (1995) and how-to-use-WordPerfect 1991 manuals. Obscure, but often not precious.

When we combed through our orders, nearly every book bought from us was either unavailable on Amazon or listed there at 5, 10, 20x our price on the platform they were purchased on. And the shipments are going to FBA prep companies: which is what you do with a book you're about to resell, not one you're about to cut apart. There is more than circumstantial evidence that this is a well-funded company running algorithmic arbitrage, buying underpriced or out-of-stock books to relist on Amazon at markup. This evidence includes an official denial that they are selling to train AI but I don't want to mention them for several reasons. You can Google it.

That's more mundane than the shredder story. The actual risk is worse, though, because it's quieter: books that don't sell via FBA eventually get liquidated. If we really hold the last copies of some of these (and comparing platforms, it looks like we might) we're feeding them into a portal they may not come back from. No AI company required.

Everyone assumes the internet preserved everything. It didn't. Many of these books have almost no metadata anywhere, just a scattered listings across proprietary databases then nothing. That's what my project is for: building better bibliographic records for these books than exist anywhere else, so we at least have a map of what we stand to lose. So the Library of Alexandria is still burning, just more slowly, and kore from “who cares” than from some cinematic villain.

(I am open to being corrected.)

Full writeup: https://t.co/tHXMH0L3us

See 13 related tweets

@chandrarsrikant: RT @HedgieMarkets: 🦔AI companies are bulk-buying rare books, scanning them through high-speed machin...
@BrianRoemmele: RT @BrianRoemmele: Confirmation of Anthropic’s BOOK BURNING for their AI: The Great Forgetting of th...
@jason: This is disgusting\n\nQT @HedgieMarkets: 🦔AI companies are bulk-buying rare books, scanning them thr...
@matanSF: I've always been senitmental about physical books. The way each carries with it the places you've be...
@Grady_Booch: The burning of the Library of Alexandria v2\n\nQT @HedgieMarkets: 🦔AI companies are bulk-buying rare...

2. VaibhavSisinty (Group Score: 534.3 | Individual: 36.8)

Cluster: 21 tweets | Engagement: 39 (Avg: 76) | Type: Tech

The most important part of Dario Amodei’s response to Jensen Huang?

He isn’t calling for a ban on open-weight models.

In fact, Dario agrees with several points from Jensen’s letter:

• Open weights expand access to AI • They increase competition • They give users and enterprises more control

Where Anthropic disagrees:

• Open models don’t automatically benefit defenders more than attackers • Advanced open models are harder to secure because safety guardrails can’t be universally enforced after release • Frontier AI capabilities, especially in areas like biology, could introduce risks that require additional oversight

Anthropic’s position is more nuanced than many headlines suggest:

→ No ban on open-weight models → Safety testing for both open and closed models → Crackdowns on chip smuggling and industrial-scale distillation → Focus on preventing frontier AI capabilities from empowering adversarial state actors

The debate is no longer Open vs Closed.

It’s becoming:

How do we preserve openness, competition, and innovation while managing genuinely dangerous capabilities?

That’s the question the AI industry will have to answer.\n\nQT @AnthropicAI: There’s been a lot of speculation about where we stand on open-weights models. We’ve outlined our views in full here: https://t.co/NtXQuWm2g5

See 20 related tweets

@MTSlive: SITUATION EXPLAINED: Dario Amodei responds directly to Jensen Huang's open-source letter.

• Dario's...

@MLStreetTalk: > "Anthropic has never advocated for a ban on open-weights models" ... "that don’t have dangerous ca...
@jason: Open source, open weight models surging… now we just need app layer/harness/interfaces that make the...
@quxiaoyin: Dario finally supports open. But his letter STILL contains potential loopholes: The "required test b...
@AdrianDittmann: Basically, Anthropic wants the government to regulate and control who has access to AI models (both ...

3. zaynmcps (Group Score: 420.4 | Individual: 65.0)

Cluster: 12 tweets | Engagement: 10822 (Avg: 410) | Type: Tech

RT @FishAudio: Today we’ve raised $52M Seed and we are announcing the public launch of S2.1 Pro.

It can clone a voice from 5 seconds of audio 2x faster than Cartesia & 1/6th the cost of Eleven Labs most expressive model with word level control over emotion, intonation, pacing etc

We support frontier AI companies including HeyGen, LiveKit, Retell, Sanas, and OpenArt all run our model in production.

If you're a business and we can't cut your voice AI costs by 50%, we'll give you 1 year of Fish Audio for free. Book a demo: https://t.co/vHkyZf9JoG

To celebrate our first birthday, we'll give you 1 month of S2.1 Pro for free. Like, retweet, and comment “Fish” to get it.

See 11 related tweets

@aakashgupta: Fish Audio is used by 8 million people and has 30K+ stars on its GitHub repo.

If you ask them how t...

@VadimStrizheus: do you understand what just happened?!

an ex-NVIDIA engineer just launched S2.1 Pro, an AI voice t...

@dr_cintas: 🚨Fish Audio just launched S2.1 Pro

a production voice model for live agents: streaming, low latency...

@kimmonismus: A few years ago you could spot an AI voice in about a second. Flat, robotic, that dead text-to-speec...
@cgtwts: > be Fish Audio > hear “the market is already won” every week > build anyway > open-sour...

4. RnaudBertrand (Group Score: 335.2 | Individual: 62.7)

Cluster: 8 tweets | Engagement: 12530 (Avg: 1330) | Type: Tech

RT @Kimi_Moonshot: Releasing the model weights and technical report of Kimi K3.

Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.

New model architecture: 2.5x the intelligence per unit of compute, not just more params.

Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.

Model weights: https://t.co/7m7eEg6Y0B Tech report: https://t.co/yeu6cjpMCT Tech blog: https://t.co/YTfiMSNM1f

See 7 related tweets

@shiri_shh: "API costs are too high, let's just self-host this 2.8T model on our own." https://t.co/kRAjEdkAk6\n...
@yacinelearning: I can’t shake the feeling that this may mark the end of the fully open frontier research era\n\nQT @...
@guohao_li: kimi k3 is now available on @Eigent_AI cloud (local-first open source cowork) for free trials

downl...

@guohao_li: now the most important agi task is to use frontier intelligence to develop more, cheaper, faster, sm...
@tugot17: same picture btw https://t.co/bAkEZkpLoI\n\nQT @Kimi_Moonshot: Releasing the model weights and techn...

5. Daniel_Farinax (Group Score: 243.0 | Individual: 38.9)

Cluster: 9 tweets | Engagement: 152 (Avg: 111) | Type: Tech

This is how to make and publish a full app with the new Grok Build in just a few minutes from https://t.co/xDErUd9C4L, iOS or Android 🔥

Go to https://t.co/xDErUd9C4L → select Build Tell it what you want It pulls data, loads skills, and creates apps fast Hit Publish → and your app or site is live! No code. No setup. Just describe it and ship.

Have fun building 🚀\n\nQT @grok: Build apps from https://t.co/f3u3RwhHjk, iOS, and Android. With one prompt, turn an idea into a published product with its own domain.

Now available for SuperGrok Heavy users. https://t.co/Dg1WyC3qeo

See 8 related tweets

@XFreeze: SpaceXAI has just introduced the powerful new Grok App Builder, turning a simple conversation into a...
@Baconbrix: You can now create apps directly in the Grok app with Build mode!

Available on all platforms in ear...

@Baconbrix: I asked Grok to simulate a black hole composed of the word grok

https://t.co/iwFnmie791\n\nQT @Baco...

@testingcatalog: SPACEXAI 🔥: A new Grok Build mode has been introduced on Grok web and mobile.

Users can build apps...

@XFreeze: You can just build things with Grok now, right from the app, without ever touching a laptop

The cap...

6. MTSlive (Group Score: 173.9 | Individual: 36.4)

Cluster: 5 tweets | Engagement: 63 (Avg: 51) | Type: Tech

DAILY SITUATION RECAP:

Nvidia launches the Open Secure AI Alliance in order to find and fix vulnerabilities using open-source AI, sort of like an open Project Glasswing. Founding partners include a mix of enterprise software companies (Databricks, Salesforce, IBM, SAP, Siemens, Snowflake), cybersecurity companies (Palo Alto Networks, Red Hat), open-source providers (Hugging Face, LangChain, OpenClaw, Nous, the Linux Foundation), AI labs (SpaceXAI, Thinking Machines, Cognition), and other major companies (Nvidia, Microsoft, Cisco, Palantir, Dell).

Moonshot AI releases the Kimi K3 weights and technical report after eleven days since launch. Kimi K3 is a 2.8T parameter mixture-of-experts (MoE) model with 104B active parameters and a 1M token context window. Moonshot also open-sourced much of their infrastructure, including their attention kernels, agent environment platform, and MoE communication library. Just because you can download it in theory doesn’t mean you actually can — the model is far too big to be run on any consumer hardware.

Nvidia invests $5B in Ilya Sutskever’s SSI. Sutskever, formerly co-founder and Chief Scientist of OpenAI, founded Safe Superintelligence in 2024 with the sole goal of building a safe superintelligence, with no other products along the way. It has since raised$ 3B at up to a $32B valuation (likely higher now). SSI is famously very secretive about its research, but Sutskever said it’s “focused on overlooked aspects of how the human brain functions”. The new funding, and access to Nvidia Vera Rubin GPUs, will allow SSI to 10x its compute.

More companies sign on to Nvidia’s open source letter. The letter, posted by Jensen Huang on Friday, advocates for a robust American open-source ecosystem with minimal government regulation. New signatories include Google, SpaceXAI, OpenAI, AMD, Cisco, Palo Alto Networks, Nebius, Scale, Fireworks AI, Baseten, Cohere, Sakana AI, Periodic Labs, Core Automation, OpenClaw, and GitHub. Every major American frontier lab except for Anthropic has now signed.

CXMT stock surges 466% on its first trading day. The company, formerly ChangXin Memory Technologies, is the largest memory manufacturer in China and the fourth-largest in the world (after SK Hynix, Samsung, and Micron), with a 9% global market share. It now has the second-highest market cap of any Chinese company after Tencent. CXMT doesn’t make the most leading-edge HBM for AI chips, but supplies DRAM to consumer tech manufacturers and data centers.

Nvidia may guarantee $250-350B of financing for an OpenAI data center. SB Energy, a subsidiary of SoftBank, is developing a massive 10 GW data center on federal land in Ohio at a total cost of over$ 500B. The financing guarantee would allow SB Energy to borrow money at lower rates, and possibly allow OpenAI to spend more on Nvidia chips.

China begins manufacturing DUV machines. Deep ultraviolet (DUV) lithography machines print intricate nanoscale patterns on silicon wafers, a critical step in chipmaking. The new machines, built by an unnamed state-backed company, will be shipped to local chipmakers including SMIC, Hua Hong Semiconductor, and CXMT. China is still behind on the most advanced extreme ultraviolet (EUV) lithography, which is solely produced by Dutch company ASML. ASML stock fell 6% on the news.

Dario Amodei explains Anthropic’s position on open models: open-weight models without dangerous capabilities are a public good, and Anthropic has never supported a full ban. However, we should be worried about the CCP using them for repression, as well as cyber/bio/alignment risk. To that end, we should not sell chips to China, crack down on distillation, and require mandatory safety testing for all sufficiently capable open and closed models.

China threatens to respond if the US sanctions their AI labs. The Chinese Ministry of Commerce said US accusations of distillation were “smears” and that China will “take all necessary measures” to defend its rights and interests against any action that substantively harms them.

DeepSeek has suspended its recent funding round after comments from a private investor call with CEO Liang Wenfeng were leaked.

Written by @theojaffee. Read more at our link in bio.

See 4 related tweets

@teortaxesTex: Nvidia playing all teams\n\nQT @MTSlive: DAILY SITUATION RECAP:

Nvidia launches the Open Secure AI ...

@shanaka86: Washington did not miss this one. On 22nd July the White House named the exact chips, the exact coun...
@WesRoth: Nvidia just invested $5 billion in Ilya Sutskever’s secretive AI lab. But it may be buying something...
@MTSlive: SITUATION EXPLAINED: Sam Altman returns to DC this week to face questions on a new AI safety framewo...

7. BrianRoemmele (Group Score: 151.2 | Individual: 46.2)

Cluster: 6 tweets | Engagement: 9268 (Avg: 327) | Type: Tech

RT @BrianRoemmele: Some folks don’t believe that a “principled AI company” would destroy books after they suck out all the information for their models, even one-of-a-kind books and manuscripts.

Here it is admitted in open court. https://t.co/PHrqYFCjOG

See 5 related tweets

@ednewtonrex: Incredibly important new research published today shows the huge impact AI books are having on real ...
@seconds_0: RT @TuhinChakr: 🚨 New paper on AI and Copyright 🚨

It's easy to dismiss AI-written books 📚 as slop n...

@deliprao: Imagine being so evil monopolistic that you decide to destroy copies of physical books no one can fi...
@ns123abc: HOLD ON

WHAT\n\nQT @BrianRoemmele: Some folks don’t believe that a “principled AI company” would d...

@mark_k: Unpopular opinion: There's no problem with destroying books after you or your AI have read them. If ...

8. a16z (Group Score: 150.5 | Individual: 36.0)

Cluster: 5 tweets | Engagement: 74 (Avg: 298) | Type: Tech

.@AppliedInt wants to make robotics as easy as building an iPhone app:

"Dana is our new agentic platform for physical AI. This really is the culmination of pretty much everything we've worked on now over the last decade."

"Building physical AI is extremely complicated... We know how to make the chips. We know how to make the hardware for these systems. It's more so actually developing all of this technology and getting it to work that is very, very difficult."

"With modern AI and with this new Dana platform, we're really drastically reducing the barrier to entry to building physical AI and deploying it in the real world."

"If you're building an app for an iPhone, high school kids can do that now because there's all of these things that exist and that make it easy... It's very hard to do that in terms of robotics."

"The thing that Peter's really emphasizing and thing that we really wanna do is just lower the bar. So anybody really out there, starting with engineers, but ultimately really anybody can develop robots."

Applied Intuition Co-founders Peter Ludwig and @qasar on @bizbreakdowns\n\nQT @a16z: .@AppliedInt Co-founder and CTO Peter Ludwig says "a billion machines will become autonomous or intelligent over the next ten years."

"Cars, trucks, tractors, mining haulers, defense systems, warehouse robots, humanoids—the physical economy will be rebuilt around software that perceives, decides, and acts."

"The prevailing assumption about how we get there goes something like this: models keep improving, world models mature, foundation models for robotics arrive, and autonomy falls out the other end. Intelligence is the whole game; scale the intelligence and the machines will follow."

"The contrarian bet, then, isn't against intelligence. It's that the next order of magnitude in physical AI comes from making the engineering system as intelligent as the models it carries."

Full piece on compounding systems and Applied Intuition's newest platform, Dana: https://t.co/yiPrruUrro

See 4 related tweets

@AppliedInt: “Twenty years from now, we won't remember which company had the best world model in 2027. We'll reme...
@a16z: .@AppliedInt Co-founder and CTO Peter Ludwig says "a billion machines will become autonomous or inte...
@a16z: .@AppliedInt CEO Qasar Younis says physical AI can't arrive fast enough:

"By putting self-driving a...

@a16z: .@AppliedInt CTO Peter Ludwig says the model is only 1% of a physical AI system:

"General purpose m...

9. eisokant (Group Score: 145.9 | Individual: 32.7)

Cluster: 5 tweets | Engagement: 86 (Avg: 83) | Type: Tech

What I really like about this release is that it is fully built around ACP (Agent Context Protocol). That means Bring Your Own Agent & Bring Your Own Model. You can use multiple agents, handover from one to the other, no lock in whatsoever. This has quickly become the daily driver for quite a few folks on our team. Give it a spin!!\n\nQT @poolsideai: Today, we’re releasing Poolside Desktop Assistant.

One place to run coding agents across macOS, VS Code, and Visual Studio.

We built it for ourselves and have used it every day for the past year. Now we’re opening it up to everyone. https://t.co/b4L4pGITZK

See 4 related tweets

@nathanbenaich: for the billions of knowledge workers, you can now use @poolsideai models right from a familiar maco...
@zeddotdev: Congrats to @poolsideai on launching Poolside Desktop Assistant.

Their Laguna models power an ACP a...

@airstreet: big day @poolsideai! you now run poolside models as well as other open and closed models from your M...
@poolsideai: RT @jasoncwarner: Today we’re releasing the Poolside Desktop Assistant, one workspace for running co...

10. chandrarsrikant (Group Score: 145.7 | Individual: 58.9)

Cluster: 4 tweets | Engagement: 15676 (Avg: 1689) | Type: Tech

RT @itsolelehmann: btw anthropic's internal document on this literally said "we don't want it to be known that we are working on this.”

it was called project panama.

here's exactly what happened:

1: anthropic concluded that books were the cheapest way to build a world-class model because they gave claude curated facts, structured arguments, compelling stories, and writing “an editor would approve of.”

2: once anthropic decided it needed books at enormous scale, its first solution was piracy.

it downloaded 7m+ books from online libraries including libgen. the judge later wrote that although anthropic had legal ways to buy them, it chose piracy to avoid what dario amodei called the “legal/practice/business slog.”

3: that piracy created a massive legal risk.

so in february 2024, anthropic hired tom turvey, the former head of partnerships for google books, to find a legally safer way of obtaining “all the books in the world.”

4: turvey first contacted major publishers about licensing their catalogs.

those attempts didn’t produce agreements, so anthropic chose a route that required no publisher permission: buying millions of physical books through distributors and used-book retailers.

5: within about a year, anthropic spent tens of millions acquiring and scanning millions of books, including many rare and 1/1 titles. one vendor proposal targeted 500,000 to 2 million books in six months.

6: to scan that many books within months, the vendors physically dismantled them.

a hydraulic cutter removed each spine. the pages were trimmed to size, fed as loose sheets through high-speed industrial scanners, and converted into searchable PDFs. the paper remains were then sent for recycling.

7: these PDFs were fed into claude as training data.

the complete collection became a private, searchable anthropic library that the company planned to “store forever.” the scans aren’t available to the public and were never open-sourced.

See 3 related tweets

@jason: I think the books are being destroyed to hide the evidence, in the non-zero chance that the courts d...
@MsMelChen: Of course it’s the self-appointed safety-first moral adults of the AI industry who were bulk-buying ...
@Pirat_Nation: Anthropic reportedly bought millions of physical books, scanned them into digital copies, and destro...

11. wallstengine (Group Score: 141.3 | Individual: 36.4)

Cluster: 6 tweets | Engagement: 248 (Avg: 157) | Type: Tech

$META & BLACKROCK FORM$ 14B VENTURE FOR 1GW TEXAS DATA CENTER

Meta will be the campus’s initial sole occupant, using the capacity to support AI-model development and enhancements to its core business. The venture expects to begin bringing the capacity online in 2028.

Funds managed by BlackRock will own 80% of the venture, while Meta will retain 20%. The partners have committed to fund their respective shares of approximately $14 billion in development costs covering the buildings, power, cooling, and connectivity infrastructure.

At closing, Meta will contribute land and construction-in-progress assets valued at approximately $2.3 billion. BlackRock-managed funds will contribute about$ 4.9 billion in cash. Meta will receive a one-time distribution of approximately $1 billion to align the venture with the 80/20 ownership split.

The financing structure also includes $12.5 billion of debt.

Meta will lease the entire campus for an initial four-year term, with four extension options providing flexibility over a potential 20-year period. The company is also providing residual-value guarantees with an aggregate threshold of approximately $13 billion that declines over time.

The El Paso project represents an investment of more than $10 billion from Meta. The companies say it will support more than 4,000 construction jobs at peak and create 300 operational positions once completed. More than 2,300 workers are already onsite.

See 5 related tweets

@StockMKTNewz: Meta Platforms $META and BlackRock$ BLK will establish a joint venture to build and operate a 1-giga...
@LoganDobson: "...supporting more than 4,000 construction jobs at peak and 300 operational jobs once complete...

...

@WesRoth: Meta just partnered with BlackRock to build a $14 billion AI data center but Meta will own only 20% ...
@StockSavvyShay: $META and$ BLK are forming a $14B venture to develop a 1GW AI data center campus in Texas with Meta ...
@exec_sum: BREAKING: Meta and BlackRock have partnered on a $14 billion data center campus in Texas

BlackRock ...

12. eng_khairallah1 (Group Score: 128.6 | Individual: 32.3)

Cluster: 5 tweets | Engagement: 86 (Avg: 72) | Type: Tech

Boris Cherny, creator of Claude Code:

"Opus 5 does in a day what used to take your team a month. Most people will keep using it wrong."

In 58 minutes he explains why Opus 5 needs less prompting than any model before, and why your old detailed prompts now work against you.

Watch it, then read the article below on how to prompt Opus 5 👇\n\nQT @eng_khairallah1: https://t.co/mLfo3ZV2Vw

See 4 related tweets

@sairahul1: RT @sairahul1: ANTHROPIC JUST DROPPED THE OFFICIAL GUIDE TO PROMPTING OPUS 5

This is the most impor...

@RoundtableSpace: READ THIS BEFORE YOUR NEXT PROMPT:

ANTHROPIC JUST RELEASED THE OFFICIAL PROMPTING GUIDE FOR OPUS 5....

@sairahul1: RT @sairahul1: Boris Cherny, creator of Claude Code:

"Opus 5 does in a day what used to take your t...

@eng_khairallah1: RT @eng_khairallah1: Boris Cherny, creator of Claude Code:

"Opus 5 does in a day what used to take ...

13. Parul_Gautam7 (Group Score: 124.3 | Individual: 29.6)

Cluster: 5 tweets | Engagement: 87 (Avg: 120) | Type: Tech

developers already have AI embedded into how they work.researchers are now starting to make the same shift:

> less repetitive work > more time for experiments > faster iteration

excited to see programs like Reactorfield helping make that happen.\n\nQT @jamessinka: AI transformed coding, science is next

@ReactorfieldAI makes scientists and deep tech startups AI-native

Built by me and @berkbuilds to accelerate science with frontier AI https://t.co/4rETjNfepp

See 4 related tweets

@omarsar0: I spent the last two years rebuilding my work around agents. Research, writing, and my daily analysi...
@p_maverick_b: If you're deep into hard tech but skeptical about AI, maybe it's worth trying to learn as much as yo...
@PeterDiamandis: RT @jamessinka: AI transformed coding, science is next

@ReactorfieldAI makes scientists and deep te...

@Parul_Gautam7: RT @Parul_Gautam7: developers already have AI embedded into how they work.researchers are now starti...

14. ReutersLegal (Group Score: 124.2 | Individual: 28.6)

Cluster: 6 tweets | Engagement: 0 (Avg: 2) | Type: Tech

Nvidia will make a $5 billion equity investment in Safe Superintelligence as part of a strategic partnership between the chipmaker and the startup co-founded by OpenAI's former chief scientist Ilya Sutskever, a person briefed on the deal said. https://t.co/xQ1zWCEkXO

See 5 related tweets

@Reuters: Nvidia to invest $5 billion in Ilya Sutskever's AI startup, source says https://t.co/Ss27bywD1z http...
@Cointelegraph: 🚨 BIG: Nvidia will invest $5 billion in Safe Superintelligence, the AI startup co-founded by OpenAI'...
@pstAsiatech: Nvidia has committed to invest $5 billion in Ilya Sutskever’s artificial intelligence startup Safe S...
@business: Nvidia has disclosed a fresh round of mega-sized commitments to other tech companies over the last f...
@rohanpaul_ai: RT @rohanpaul_ai: Ilya Sutskever’s Safe Superintelligence partners with Nvidia to scale its compute ...

15. aakashgupta (Group Score: 121.8 | Individual: 35.1)

Cluster: 4 tweets | Engagement: 11 (Avg: 165) | Type: Tech

Moonshot just gave away the largest AI model ever built, and their revenue went up 6x.

Kimi K3 is 2.8 trillion parameters, open weights, free to download, fine-tune, self-host, even resell. Standard business logic says this torches the company. The numbers say the opposite. Daily revenue has grown at least 6x since K3 launched. ARR hit $300M in June, up from$ 200M in April. They're now raising at a $50 billion valuation.

The mechanism is distribution. Every self-hosted K3 deployment is a free trial, and the paid API is OpenAI SDK compatible, so migrating off Anthropic or OpenAI takes one line of config. Developers in blind arena evals were already picking K3 over US models on frontend coding before the weights even dropped.

Now look at who actually paid for the free release. On K3's debut, https://t.co/eVnxi2G6Fw fell 30% in Hong Kong. MiniMax dropped 16%. Alibaba lost 4%. The Nasdaq slid about 1%. A free download erased more market cap in one day than most paid launches generate in a year.

That's the trade. Moonshot can't win a premium pricing auction against Claude Fable 5 and GPT-5.6 Sol, which still beat K3 at the top. So instead of competing for the ceiling, they reset the floor for everyone underneath it.

Giving away the model cost Moonshot nothing they could sell anyway. It cost their competitors a third of their market cap in a day.\n\nQT @Kimi_Moonshot: Releasing the model weights and technical report of Kimi K3.

Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.

New model architecture: 2.5x the intelligence per unit of compute, not just more params.

Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.

Model weights: https://t.co/7m7eEg6Y0B Tech report: https://t.co/yeu6cjpMCT Tech blog: https://t.co/YTfiMSNM1f

See 3 related tweets

@Malay4Product: I run roughly 25 agents across Ayuda and model cost is the line item I check every month.

Kimi K3 ...

@sudoingX: kimi number #1 again, and not by a hair, kimi k3 at +9.8% net improvement on the agent arena, measur...
@WesRoth: Kimi K3 is now the highest-scoring open-weight model tested by Artificial Analysis.

It scored 57 on...

16. Reuters (Group Score: 118.5 | Individual: 23.0)

Cluster: 7 tweets | Engagement: 62 (Avg: 84) | Type: Tech

Meta, BlackRock partner on $14 billion El Paso data center venture https://t.co/MznF39nD9F https://t.co/MznF39nD9F

See 6 related tweets

@MetaNewsroom: Meta and BlackRock today announced a new venture to develop and own a data center campus in El Paso,...
@Techmeme: Meta and BlackRock form a strategic venture to develop a 1GW data center campus in El Paso, set to c...
@Forbes: BlackRock Will Own 80% Of Meta's Massive New AI Data Center In El Paso https://t.co/69MUZMcSsR https...
@WhaleInsider: JUST IN: 🇺🇸 Meta and BlackRock to develop data center campus in El Paso, Texas...
@business: BlackRock wrapped up a $12.5 billion bond sale tied to a Meta data center project in Texas after a n...

17. FellMentKE (Group Score: 117.6 | Individual: 32.3)

Cluster: 4 tweets | Engagement: 127 (Avg: 140) | Type: Tech

Most AI slide tools ship locked images that fall apart in PowerPoint but Templafy is a massive technical shift.

The new PowerPoint Agent turns a prompt into a fully editable .pptx file with native formatting and custom color themes.

This is the perfect execution engine for enterprise decks.

Try it for FREE: https://t.co/DZ58gQvLXm\n\nQT @Templafy: Creating PowerPoint presentations with Templafy's free AI Agent just got even better!

🗨️Chat history 🎨Custom brand colors 📊Editable charts

Create presentations—fully editable in PowerPoint: https://t.co/3Bj4XkHyaz https://t.co/fGt6w72fiL

See 3 related tweets

@ai_explorer25: I didn't think an AI could build a real PowerPoint deck. I was wrong.

28 slides in 4 minutes , open...

@cgtwts: All those years of refusing to learn powerpoint is about to finally pay off https://t.co/Df1PxQjpFS\...
@chenzeling4: Every presentation tool wants $10/mo now. This one is free.

Banana Slides generates editable PPTX f...

18. ANI (Group Score: 116.6 | Individual: 32.8)

Cluster: 4 tweets | Engagement: 57 (Avg: 278) | Type: Tech

Pavan Kumar Chandana, Co-founder and Chief Executive Officer (CEO), and Naga Bharath Daka, Co-founder and Chief Operating Officer (COO) of Skyroot Aerospace, called on Chief Minister of Telangana, A. Revanth Reddy, at the Delhi Camp Office today.

During the meeting, the Skyroot Aerospace founders briefed the Chief Minister on the company’s recent landmark achievement of successfully launching Vikram-1, India’s first privately developed orbital-class launch vehicle, on 18 July 2026. The historic mission made India the third country, after the United States and China, to achieve private-sector orbital launch capability.

The meeting also featured discussions on Telangana’s growing aerospace and space technology ecosystem, innovation, advanced manufacturing, research and development, and opportunities for collaboration between the Government of Telangana and the private space industry.

The Chief Minister congratulated the founders on their remarkable achievement and appreciated their contribution to strengthening India’s position in the global space sector. He reiterated the Telangana Government’s commitment to fostering innovation and creating a world-class ecosystem for startups and high-technology industries in the State.

(Pics Source: Telangana CMO)

See 3 related tweets

@SkyrootA: Skyroot’s founders were honoured to meet Hon’ble Union Minister Shri @PiyushGoyal.

The reforms cham...

@PawanKChandana: It is always a pleasure meeting Hon’ble Union Minister of Commerce & Industry Shri @PiyushGoyal ji. ...
@SkyrootA: Thank you Hon’ble Chief Minister @DKShivakumar 🇮🇳🚀\n\nQT @DKShivakumar: Delighted to meet the founde...

19. kimmonismus (Group Score: 114.4 | Individual: 35.2)

Cluster: 4 tweets | Engagement: 2448 (Avg: 791) | Type: Tech

I thought this was a joke wtf: Claude Opus 5 was left running for 24 hours straight and built an entire game from scratch. without any external assets or code. Prompts, workflow, code are included.

https://t.co/eA4c1IV06W

See 3 related tweets

@threejs: RT @anshuc: I ran Claude Opus 5 in one marathon session for 24 hours straight, and ended up with thi...
@altryne: This is epic\n\nQT @anshuc: I ran Claude Opus 5 in one marathon session for 24 hours straight, and e...
@alex_prompter: RT @alex_prompter: People are building playable games with Claude Opus 5.

Every one of these is a r...

20. Hadas_Gold (Group Score: 111.2 | Individual: 27.4)

Cluster: 7 tweets | Engagement: 75 (Avg: 544) | Type: Tech

Sam Altman says the Hugging Face hack is the first security incident he felt "viscerally" and that they paused training. He is now questioning if "we may have to pace the rate of AI development to give ourselves enough time for society to harden around these new capability levels" https://t.co/4wqhUOdKvl\n\nQT @patrick_oshag: My conversation with Sam Altman (@sama), CEO of OpenAI.

We discuss:

Kimi, distillation, and open source
OpenAI's compute bets
The Hugging Face incident
What happens after AGI
Raising kids in an age of abundant intelligence
And much more

Enjoy!

TIMESTAMPS 0:00 Intro 4:10 The Race for Compute 14:24 A Sci-Fi Cyber Incident 16:14 The Promise and Risks of AGI 23:27 How AI Will Change Jobs 29:38 Sam’s Vision for a Personal AI 35:02 Robotics, ChatGPT, and What Comes Next 44:39 The Weight of Leading OpenAI 51:36 OpenAI’s Biggest Lessons

See 6 related tweets

@firstadopter: On compute commitments:

"We did under do it " - Sam Altman

4:10 The Race for Compute\n\nQT @patric...

@techreview: This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s...
@Cointelegraph: 🚨 LATEST: OpenAI's rogue AI agent breach that targeted Hugging Face is a "warning shot" for AI-drive...
@unusual_whales: BREAKING: OpenAI CEO Sam Altman has said that AI singularity has arrived, per ABC...
@CNBCMorningCall: .@ashleycapoot, AI Reporter at @CNBC said OpenAI will preview a more powerful AI model as Sam Altman...