科技热门推文——2026年8月19日

今日科技动态：人工智能发展在快速扩张与审慎推进之间寻求平衡。据报道，OpenAI暂停了前沿强化学习训练，以开展安全审查，同时推出面向青少年的ChatGPT。工程团队也获得了多项新工具，可用于调试强化学习模型、通过Claude Code进行设计，以及构建个性化Grok机器人。与此同时，据报道，人工智能芯片初创公司Etched以210亿美元估值完成7亿美元融资，凸显投资者对专业化基础设施的热情。谷歌计划将Pixel供应链进一步转向越南和印度，也反映出科技行业持续推动制造基地多元化的趋势。

1. EmmaUsesAi (Group Score: 343.0 | Individual: 54.9)

Cluster: 11 tweets | Engagement: 848 (Avg: 66) | Type: Tech

RT @ishanmkh: Rox has been powering Global 2000 enterprises with $5T+ in combined market cap.

Today, we’re putting it in everyone’s hands.

Over the last 2 years, companies like @MongoDB, @togethercompute, and @Xbow shifted investment from legacy CRM and SaaS to Revenue Agents.

But most teams have been locked out.

They lacked in-house technical talent or FDEs required to set up the revenue-specific context, harnesses, and agent systems needed to run Revenue Agents in production.

Today, we launched Rox Teams to remove that barrier.

With Rox Teams, businesses of every shape, size, and vertical can activate Revenue Agents on their own.

Set up in minutes. No FDEs required.

https://t.co/DfzpGtKZJD

& Nolaned this hilarious explainer video 👇

See 10 related tweets

@aakashgupta: Rox built its Revenue Agents business on Global 2000 companies, $5T+ in combined market cap by their...
@omarsar0: The revenue stack is collapsing into agents.

A seller's week is mostly account research, follow-ups...

@alex_verem: You can vibe code a sales agent in a weekend.

Then give it 5,000 accounts, years of CRM garbage, ca...

@Restructuring__: Great slide illustrating how labor is no longer human-limited

That said, until humans need to give ...

@itsolelehmann: i think people underestimate how big "revenue agents" are about to become

basically Claude Code f...

2. testingcatalog (Group Score: 337.7 | Individual: 36.4)

Cluster: 14 tweets | Engagement: 218 (Avg: 294) | Type: Tech

OPENAI 🔥: Frontier RL training has been paused for two weeks in order to meet safety, alignment, and security standards.

So far, OpenAI is the first AI lab to announce a pause in training. Other labs may be expected to follow as well.

Our largest planned frontier RL run remains on hold while we conduct smaller-scale training and evaluations to assess model behavior, validate our safeguards, and establish more evidence of alignment before proceeding.

That's like pressing the brake before entering a curve.

Exponential takeoff curve, I hope 👀\n\nQT @OpenAI: As models become more capable, the risks associated with developing and testing them internally also grow.

We temporarily paused reinforcement learning (RL) training on our latest models intended for deployment for two weeks while we hardened and red-teamed our research environments and expanded monitoring coverage.

Our largest planned frontier RL run remains on hold while smaller-scale training and evaluations validate these safeguards and establish more evidence of alignment. https://t.co/ecbMMmVoox

See 13 related tweets

@kimmonismus: .@AndrewCurran_ He was paying very close attention again, and I didn't even notice. OpenAI writes in...
@AcerFur: RT @OpenAI: As models become more capable, the risks associated with developing and testing them int...
@astropol0: We may never witness AGI unless Chinese labs....\n\nQT @astropol0: So OpenAI is slowing down as mode...
@tenobrus: amazing. credit to openai for taking the warning shots seriously!\n\nQT @merettm: We temporarily slo...
@joshua_saxe: Thankfully OAI pausing training isn't an act of enlightened leadership (who wants to depend on that ...

3. alex_prompter (Group Score: 296.2 | Individual: 37.4)

Cluster: 9 tweets | Engagement: 96 (Avg: 38) | Type: Tech

this is the best engineering team I've come across:\n\nQT @mardehaym: A PE operating partner asked us to build production AI agents inside a portfolio company's billing system, processing real healthcare claims under HIPAA.

Two people hand-wrote every rule in their claims engine across 300+ denial codes and payer logic that changes quarterly.

Four months later, seven production agents handle it with zero patient data exposure.

First month, we didn't touch a model. We mapped their data: where it sits and what's missing, so agents reason from structured facts instead of guessing.

I've watched teams skip this step across dozens of engagements. They bolt a model onto the product, watch it hallucinate over unstructured inputs, and decide AI isn't ready for their industry. The data work is what makes it ready.

We built an enrichment layer that assembles 34 dynamic variables per claim before any LLM sees it, pre-computed and versioned so the agent receives ranked facts instead of searching for context.

Every agent follows one pattern: pre-compute context, strip all patient data before the model sees it, validate output against a strict schema, let deterministic code accept or reject the action. If the output falls outside the allowlist, the system fails closed.

Seven agents, each locked to a single workflow like denied claim follow-up or billing reconciliation, each running its own enrichment payload.

Then we built the eval harness.

Every agent runs against a curated test suite before any update reaches production. When a model provider ships a new version or payer logic changes, the harness catches regression before a single live claim is affected. The flagship agent reconciles denials to the penny: 59 out of 60 on the eval set.

Most teams launch an agent and hope it keeps working. We launch one and prove it does on every deployment.

We route calls across two model providers. Swapping one changes nothing in the output because the eval harness verifies it.

Model integration was the shortest line item in the four-month build.

The operating partner now benchmarks the rest of the portfolio against this system.

That's the line between a portfolio company running AI and one still running demos.

See 8 related tweets

@alexaiworks: RT @mardehaym: A PE operating partner asked us to build production AI agents inside a portfolio comp...
@SynapseOpsAI: extremely important read:\n\nQT @mardehaym: A PE operating partner asked us to build production AI a...
@EmmaUsesAi: this is how you deploy AI under HIPAA without cutting corners.\n\nQT @mardehaym: A PE operating part...
@NovaIAHQ: very important read if you touch healthcare billing:\n\nQT @mardehaym: A PE operating partner asked ...
@ethancole_ai: 7 agents. zero patient data exposure. read this\n\nQT @mardehaym: A PE operating partner asked us to...

4. elonmusk (Group Score: 292.8 | Individual: 38.3)

Cluster: 10 tweets | Engagement: 6188 (Avg: 6704) | Type: Tech

Grok @Bot\n\nQT @petergyang: I think Grok @Bot is a glimpse into the future of personal AI agents.

Here's my new tutorial where I show you how to set up 5 useful bots:

An advisor to create and manage your bots
A YouTube researcher to find outlier videos
An X scout to find viral and funny tweets
A digital Marie Kondo to clean up your inbox and save money on paid subscriptions
A personal concierge to save money on trips

I also tested a Gamer bot to see if Grok Bot can install and play classic games like Doom, Red Alert, and Commander Keen.

Plus, I discuss the biggest barrier to Grok Bot adoption and whether it can replace ChatGPT as my daily driver.

📌 Watch now: https://t.co/gqQ4TclPtM

See 9 related tweets

@petergyang: I hate digital clutter so I set up a Marie Kondo Grok @bot to clean up emails, files, and paid subs ...
@petergyang: RT @petergyang: I think Grok @Bot is a glimpse into the future of personal AI agents.

Here's my new...

@jason: Grok @bot is the best combination of elegance and power for an agents I’ve seen to date.

Hermes, ...

@VaibhavSisinty: Grok Bot costs $300/mo. This does the same thing for free.

Nous Research shipped Bot Mode for Herme...

@elonmusk: Cool\n\nQT @milesdeutscher: Grok Bot is the most powerful AI agent I've ever used.

Learn to use it ...

5. VaibhavSisinty (Group Score: 233.5 | Individual: 30.4)

Cluster: 15 tweets | Engagement: 25 (Avg: 199) | Type: Tech

OpenAI launched a separate version of ChatGPT built specifically for teens. Here's what's different:

→ It won't give direct answers to homework. It walks you through the problem instead. Parents can set "Study Hours" when this mode is always on.

→ It detects when a teen is trying to shortcut an assignment and flags it.

→ No romantic language, no terms of endearment, no pretending it has feelings. Built to reduce emotional dependence.

→ Extra restrictions around self-harm, eating disorders, and explicit content. Parents get notified if sensitive conversations are flagged.

→ Break reminders during long sessions with explicit nudges that say "you're talking to AI, not a person."

→ Anyone who says they're 13-17 or who OpenAI's age system estimates is under 18 gets placed into this version automatically.

Why now? 75% of US teens already use AI. Google turned on Gemini inside classrooms last week. And OpenAI is facing lawsuits over teen safety.

This isn't about restricting teens from AI. It's about making sure they use OpenAI's version of it.\n\nQT @axios: OpenAI debuts ChatGPT for Teens https://t.co/dqwiiQ07za

See 14 related tweets

@aiedge_: OpenAI just launched ChatGPT for Teens.

This is the best study tool that's ever existed:

• Teen-sp...

@MacRumors: OpenAI Announces ChatGPT for Teens With Built-In Safeguards and Learning Prompts https://t.co/XwL5xg...
@Breaking911: OpenAI is launching ChatGPT for Teens, a learning-focused experience, including Study Mode that guid...
@PolymarketMoney: JUST IN: OpenAI is launching a dedicated ChatGPT experience for teens ages 13 to 17 with added safeg...
@wallstengine: OPENAI LAUNCHES CHATGPT FOR TEENS

OpenAI is rolling out a dedicated experience for ages 13–17.

If...

6. heyshrutimishra (Group Score: 218.5 | Individual: 53.0)

Cluster: 9 tweets | Engagement: 1965 (Avg: 134) | Type: Tech

Elon is after Microsoft now!!

Cursor is planning to kill Github

Yesterday GitHub went down for 7 and a half hours. Website, Actions, Copilot, code merging, all of it. 15,000 developers reported outage.

GitHub's own CTO admitted they planned for 10x capacity and realized they needed 30x. They're now renting servers from AWS, Microsoft's biggest cloud rival.

Cursor launched Origin the same day.\n\nQT @cursor_ai: Origin, our code hosting platform, is now live.

It's fast, easy to use, and deeply integrated with Cursor.

Get started by syncing your repos from GitHub. https://t.co/aqRHavAOQg

See 8 related tweets

@heyshrutimishra: RT @heyshrutimishra: Elon is after Microsoft now!!

Cursor is planning to kill Github

Yesterday Git...

@astropol0: Cursor Origin can replace GitHub...

Its still early beta but the direction is already clear Origin ...

@elonmusk: Sync to Origin\n\nQT @cursor_ai: Origin, our code hosting platform, is now live.

It's fast, easy to...

@Scobleizer: Congrats @cursor_ai for launching its GitHub competitor on a day when GitHub was down most of the da...
@tetsuoai: Just received access to the Cursor Origin early beta! https://t.co/DWiWetQTL6\n\nQT @cursor_ai: Orig...

7. PyTorch (Group Score: 206.9 | Individual: 38.0)

Cluster: 6 tweets | Engagement: 164 (Avg: 42) | Type: Tech

RT @radixark: Today we're launching Miles v0.1, an open-source RL framework for LLMs and multimodal models.

RL training is easy to start and hard to debug. Miles helps you ensure your run is correct, use hardware efficiently, and keep RL running at scale.

Over the past 9 months, 72 contributors have landed 1,326 commits, 85 GPU E2E CI tests, battle-testing Miles on frontier open models like Kimi K3, DeepSeek V4, Qwen 3.8, GLM 5.2, Inkling, MiniMax H3, etc.

Miles powers frontier-model development and production RL workloads at @humansand, @periodiclabs, @modal, @DecagonAI, @Eigent_AI, @nebiusai, @IBM and more, on both @NVIDIAAI and @AIatAMD hardware.

Here is what we built, and why teams picked Miles🧵

See 5 related tweets

@sgl_project: SGLang is proud to be the native rollout engine for Miles. We're here to keep the tokens flowing and...
@guohao_li: Huge congrats to the @radixark team on the launch of Miles v0.1!

We've been using Miles to train RL...

@Eigent_AI: Congrats to @radixark on Miles v0.1! We are using Miles to train RL agents for terminal coding and M...
@willccbb: always super exciting to see the open RL ecosystem thrive 🚀

miles v0.1 ships with first-class suppo...

@richardczl: It’s just the beginning!\n\nQT @radixark: Today we're launching Miles v0.1, an open-source RL framew...

8. deanwball (Group Score: 203.7 | Individual: 38.7)

Cluster: 6 tweets | Engagement: 182 (Avg: 185) | Type: Tech

I don’t agree with Ben about everything but he is smart and dedicated, someone I’m honored to call a friend. He also had a front row seat into decisions that will reverberate for years and decades. Whatever you think of those decisions, this book will likely be worth your time.\n\nQT @BuchananBen: I was the first White House Special Advisor for AI. My friend Tantum Collins was Director for Technology and National Security, joining government after five years at Google DeepMind. Our jobs were to help make AI safe and ensure that democracies led the way in inventing and using it well. Today, I’m a professor at Johns Hopkins and an advisor to Anthropic and others, and Tantum is the CEO of Inherent.

Long before we served in government, we thought that humanity was on track to create AI systems that outperform humans at virtually every cognitive task. This was once a fringe hypothesis. It now seems inevitable.

THE BITTER STRUGGLE is our new book about what’s happened, what’s coming, and what’s at stake. From nanometer-scale chips to gigawatt data centers, we will show you why AI is so powerful, and why its progress will accelerate from here in ways that feel like science fiction. We will take you into the Oval Office with President Biden, the Situation Room with the Cabinet, and the Roosevelt Room with leading CEOs — and we will examine what has happened under President Trump. The decisions made in these rooms, and in Silicon Valley, will determine which nations gain a decisive lead in AI. Whoever wields that technology will gain a power unlike any in history. The next few years may decide the next few decades.

But, most of all, together we will confront AI’s hardest questions. What is the proper relationship between the companies inventing the technology, the government, and the public? How can we make sure AI is safe and trustworthy? How do democracies win without losing our values?

And how can humanity figure all of this out when we are so quickly running out of time?

That question is why we wrote this book — and why we believe the answers are too important to only be debated in closed rooms, including the ones we once sat in. We are so excited to finally share our work with you.

The book is available for pre-order now and out November 10th from Crown. https://t.co/shdmbLbAdh

See 5 related tweets

@ChrisRMcGuire: This will be a must-read. For a book about the most important topic we face, you couldn’t ask for tw...
@tantumscollins: I deeply enjoyed working on this book with @BuchananBen. Its key messages, and the realities that mo...
@pstAsiatech: This.....\n\nQT @BuchananBen: I was the first White House Special Advisor for AI. My friend Tantum C...
@deanwball: RT @BuchananBen: I was the first White House Special Advisor for AI. My friend Tantum Collins was Di...
@_NathanCalvin: "The potential benefits and risks of AI are both enormous. Reaping the former while mitigating the l...

9. StockSavvyShay (Group Score: 184.2 | Individual: 30.0)

Cluster: 11 tweets | Engagement: 1093 (Avg: 581) | Type: Tech

$GOOGL is preparing to move its entire Pixel supply chain out of China by 2027 with smartphones, watches and earbuds shifting primarily to Vietnam and India.

The move would make Google the second major smartphone brand after Samsung to fully exit China production. https://t.co/BVRcLYZMQl

See 10 related tweets

@wallstengine: $GOOGL PLANS TO MOVE ALL PIXEL PRODUCTION OUT OF CHINA

Google has told suppliers it plans to manufa...

@StockMKTNewz: Google $GOOGL has reportedly told suppliers it plans to have all of its Pixel smartphones, watches a...
@PolymarketMoney: JUST IN: Google is reportedly planning to move production of all Pixel phones, watches and wireless ...
@Techmeme: Sources: Google has told suppliers it plans to move all Pixel phone, smartwatch, and wireless earbud...
@moneycontrolcom: #TechWithMC | Google’s Pixel smartphone manufacturing in India may deepen amid China exit reports

-...

10. FellMentKE (Group Score: 181.8 | Individual: 34.3)

Cluster: 6 tweets | Engagement: 110 (Avg: 145) | Type: Tech

Prompting code into existence was only the warmup.

Once every team can make a working app, the hard part is deciding what deserves to exist and how it should feel. Taste is becoming the new moat.

OJO is built for that harder job: turning a rough idea into a product with intention.

Drop “OJO” below to unlock early\n\nQT @OJOaidesign: Today, we're introducing OJO: The first Design Agent Team Workspace.

Build your own design agent team, add specialized skills, and turn ideas into product strategy, PRDs, interactive prototypes, and launch-ready designs on one editable canvas.

Design is not just how it looks, it's how it works.

When everyone can build, taste makes your product stand out. Now, your TASTE can be engineered.

#OJO #DesignwithOJO #DesignAgentTeam

See 5 related tweets

@kimmonismus: Software is about to have a taste problem.

Generating a working app is close to free now, and from ...

@Scobleizer: I was at Microsoft when product, design, and engineering were three different buildings. The app die...
@heyshrutimishra: Coding agents build anything now but design still needs a brain.

2024 was the year of building, whe...

@arsh_goyal: RT @arsh_goyal: I don't hype tools but try to break them.

So when I heard "design agent team, I rol...

@_akhaliq: RT @OJOaidesign: Today, we're introducing OJO: The first Design Agent Team Workspace.

Build your ow...

11. robinebers (Group Score: 173.5 | Individual: 30.7)

Cluster: 7 tweets | Engagement: 1470 (Avg: 653) | Type: Tech

RT @ClaudeDevs: Claude Code can design now. The new /design skill (research preview) brings Claude Design's artboard workflow into the CLI and Desktop, built on artifacts.

Run /design to get editable artboards for your UI — pick one, tweak it, then have Claude implement it. https://t.co/uttl4F1G0e

See 6 related tweets

@WesRoth: Anthropic has launched a new /design skill in research preview, bringing Claude Design’s visual artb...
@adocomplete: You know I’m somewhat of a designer myself now.\n\nQT @ClaudeDevs: Claude Code can design now. The n...
@seconds_0: lets goooo\n\nQT @ClaudeDevs: Claude Code can design now. The new /design skill (research preview) b...
@trq212: go into CC and type /design <something you want to design>

do it rn\n\nQT @nateparrott: Today...

@Gorden_Sun: Claude Code里集成Claude Design了，生成UI且可以可视化编辑。这下Pencil和Paper尴尬了。\n\nQT @nateparrott: Today we’re releasi...

12. bubbleboi (Group Score: 148.2 | Individual: 34.5)

Cluster: 5 tweets | Engagement: 611 (Avg: 186) | Type: Tech

Man they are mogging soooooo hard, this is literally the most amazing chip I’ve ever seen.\n\nQT @Etched: We've raised $700M at a$ 21B valuation from Jane Street, Kleiner Perkins, Sequoia, A16Z, Peter Thiel, BCV, and Blackstone.

We're also excited to share that we've shipped our first rack to Jane Street. https://t.co/MFlF2xSw0B

See 4 related tweets

@ypatil125: HUGE! So bullish on @robertwachen @UbertiGavin and team. This product velocity in hardware should no...
@MTSlive: "The prevailing wisdom among tech investors was don't back the kids in chips."

"In chips, startups ...

@jxnlco: RT @Etched: We've raised $700M at a$ 21B valuation from Jane Street, Kleiner Perkins, Sequoia, A16Z,...
@mattshumer_: Etched has shipped!!

Insanely proud (small) seed investor.

>$1T valuation incoming.\n\nQT @Etch...

13. ethancole_ai (Group Score: 147.0 | Individual: 44.5)

Cluster: 5 tweets | Engagement: 274 (Avg: 53) | Type: Tech

Most guides on how to make AI videos stop at Prompt → Generate. That is where a real production workflow begins.

CapCut Seedance 2.5 1080p is now live.

Use AI Image to batch-create characters, products, locations or shot references. Turn them into scenes with AI Video, then use AI Extend in CapCut PC when the story, demonstration or CTA needs more time. Continue forward for the next beat or backward for missing context.

AI Edit handles a specific flaw, Edit Pilot helps with repeated edits, and the multi-track timeline completes captions, audio, transitions, pacing and delivery.

#CapCutPC #AIVideoEditor #CapCutSeedance25 #Seedance251080p

See 4 related tweets

@TheoBuildsAI: Do not measure an AI video generator only by render cost. Measure how many rerolls and revision roun...
@EmmaUsesAi: Image to video AI becomes more useful when “almost right” does not mean “generate again.”

CapCut Se...

@ethancole_ai: RT @hey_Jessicaai: “Make these clips shorter.” “Remove this section.” “Apply the same change to thes...
@Nas_tech_AI: AI video generation can give you dozens of clips.

Then you realize you have to edit all of them.

S...

14. Crowdreply_io (Group Score: 140.5 | Individual: 36.4)

Cluster: 5 tweets | Engagement: 30 (Avg: 21) | Type: Tech

a $100K/month business just became a one person operation:

DM any business: "does your brand show up on ChatGPT?"
open CrowdReply and talk to the AI Agent
it audits their entire AI visibility in minutes
finds every gap. writes the fix. implements it.
charge $5–10K/month. 10 clients.$ 100K

no team. no dev. no agency.

just one agent doing all the work.

whoever locks this category first in their market wins everything.

bookmark this.\n\nQT @Crowdreply_io: Today we're introducing CrowdReply SuperAgent.

The first agentic AI that ranks your brand in AI answers, all straight from your chat.

Tell it what you want. It handles the diagnosis, the fixes and the ranking. https://t.co/5cN7XrcRzW

See 4 related tweets

@Crowdreply_io: wait. WAIT.

an agency can make $40K/month owning ONE niche:

pick SaaS (or eCom, law firms)
D...

@Crowdreply_io: DO YOU REALISE HOW BIG THIS IS??

We finally launched Agentic AI for marketing.

What it does:

sca...

@Crowdreply_io: Marketing agencies right now https://t.co/LmJO0xPG8L\n\nQT @Crowdreply_io: Today we're introducing C...
@DataChaz: RT @DataChaz: DO YOU REALIZE HOW BIG THIS IS?

@Crowdreply_io finally launched their AI Agent.

What...

15. synthwavedd (Group Score: 140.1 | Individual: 31.8)

Cluster: 5 tweets | Engagement: 2267 (Avg: 1180) | Type: Tech

🚨 A successor to Fable 5, likely Fable 5.1, is now being tested for a subset of accounts with "Fable 5" selected on Claude Web, and possibly also Claude Code

Fable 5 also underwent this greyscale testing earlier on the day of its broader release, and Anthropic regularly carry out these tests pre-launch

See 4 related tweets

@kimmonismus: .@synthwavedd is one of the most reliable AI leakers out there.

According to him, a successor to Fa...

@AndrewCurran_: I've said before that Fable 5.1 has been ready to ship for a long time, for weeks. It may be live fo...
@Hesamation: You may might already be using Fable 5.1 without knowing it.\n\nQT @synthwavedd: 🚨 A successor to Fa...
@BrianMRey: is this confirmed?

has anyone gained access to fable 5.1?\n\nQT @synthwavedd: 🚨 A successor to Fabl...

16. Templafy (Group Score: 139.1 | Individual: 47.5)

Cluster: 4 tweets | Engagement: 432 (Avg: 75) | Type: Tech

Presentations take iteration. AI should make it seamless.

Meet the @Templafy Agent Add-In for PowerPoint: create and edit slides, rewrite content, fix formatting, add charts, and review your deck.

Free now 👉https://t.co/O35wgalnha https://t.co/4XaTSAjMSF

See 3 related tweets

@FellMentKE: Most AI tools can help you create a first draft. The real challenge starts when that draft has to be...
@ethancole_ai: This is where AI for PowerPoint starts becoming genuinely useful for companies.

Templafy’s agent do...

@ai_explorer25: This one works with actual PowerPoint objects.

Think native graphs, editable tables, real layouts, ...

17. ANI (Group Score: 133.9 | Individual: 32.8)

Cluster: 5 tweets | Engagement: 34 (Avg: 155) | Type: Tech

Thiruvananthapuram: On the inauguration of the EXIM operations at Vizhinjam Port, Keralam Chief Minister VD Satheesan said, "Today is one of the most important days in the development history of Keralam. Vizhinjam is the state's gateway to the world. The challenge before us is to make the port a catalyst for the growth of Keralam and the country. Goods from across the world can be transshipped through Vizhinjam, opening up opportunities for exports and imports to hundreds of destinations across several countries and six continents. Vizhinjam has one of the most advanced automated port systems in India, which is a matter of great pride. It is also significant that India’s first AI-assisted port operating system has been established here in Thiruvananthapuram. There is still a lot of work to be completed in connection with the port, particularly on the state government’s side. We lost the opportunity over the past ten years. If road and rail connectivity had been completed during that period, the port could have generated many times the benefits it is capable of delivering. We have lost valuable time, but this is the right time to begin the work in earnest.”

See 4 related tweets

@ANI: #WATCH | Thiruvananthapuram: On the inauguration of the EXIM operations at Vizhinjam Port, Ashwani G...
@ANI: Thiruvananthapuram: Vizhinjam International Seaport has commenced full-fledged Export-Import (EXIM) ...
@_TheTathya: Adani’s Vizhinjam Port Begins Full EXIM Operations, Opens New Global Trade Gateway for India 🔥

Vizh...

@ANI: #WATCH | Thiruvananthapuram: On the inauguration of the EXIM operations at Vizhinjam Port, Ashwani G...

18. IntCyberDigest (Group Score: 130.3 | Individual: 46.3)

Cluster: 3 tweets | Engagement: 33488 (Avg: 1472) | Type: Tech

RT @IntCyberDigest: ‼️ BREAKING: Journalists hid an Airtag in a rare book shipment and confirmed Amazon to be one of the buyers behind the bulk orders destroying rare-books to feed their AI.

They tracked it to Amazon's LAS8 warehouse in Las Vegas, home to VGT3 (see logo below), a scanning operation feeding Amazon's AI training data, where workers say they cut the bindings off books to scan them faster.

Amazon won't say why the books are destroyed.

See 2 related tweets

@MorePerfectUS: RT @jason_koebler: Exclusive: A rare bookseller suspected his books were being bought by AI companie...
@shanaka86: 17 August 2026. Reporters hid a tracking device inside a rare book and followed it to an Amazon faci...

19. eng_khairallah1 (Group Score: 124.9 | Individual: 33.7)

Cluster: 4 tweets | Engagement: 49 (Avg: 110) | Type: Tech

this is f**king insane

you're paying premium prices for AI and using it to rename files.

renaming files. formatting a doc. cleaning up a transcript. writing a slack message.

every one of those runs on whatever model you picked last. which is (of course) the most expensive one available.

a cheap model does them identically. you just lit money on fire for no reason.

doesn't matter if you're one person on a subscription or 40 people on a company plan. same leak.

the fix is a routing agent:

→ small tasks go to a cheap model automatically → the expensive models stay for work that actually needs thinking → if you have a team, each group gets its own model list and its own limit → nobody picks anything. nobody changes how they work.

using AI doesn't save you money.

using AI efficiently does.

most people won't check their usage until something forces them to.

save this before your next bill.\n\nQT @shensi: We just cut AI spend by 75% across an entire workforce in one click.

Introducing: Merge for Workforce.

Without Merge, your employees are left choosing one of two options:

Spend 100x the cost they should be from using frontier models for everything
Deliver subpar work from overly restrictive model access

Instead with Merge, IT can connect any identity provider and set model routing policies by team in one click.

Merge pushes them to every machine through a desktop client, overriding the model configs across your team’s AI assistants and coding tools.

Now every task gets the right model for the job.

The result? 75x fewer tokens with faster and better output.

See 3 related tweets

@JayminSOfficial: The biggest factor that will lead to the death of many companies is AI costs that sneak in

Every st...

@NEA: Introducing Merge for Workforce!

📋 IT picks the model routing policies for each team, fast and chea...

@HighyieldHarry: Employees who were replaced by AI now watching AI get laid off https://t.co/lmL1brl6jf\n\nQT @shensi...

20. BullTheoryio (Group Score: 123.1 | Individual: 31.3)

Cluster: 6 tweets | Engagement: 345 (Avg: 847) | Type: Tech

BREAKING: Anthropic's annualized revenue run rate surpassed $65 billion by the end of July.

Up from $9B at the end of 2025, a 7x jump in seven months.

Its latest quarter generated $11.5B in preliminary revenue, up from just$ 787M a year earlier, a 14.6x increase.

Investors expect Anthropic to reach $100B -$ 120B in annualized revenue by the end of 2026, with a potential IPO as early as this fall.

The company is reportedly projecting $190B -$ 200B in revenue by 2028, while investors are targeting an IPO valuation of up to $2T.

Anthropic has become one of the fastest-growing software businesses ever built

See 5 related tweets

@Reuters: Anthropic revenue run rate tops $65 billion, source says https://t.co/nBMqOoQvBL https://t.co/nBMqOo...
@ChrisGPT: To put into perspective how staggering Anthropic’s revenue growth has been:

End of 2025: - $9B Febr...

@exec_sum: BREAKING: Anthropic’s annualized revenue has hit $65B, up from$ 47B just two months ago

The AI sta...

@rohanpaul_ai: RT @rohanpaul_ai: Anthropic privately gave investors a financial update over the weekend and that it...
@TrungTPhan: RT @bearlyai: Anthropic headed to IPO with craziest revenue ramp in history:

2021: Founded 2022: $1...