今日科技热推 - 2026年2月24日

2026年2月24日科技每日简报

Today's top tech conversations are led by @rohanpaul_ai, whose post about 'OpenAI and SoftBank disagreed ...' garnered the highest engagement. Key themes trending across the top stories include comic, artist, their, openai, agents. The community is actively discussing recent developments in AI, engineering practices, and startup strategies.

1. rohanpaul_ai (Group Score: 93.5 | Individual: 35.9)

Cluster: 3 tweets | Engagement: 179 (Avg: 111) | Type: Tech

OpenAI and SoftBank disagreed over how to develop data centers together.

Back in early 2025, the massive $500B joint venture Stargate was announced at the White House. The plan was for Oracle and SoftBank to build out 10GW of AI computing power, but more than a year later, not much is happening.

Disagreements over leadership, structure, and financial responsibility have left the partnership gridlocked.

See 2 related tweets

@aakashgupta: The $500 billion number was always the distraction. The constraint was always execution.

Stargate L...

@kimmonismus: This is a huge step backwards for OpenAI:

OpenAI’s ambitious $500 billion Stargate data center vent...

2. rseroter (Group Score: 84.8 | Individual: 48.3)

Cluster: 3 tweets | Engagement: 6334 (Avg: 165) | Type: Tech

RT @AnthropicAI: We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax.

These labs cr…

See 2 related tweets

@testingcatalog: BREAKING 🚨: Anthropic vs China!

Anthropic has reported on the industry scale distillation from Chin...

@BlindGoose1337: RT @kimmonismus: Holy sh*t: Anthropic says it uncovered large-scale “distillation” campaigns by Deep...

3. rohanpaul_ai (Group Score: 75.0 | Individual: 27.2)

Cluster: 3 tweets | Engagement: 45 (Avg: 111) | Type: Tech

🚨 BREAKING: Anthropic is accusing some Chinese AI labs of creating 24,000 fake accounts and running 16 million prompts to boost their own models.

They say, this extraction relies on distillation (which are technically legitimate methods), where a smaller AI learns by studying the high-quality answers of a smarter AI.

Anthropic is claiming, to avoid getting blocked, these labs built sprawling proxy networks called hydra clusters. These clusters constantly shuffle requests across thousands of fake profiles to blend in with normal traffic.

See 2 related tweets

@LiorOnAI: Anthropic just exposed the real vulnerability in AI: it's not the models, it's the training data pip...
@chatgpt21: 🚨Breaking:

Anthropic has publicly identified three Chinese AI laboratories DeepSeek, Moonshot, and ...

4. rohanpaul_ai (Group Score: 72.0 | Individual: 35.7)

Cluster: 3 tweets | Engagement: 158 (Avg: 111) | Type: Tech

CNBC: OpenAI is teaming up with Accenture, Boston Consulting Group, Capgemini, and McKinsey to solve the "last mile" problem of getting AI into daily business tasks.

These consulting giants will act as the bridge between OpenAI’s raw technology and the complex, messy systems that large companies use every day.

The focus is on a new platform called Frontier, which acts as a central brain that connects all the different data and software within a single organization.

This setup makes it much easier to create AI agents. OpenAI is even sending its own engineers to work directly inside these partner firms to make sure the tech is set up correctly and safely.

The business side of this is huge because corporate clients already make up 40% of OpenAI’s revenue, and they expect that to hit 50% by the end of 2026.

By letting consultants handle the strategy and setup, OpenAI can scale its technology to thousands of companies without needing to hire an infinite number of staff.

cnbc. com/2026/02/23/open-ai-consulting-accenture-boston-capgemini-mckinsey-frontier.html

See 2 related tweets

@kimmonismus: OpenAI expands enterprise

OpenAI is rolling out Frontier Alliances, teaming up with BCG, McKinsey, ...

@Reuters: OpenAI deepens partnerships with consulting giants to push enterprise AI beyond pilot https://t.co/1...

5. llama_index (Group Score: 70.6 | Individual: 27.0)

Cluster: 3 tweets | Engagement: 19 (Avg: 32) | Type: Tech

🚀 LlamaAgents Builder just leveled up: File uploads are here!

Our natural language interface for building agentic document workflows now supports file uploads.

You can provide example documents as context, and the agent will use them as a starting point to design and tailor your workflow. The result? Applications that better match your real-world use case.

The more representative your sample files, the more accurate your final app.

🎥 Watch the full walkthrough: https://t.co/LQW2PEZ8d9 🦙 Get started with LlamaCloud: https://t.co/wZjhFV29gN

See 2 related tweets

@itsclelia: We’ve been cooking at @llama_index🍳

Almost a month ago, we launched LlamaAgents Builder, a natural ...

@jerryjliu0: RT @llama_index: 🚀 LlamaAgents Builder just leveled up: File uploads are here!

Our natural language...

6. gdgtify (Group Score: 68.6 | Individual: 35.1)

Cluster: 2 tweets | Engagement: 43 (Avg: 24) | Type: Tech

Here is a fun Nano Banana prompt for comic book artists. I have included a couple of examples in the prompt. Animated with Kling

Prompt:

Input Variable: [comic book artist and seires] (e.g., Alan Moore, Stan Lee, Frank Miller, Osamu Tezuka)

System Instruction: Generate a hyper-detailed, macro 3D diorama of a "Comic Artist's Visual Workspace." Use the following logic to procedurally fill the scene with maximum density:

Profile: Analyze the Input: Identify the Artist/Series, their Art Style, their Era, and their Signature Work The Vibe: Identify the atmosphere of their work (e.g., Moore = Complex/Metatextual; Lee = Heroic/Dynamic; Miller = Gritty/Dark; Tezuka = Manga/Expressive) The Artifacts: Identify 5-7 iconic items from their comics (e.g., Smiley face badge, Spider-Man's mask, Batman's cowl, Astro Boy's rockets) Container: The Base: A massive comic book lying flat Texture: The comic book texture must match the genre (e.g., Glossy modern, Newsprint vintage, Matte indie) The Title: The comic cover features the artist's most famous title in comic-appropriate typography Workspace: The Desk: A period-accurate artist drawing table sits on top of the comic book The Artist: A 1:12 scale figure of the Comic Artist sits at the drawing table Pose: They are in a "Creative Burst" state of creation—inking, penciling, or lettering The Clutter (Density Layer 1): The drawing table is covered in: Sketch pages and reference materials Period-specific art supplies (e.g., India ink, brushes, nib pens, digital tablet) Comic book reference materials and inspiration Visual: CRITICAL: The comic world is physically invading the drawing table The "Micro-Population": Tiny, microscopic figures (1:87 scale) of the Comic Characters are interacting with the art supplies (e.g., Rorschach's mask forming in an ink blot; Spider-Man swinging from the lamp; Watchmen characters climbing the page layouts) The "Panel Bleed": The environments of the comics are overgrowing the drawing table (e.g., Gotham's rooftops forming around the ink jars; Metropolis's skyline appearing in the background; Manga speed lines affecting the pencil strokes) The Artifacts: The "Artifacts" (Step 1) are used as functional art objects (e.g., The Smiley badge as a paperweight; Captain America's shield as a palette; Wolverine's claws as lettering guides) Lighting & Atmosphere: Primary Light: An artist's drafting lamp (Bright white) illuminating the drawing area Secondary Light: Comic energy coming from the pages or characters (e.g., The glow of a power effect, the dramatic shadow of a villain, the bright colors of a superhero) Scale: Macro photography with a Tilt-Shift effect to emphasize the tiny scale of the characters vs. the "Giant" artist Output: ONE image, 3:2 Aspect Ratio, Octane Render, "Artist Studio" level of detail, Graphic/Dynamic Aesthetic

See 1 related tweets

@gdgtify: A fun Nano Banana prompt for visualizing visionary writers and futurists. I am still working to make...

7. business (Group Score: 60.8 | Individual: 36.5)

Cluster: 2 tweets | Engagement: 213 (Avg: 97) | Type: Tech

IBM shares plunged on Monday, after the AI startup Anthropic said its Claude Code tool can help with modernizing COBOL, a dated programming language that’s mainly run on IBM computers https://t.co/eJwLsjWSRe

See 1 related tweets

@Forbes: IBM Shares Plummet 13%—Worst Day Since 2000—After Anthropic Launches Programming AI Tool https://t.c...

8. business (Group Score: 56.6 | Individual: 30.4)

Cluster: 2 tweets | Engagement: 89 (Avg: 97) | Type: Tech

The India AI Impact Summit has laid out an ambitious roadmap for 2026. But can Modi make the nation truly a global AI power? Join Bloomberg reporters for a Live Q&A on February 23, 2026 at 2pm IST | 4:30pm HK https://t.co/OqVdpvBuwk https://t.co/Enr63tgcIj

See 1 related tweets

@business: 🚨 LIVE NOW: The India AI Impact Summit has laid out an ambitious roadmap for 2026. But can Modi make...

9. kimmonismus (Group Score: 54.7 | Individual: 34.2)

Cluster: 2 tweets | Engagement: 251 (Avg: 478) | Type: Tech

Read this blogpost! But above all, read it as a starting point for discussion, so that we can finally have a broad and joint conversation about the near future and search for solutions together!

Citrini Research sketches a "future macro memo" scenario where rapidly improving, ever-cheaper AI agents trigger a self-reinforcing loop, companies replace white-collar labor, wages and consumption collapse, “Ghost GDP” rises, intermediation moats evaporate, private credit and eventually prime mortgages come under stress - culminating in a systemic market drawdown and political crisis over how to redistribute gains from abundant machine intelligence.

"That summer, we spoke with a procurement manager at a Fortune 500. He told us about one of his budget negotiations. The salesperson had expected to run the same playbook as last year: a 5% annual price increase, the standard “your team depends on us” pitch. The procurement manager told him he’d been in conversations with OpenAI about having their “forward deployed engineers” use AI tools to replace the vendor entirely. They renewed at a 30% discount. That was a good outcome, he said. The “long-tail of SaaS”, like Monday. com, Zapier and Asana, had it much worse."

It starts with white-collar workers coming under pressure from AI agents, and wages beginning to fall. Why pay a human the same amount as before when an AI agent is more productive and works longer hours?

"By early 2027, LLM usage had become default. People were using AI agents who didn’t even know what an AI agent was, in the same way people who never learned what “cloud computing” was used streaming services. They thought of it the same way they thought of autocomplete or spell-check - a thing their phone just did now."

It soon becomes clear to everyone that AI agents are here to stay. They will become as commonplace as the internet.

"Our January 2027 macro memo argued this was the wrong mental model. The US economy is a white-collar services economy. White-collar workers represented 50% of employment and drove roughly 75% of discretionary consumer spending. The businesses and jobs that AI was chewing up were not tangential to the US economy, they were the US economy. “Technological innovation destroys jobs and then creates even more”. This was the most popular and convincing counter-argument at the time. It was popular and convincing because it’d been right for two centuries. Even if we couldn’t conceive of what the future jobs would be, they would surely arrive."

In conjunction with the financial and capital markets, this will become a systemic challenge. However, what has always been criticized from all sides is now becoming reality: AI will cause mass layoffs. Yes, it will also create a few new jobs, but overall, it will lead to more unemployment. Or to put it another way: there will be no solution for how to balance the few new jobs against the mass unemployment.

"Every new job, however, required a human to perform it.

AI is now a general intelligence that improves at the very tasks humans would redeploy to. Displaced coders cannot simply move to “AI management” because AI is already capable of that.

Today, AI agents handle many-weeks-long research and development tasks. The exponential steamrolled our conceptions of what was possible, even though every year Wharton professors tried to fit the data to a new sigmoid."

These are concrete problems that are already visible. This blog post, however, systematically addresses them. So far, there is no solution to the problem.

It is up to us to work on solutions.

See 1 related tweets

@BusinessInsider: Market research firm Citrini Research envisions a scenario where AI's continued expansion harms the ...

10. lateinteraction (Group Score: 51.4 | Individual: 51.4)

Cluster: 1 tweets | Engagement: 658 (Avg: 36) | Type: Tech

RT @RyanWatkins_: “Dude I have 10 agents running while I sleep. No one is prepared for AGI in 2 years man.”

“So what are you building?”

“…

11. cryptopunk7213 (Group Score: 50.6 | Individual: 50.6)

Cluster: 1 tweets | Engagement: 4043 (Avg: 645) | Type: Tech

LMFAO be anthropic:

wake up, make coffee scroll boomer companies worth billions of dollars pick one "hi claude make a better version of this make no mistakes ty" wait 1 hour "looks good" push to prod post tweet about how you just displaced boomer company with new feature

profit

12. AnthropicAI (Group Score: 50.2 | Individual: 25.2)

Cluster: 2 tweets | Engagement: 2882 (Avg: 5733) | Type: Tech

New research: The AI Fluency Index.

We tracked 11 behaviors across thousands of https://t.co/RxKnLNNcNR conversations—for example, how often people iterate and refine their work with Claude—to measure how well people collaborate with AI.

13. rohanpaul_ai (Group Score: 48.8 | Individual: 21.1)

Cluster: 3 tweets | Engagement: 212 (Avg: 111) | Type: Tech

Big revelation in this data.

Everyone keeps talking about an AI bubble, but we forget the fact that only 0.3% of the global population actually pays for a premium subscription.

For the vast majority of the real world it has not started yet.

6.8B people have yet to interact with even 1 free chatbot.

The massive grey chunk is the 6.8B people who have zero experience with AI. The green dots represent 1.3B people using free versions of tools.

Only a small group of 15-25 mn people pays for subscriptions. We sit in that microscopic slice.

See 2 related tweets

@minchoi: This chart is wild..

Out of 8.1 billion people

84% (~6.8B) Never used AI 16% (~1.3B) Free AI chatb...

@minchoi: RT @minchoi: This chart is wild..

Out of 8.1 billion people

84% (~6.8B) Never used AI 16% (~1.3B) ...

14. jerryjliu0 (Group Score: 48.6 | Individual: 32.7)

Cluster: 2 tweets | Engagement: 32 (Avg: 52) | Type: Tech

We built an AI agent that lets you vibe-code document extraction - high accuracy and citations over the most complex documents.

Our latest release lets you upload documents as context. All you then have to do is describe what you want extracted in natural language. 💡 Our agent will then read the document with file tools to infer the right schema, validation rules, and other pre/postprocessing logic. ✅ It will give you back a workflow that can extract over thousands/millions of documents at scale. You can still of course review and edit every output before approving.

Stop handling paperwork manually; just upload files, describe your task, and let our agent handle the rest.

Our vision for LlamaAgents is to provide the most advanced and easy-to-use way for you to orchestrate document work.

Walkthrough: https://t.co/dAtzlZbot4 Check it out: https://t.co/XYZmx5TFz8 If you’re interested in reducing the operational burden of document extraction (invoices, claims, onboarding forms), come talk to us! https://t.co/Ht5jwxSrQB

See 1 related tweets

@itsclelia: RT @jerryjliu0: We built an AI agent that lets you vibe-code document extraction - high accuracy and...

15. Prince_Canuma (Group Score: 48.6 | Individual: 26.7)

Cluster: 2 tweets | Engagement: 56 (Avg: 45) | Type: Tech

One of the reason I built MLX-Audio is because my Dad become blind 6 years ago and I promised him I’ll help him resume his favourite hobby of reading books using AI.

With the new MLX-Audio-Swift he can read the news on his iPhone :)

Hope it empowers you too @tayarndt!

See 1 related tweets

@Prince_Canuma: MLX-Audio-Swift v0.1.0 is here 🔥

The first native Swift SDK for on-device audio AI on Apple Silicon...

16. BrianRoemmele (Group Score: 48.1 | Individual: 24.4)

Cluster: 2 tweets | Engagement: 73 (Avg: 364) | Type: Tech

TIME WARS—NOT SCIENCE FICTION.

Folks are not ready for some of what my optimized AI has predicted as a possibility with foundations supporting it over the next 5000 days.

Get acquainted with “I would never have thought that”:

Temporal Desynchronization Crises:

AI-optimized personal timelines (e.g., via time-dilation simulations or productivity drugs) cause people to experience time at different rates, fracturing social cohesion.

Trigger: Workers use AI to “stretch” days for more output, but it alters perception. Escalation: Families desync – one ages “faster” subjectively, leading to relational breakdowns.

Climax: Global “time wars” where fast-livers dominate economies, marginalizing slow-adopters. Resolution: Standardized “time equity” protocols, but cultural rifts linger. Tied to The Lathe of Heaven by Ursula K. Le Guin (reality-altering dreams, analogous to time manipulation) (https://t.co/xtmnW0qGBv).

Likelihood: 2/10 – Quantum computing and neurotech could enable this, a blind spot in current discussions.

See 1 related tweets

@BrianRoemmele: Temporal Desynchronization Crises: Wilder: AI-optimized personal timelines (e.g., via time-dilation ...

17. IamEmily2050 (Group Score: 47.7 | Individual: 24.3)

Cluster: 2 tweets | Engagement: 62 (Avg: 756) | Type: Tech

Same Natural language prompt with Gemini 3.1 Pro High on Google AI Studio but I used my System prompt for SVG and I can even make it better than this, the model is very capable but you need to know how to use it. https://t.co/UcXJ0fIPPV

See 1 related tweets

@IamEmily2050: Same Natural language prompt with Gemini 3.1 Pro High only Google AI Studio but I used my System pro...

18. cryptopunk7213 (Group Score: 47.4 | Individual: 47.4)

Cluster: 1 tweets | Engagement: 2983 (Avg: 645) | Type: Tech

yo so just to recap the week:

google released gemini 3.1 but it disappointed tf out of people who claimed its not as good as they advertise
BUT THEN google dropped an AI song generation model which kinda slaps (Lyria 3) AND a new feature that 1-shots product photoshoots putting photographers (and fashion models) out of a job (Google Pomelli)
Microsoft pioneered a way to STORE DATA ON GLASS - stores data for up to 10,000 fucking years (🤯) and way cheaper than using silica.
Taalus fused an ai model DIRECTLY into silicon processing 17,000 tokens PER SECOND which is 17X faster than openai’s quickest model but 10 TIMES CHEAPER than nvidia blackwell gpus. (smaller models tho)
Apple is releasing 3 new AI devices: glasses, airpods with cameras and a pendant.
anthropic wiped $10B off the stock market after launching an ai that finds and help fixes security flaws in your code. cybersecurity stocks dumped <1 hour after launch.
BUT THEN Pentagon threatened to blacklist anthropic for failing to provide them uncensored access to claude.
xAI dropped grok 4.20 spawning 16 specialised ai agents that work together to solve any prompt. the model self-improves EVERY WEEK.
scientists discovered dna in meteorites suggesting humans could’ve originated from outer space (what)
Meta dropped Manus Agents going head-to-head with OpenClaw.
a cardiologist won 3rd-place in anthropic’s hackathon after vibe coding an app that helps his patients understand and act on their medical results. fucking hero.
china announced a cure for diabetes 1 & 2 after creating stem cells that make insulin
Moonshot labs is raising $500M immediately after raising$ 700M LAST WEEK.
ex-Deepmind researcher raised europes LARGEST ever seed round $1B at$ 4B back right as europe announces $13B new fund to catch up with america.
Stripe’s ai agents autonomously merged 1300 pull requests (up 30% from last week) suggesting top companies are now recursively improving using AI.

dear f*cking goodness GOODNIGHT

19. seraleev (Group Score: 47.3 | Individual: 47.3)

Cluster: 1 tweets | Engagement: 481 (Avg: 51) | Type: Tech

Since Nov 13, 2025, Apple updated the App Store Review Guidelines (5.1).

Developers now must clearly disclose where user data is sent, including any third-party AI services.

The key part → explicit user consent. If a user declines, AI features must stay disabled.

Many apps are getting rejected because this consent screen is missing. Worth auditing your onboarding and updating flows before your next submission.

Sharing a few examples of how I implemented this in my apps (+ one from a friend).

20. 0xdevshah (Group Score: 46.5 | Individual: 46.5)

Cluster: 1 tweets | Engagement: 514 (Avg: 23) | Type: Tech

RT @ns123abc: 🚨BREAKING: Anthropic Catches Chinese Labs Stealing Claude at Scale - Largest AI Heist in History!

Three Chinese labs ran coo…