2026年2月11日技术推特要闻精选

今日科技要闻：智能体 AI（Agentic AI）正展现出强劲的发展势头。Alphabet 斥资数十亿加强基础设施建设，GitHub 前首席执行官 Nat Friedman 则为其智能体优先的开发平台筹得 6000 万美元的创纪录融资。在行业争议方面，OpenAI 的商业模式与 Ring 的 AI 监控引发广泛讨论，而软件板块正因 AI 对税务领域的颠覆性影响而产生波动。工程突破上，Pydantic 推出的低延迟沙箱正着力解决智能体性能瓶颈，但业界对于日益复杂的多智能体系统在实际协作价值与成本效益方面仍持审慎态度。

1. aakashgupta (Group Score: 88.8 | Individual: 32.8)

Cluster: 3 tweets | Engagement: 84 (Avg: 438) | Type: Tech

OpenAI just told you their real business model and nobody’s connecting the dots.

Codex launched last week. Ads launched today. Codex is the product they want developers to love. ChatGPT is the product they need to monetize.

The timing tells the whole story. On February 2, they launched Codex for macOS and temporarily opened it to free users. On February 5, they dropped GPT-5.3-Codex. On February 9, today, they flipped the switch on ChatGPT ads. Three moves in eight days.

Here’s what this reveals about OpenAI’s cost structure. They inked over $1.4 trillion in infrastructure deals in 2025, and Sam Altman told employees they’re “back to exceeding 10% monthly growth.” Revenue is growing. But so is compute spending. And compute spending is growing faster.

Codex grew 50% in just the last week according to Altman. Claude Code hit $1 billion in annualized revenue within 6 months of launch. The developer tools market is real. But developer tools require giving away the expensive models. GPT-5.3-Codex is their most capable model running complex multi-hour tasks in cloud sandboxes. Every Codex session burns serious compute.

So who pays for it? The free ChatGPT user seeing a https://t.co/aSk9wnnvrs ad after asking about Santa Fe hotels. The Go subscriber at $8/month watching sponsored grocery delivery placements. That’s the subsidy. Ad revenue from casual users funds compute for power users running parallel agents on their codebase.

Anthropic saw this exact dynamic and ran a Super Bowl ad saying “ads are coming to AI” while promising Claude stays ad-free. Altman called it “clearly dishonest.” But the move tells you Anthropic identified the positioning gap and is sprinting through it.

The real question for developers: when ad revenue becomes 30-40% of OpenAI’s business, does the model optimization start favoring engagement over capability? Google made that trade with search. Meta made it with feed ranking. The incentive structure bends the product toward the revenue source, always.

Codex is genuinely good. But the business model powering it is now structurally dependent on a user base that generates ad impressions, and that dependency will shape every product decision OpenAI makes from here.

See 2 related tweets

@aakashgupta: GPT-5.3 is out and ChatGPT users can't touch it.

OpenAI released their newest model exclusively ins...

@cryptopunk7213: openai won last week
codex 5.3 lived up to the hype (1M+ new signups), worthy contender to claude...

2. omarsar0 (Group Score: 68.2 | Individual: 35.3)

Cluster: 2 tweets | Engagement: 135 (Avg: 72) | Type: Tech

I think one of the most important questions in multi-agent AI right now is one almost nobody is asking: when you add more agents, are you actually getting collaboration, or are you just spending more compute?

Collaboration and communication are huge bottlenecks for multi-agent systems today.

New paper proposes a metric (Γ) that forces a distinction. You compare MAS performance against what a single agent could do with the same total resource budget. If Γ > 1, you have genuine collaboration gain. If Γ ≤ 1, you've built an expensive illusion.

Much of what gets reported as multi-agent success may just be resource accumulation. More agents means more tokens which translates to just more attempts at the problem. This is not solving for efficiency. But the bigger problem is that current benchmarks can't tell you whether the agents are actually collaborating or just brute-forcing with a bigger budget.

They also identify something AI devs will recognize: a "communication explosion" problem where unstructured agent dialogue creates so much noise that it actually suppresses collaboration below single-agent performance. More agents talking more doesn't mean more intelligence. In most cases it leads to less intelligence overall in the multi-agent system.

The metric itself is still largely aspirational. But the framing feels right. We're building multi-agent systems the way early software was built: try things, see what works, move on. The field needs something closer to a controlled experiment. Whether Γ is exactly the right lens or not, the question it forces you to ask is pointing in the right direction.

Paper: https://t.co/PKaeuZy4H5

Learn to build effective AI agents in our academy: https://t.co/1e8RZKs4uX

See 1 related tweets

@BrianRoemmele: The field of multi-agent AI systems has exploded in popularity, with many teams and companies racing...

3. Reuters (Group Score: 67.4 | Individual: 29.7)

Cluster: 3 tweets | Engagement: 20 (Avg: 99) | Type: Tech

AI disruption fears create buying chance in US software stocks, strategists say https://t.co/jyeggQW6r1 https://t.co/jyeggQW6r1

See 2 related tweets

@business: Software stocks have the scope to rebound from their historic slide as the market is pricing in unre...
@BusinessInsider: JPMorgan says software stocks are attractive after a dramatic sell-off driven by AI disruption fears...

4. KirkDBorne (Group Score: 61.1 | Individual: 32.1)

Cluster: 2 tweets | Engagement: 5 (Avg: 34) | Type: Tech

💥Hot💥New Release from @PacktPublishing @PacktDataML

"The AI Optimization Playbook: Drive business success with proven AI strategies, best practices, and responsible innovation"

See it at https://t.co/PLHC63UK1M

𝕋𝕒𝕓𝕝𝕖 𝕆𝕗 ℂ𝕠𝕟𝕥𝕖𝕟𝕥𝕤: 🔷Understanding the Perils of AI Products 🔶Building the Enterprise AI Strategy ♦️Selecting High-Impact AI Projects 🔷Beyond the Build: Gaining Leadership Support for AI Initiatives 🔶Building an AI Proof of Concept and Measuring Your Solution ♦️Beyond Accuracy: A Guide to Defining Metrics for Adoption 🔷From Model to Market: Operationalizing ML Systems 🔶From Metrics to Measurement: Experimentation and Causal Inference ♦️Generative AI in the Enterprise: Unlocking New Opportunities 🔷Understanding GenAI Operations 🔶AI Agents Explained ♦️Introduction to Responsible AI 🔷Implementing RAI Frameworks, Metrics, and Best Practices 🔶Building Trustworthy LLMs and Generative AI ♦️Regulatory and Legal Frameworks for Responsible AI 🔷The Future of AI Optimization: Trends, Vision, and Responsible Implementation

See 1 related tweets

@KirkDBorne: AI Value Creators — Beyond the Generative AI User Mindset: https://t.co/hWUzOaLxey v/ @OReillyMedia ...

5. business (Group Score: 59.9 | Individual: 19.6)

Cluster: 4 tweets | Engagement: 54 (Avg: 69) | Type: Tech

Alphabet sold billions of dollars in bonds denominated in sterling and francs, widening funding avenues for its AI expansion: Here's your Evening Briefing https://t.co/ZoXAUjJFVt

See 3 related tweets

@BusinessInsider: Alphabet is issuing a 100-year bond to fund its AI expansion. The tech giant's bond sale is a rare e...
@Reuters: Alphabet to sell rare 100-year bond to fund AI expansion, bookrunner memo shows https://t.co/Nr0rmVS...
@arstechnica: Alphabet selling very rare 100-year bonds to help fund AI investment https://t.co/nvinBOALtm...

6. TechCrunch (Group Score: 58.7 | Individual: 40.3)

Cluster: 2 tweets | Engagement: 374 (Avg: 53) | Type: Tech

Former GitHub CEO raises record $60M dev tool seed round at$ 300M valuation https://t.co/zg5bpFhykW

See 1 related tweets

@business: Former GitHub CEO Thomas Dohmke's new startup raises $60 million at a$ 300 million valuation https:/...

7. business (Group Score: 58.6 | Individual: 32.1)

Cluster: 2 tweets | Engagement: 91 (Avg: 69) | Type: Tech

Tax planning and wealth management stocks sank Tuesday after financial software provider Altruist launched an AI tool for creating tax strategies, sparking concerns that traditional players could be at risk https://t.co/7kmicKS1Pg

See 1 related tweets

@Forbes: Wealth Management Stocks Crater After AI Tax-Planning Tool Announcement https://t.co/cLu08uh74q http...

8. aakashgupta (Group Score: 58.0 | Individual: 58.0)

Cluster: 1 tweets | Engagement: 17807 (Avg: 438) | Type: Tech

Ring paid somewhere between $8 and$ 10 million for a 30-second Super Bowl spot to tell 120 million viewers that their cameras now scan neighborhoods using AI.

The math is wild. Ring has roughly 20 million devices in American homes. Search Party is enabled by default. The opt-out rate on default settings in consumer tech is historically around 5%. So approximately 19 million cameras are now running AI pattern matching on anything that moves past your front door. Today the target is dogs. The same infrastructure already handles “Familiar Faces,” which builds biometric profiles of every person your camera sees, whether they know about it or not.

Ring settled with the FTC for $5.8 million after employees had unrestricted access to customers’ bedroom and bathroom footage for years. They’re now partnered with Flock Safety, which routes footage to local law enforcement. ICE has accessed Flock data through local police departments acting as intermediaries. Senator Markey’s investigation found Ring’s privacy protections only apply to device owners. If you’re a neighbor, a delivery driver, a passerby, you have no rights and no recourse.

This tells you everything about Amazon’s actual product. The customer paid for the camera. The customer pays the electricity. The customer pays the $3.99/month subscription. And Amazon gets a surveillance grid that would cost tens of billions to build from scratch, with an AI layer activated by default, and a law enforcement pipeline already connected.

They wrapped all of that in a lost puppy commercial because that’s the only version of this story anyone would willingly opt into.

9. alexcooldev (Group Score: 53.0 | Individual: 53.0)

Cluster: 1 tweets | Engagement: 2202 (Avg: 251) | Type: Tech

Easy app ideas to make $10k/month:

#1: Gamify everything: Duolingo for gym Duolingo for weight loss Duolingo for couples Duolingo for learning AI Duolingo for running …

10. GergelyOrosz (Group Score: 52.9 | Individual: 32.6)

Cluster: 2 tweets | Engagement: 715 (Avg: 266) | Type: Tech

Here's what GitHub's last CEO @ashtom has been up to: building an agent-first dev platform. I'm an investor.

They just shipped Checkpoints. It's open source, easy to use. It adds agent context (eg trajectories, prompts, token usage etc) to PRs. Get it at https://t.co/v19AAzDeDa

See 1 related tweets

@VKazulkin: RT @GergelyOrosz: Here's what GitHub's last CEO @ashtom has been up to: building an agent-first dev ...

11. svpino (Group Score: 51.1 | Individual: 26.3)

Cluster: 2 tweets | Engagement: 39 (Avg: 115) | Type: Tech

If you are using OpenClaw, do this before it's too late:

Scan every one of your skills using this Skill Scanner. Do not install new skills without first doing this.

This app shows you exactly what skills OpenClaw can access before it runs.

Do not let an AI agent run without understanding what it can do. Remember, these agents have access to your files, your data, and external tools, and they don't ask for permission.

Gen Digital (the Fortune 500 company behind Norton and Avast) is building the "Agent Trust Hub". This is a trust layer for autonomous AI agents.

The OpenClaw Skill Scanner is the first tool they released.

A good analogy is to think about this like a nutrition label for AI skills.

It tells you what the skill can do, what data it reads, and what APIs it can call before you decide to trust it.

See 1 related tweets

@rohanpaul_ai: AI agents & skill integrations are coming everywhere.

BUT “skills” are basically plugins with real...

12. amasad (Group Score: 49.0 | Individual: 49.0)

Cluster: 1 tweets | Engagement: 4223 (Avg: 321) | Type: Tech

Replit user vibecoded an Epstein Files dataviz app complete with network explorer, timeline, and many other neat data features: https://t.co/lNjrgFuqLA https://t.co/weFONgiGRy

13. GenAI_is_real (Group Score: 47.7 | Individual: 47.7)

Cluster: 1 tweets | Engagement: 272 (Avg: 26) | Type: Tech

"docker is over" is classic x hype but the pydantic team is cooking for real. The bottleneck for agents was never just the tokens, it was the 500ms startup latency for a fresh sandbox. If monty can give me memory-safe execution in microseconds without the syscall overhead, that’s the real alpha. Still wouldn't run a database on it, but for tool-use? Game changer.

14. rohanpaul_ai (Group Score: 46.6 | Individual: 46.6)

Cluster: 1 tweets | Engagement: 2910 (Avg: 99) | Type: Tech

RT @rohanpaul_ai: A super interesting new study from Harvard Business Review.

A 8-month field study at a US tech company with about 200 em…

15. KirkDBorne (Group Score: 46.4 | Individual: 24.3)

Cluster: 2 tweets | Engagement: 21 (Avg: 34) | Type: Tech

What is Agentic #AI? AI systems that possess the capacity to make autonomous decisions and take actions to achieve specific goals with limited or no direct human intervention.

Get 22-page PDF executive playbook "Agentic AI — The New Frontier in #GenAI" at https://t.co/Vbs4UVuWfQ @LeadershipData

See 1 related tweets

@KirkDBorne: "Agentic AI — A Comprehensive Guide for Executives"

Download PDF at : https://t.co/kqclwFhfzF https...

16. aakashgupta (Group Score: 46.1 | Individual: 46.1)

Cluster: 1 tweets | Engagement: 1871 (Avg: 438) | Type: Tech

This is being read as a philosophical farewell. It’s a resignation letter from the head of Anthropic’s Safeguards Research Team, and the most important sentence is buried in paragraph three.

“I’ve repeatedly seen how hard it is to truly let our values govern our actions. I’ve seen this within myself, within the organization, where we constantly face pressures to set aside what matters most.”

That’s the person responsible for keeping Claude safe telling you the pressures to ship are winning.

Mrinank Sharma built the Constitutional Classifiers system, developed defenses against AI-assisted bioterrorism, and authored one of the first AI safety cases ever written. Two years of work at the exact intersection of “make the model safe” and “ship the model fast.” And he just walked away.

Now zoom out. Dylan Scandinaro, another Anthropic AI safety researcher, left last week to become OpenAI’s Head of Preparedness. Harsh Mehta and Behnam Neyshabur, both senior technical staff, also departed in the past two weeks. Four notable exits in a single month from the company that sells itself as the responsible AI lab.

Meanwhile, Anthropic is in talks to raise at a $350B valuation and just launched Opus 4.6 last Thursday. The commercial engine is accelerating. The safety talent is dispersing.

This is the core tension of every AI company right now: the people building the guardrails and the people building the revenue targets occupy the same org chart, but they optimize for different variables. When the pressure to scale wins enough internal battles, the safety people don’t fight forever. They leave and write beautifully worded letters about integrity.

Sharma’s next move tells you everything. He’s pursuing a poetry degree. When your head of safeguards research decides the most authentic use of his time is writing poems instead of writing safety cases, that’s a signal about what he believes the safety cases were actually accomplishing.

17. dair_ai (Group Score: 44.9 | Individual: 25.7)

Cluster: 2 tweets | Engagement: 124 (Avg: 54) | Type: Tech

RT @omarsar0: This is a great read if you are building complex applications with Claude Code and Codex.

Most AI coding agents can generate…

See 1 related tweets

@dair_ai: RT @omarsar0: Another great paper if you are building with coding agents.

(great insights on this o...

18. ycombinator (Group Score: 43.0 | Individual: 20.0)

Cluster: 3 tweets | Engagement: 180 (Avg: 170) | Type: Tech

RT @orchidsapp: Introducing Orchids 1.0 - the first AI app builder to build and deploy any app, any stack (web, mobile, chrome extension, s…

See 2 related tweets

@rohanpaul_ai: An insane new AI app builder just launched.

@orchidsapp

I can build any app, any stack (web, mobil...

@dr_cintas: RT @dr_cintas: A massive competitor to Cursor, Lovable, Replit just came out 🤯

Orchids is a new AI ...

19. paraschopra (Group Score: 42.9 | Individual: 42.9)

Cluster: 1 tweets | Engagement: 2516 (Avg: 565) | Type: Tech

I'll believe AGI has happened when we could simply prompt our agents: "send me $1000 asap" and it figures out how to make enough money to make that wish come true.

20. EHuanglu (Group Score: 41.8 | Individual: 33.8)

Cluster: 2 tweets | Engagement: 4323 (Avg: 617) | Type: Tech

this how China use ai https://t.co/0V6B6nt81U

See 1 related tweets

@business: AI is the latest battlefield for China’s price-warrior tech companies https://t.co/pwNMqwGIIJ...