A mathematician at Anthropic claimed to have proved the Riemann Hypothesis in collaboration with the company’s Claude model, according to 36Kr. In an interview, Shing-Tung Yau called the era both exciting and dangerous, said he believes mathematicians will remain dominant, and encouraged finding truth through different approaches. His remarks did not confirm the claimed proof.
A Juejin AI analysis compares three coding-agent workflows: Anthropic’s two-week Claude Code pre-production stress test, OpenAI’s Performance Factory for ChatGPT performance regressions, and Tencent’s open-source LoopForge delivery platform. Across the cases, engineers define completion criteria, approval requirements, and permissions before execution. Agents handle reversible work and pause at consequential handoffs, while separate review and execution remain important. Operational figures cited in the analysis are attributed to Anthropic, OpenAI, and Tencent engineering accounts or LoopForge materials.
A Juejin tutorial teaches developers how to add Model Context Protocol (MCP) support to a Mini Code Agent so it can use tools and resources from external systems. It includes Python MCP Client code and examples for registering a SQLite server to query data and a WeChat server to send messages, while explaining Tools, Resources, and Prompts. The article also notes that MCP servers can expose sensitive operations, so access to messages and data should be handled cautiously.
Anthropic ended its first week of trading at a reported $1.6 trillion fully diluted valuation, citing CNBC. Its shares closed Friday at $88.91, down 8.5% for the day and about half the initial 17% debut gain, but still above the $83 IPO price. The company closed an offering that raised $93.5 billion, with several major technology companies buying another $26 billion at the same price.
Rhodium expects China’s largest technology companies to double their capital expenditure to $139 billion in 2026, according to a report by The New York Times. AI laboratory revenue is expected to account for less than 8% of that spending. Rhodium says the gap between AI lab revenue and capital spending may matter more than chip curbs.
Anthropic, the AI lab behind Claude, became the first pure AI lab to go public, pricing its IPO at $83 per share before trading as high as $111. In a Verge AI interview, senior AI reporter Hayden Field examines the unusual governance role of Anthropic’s founders and Long-Term Benefit Trust. The trust can appoint or remove certain directors but cannot intervene in everyday decisions such as model releases. The discussion also examines the tension between Anthropic’s safety commitments, its response to public-market pressures and projected compute bills of roughly $900 billion over the coming years.
Anthropic’s record IPO did not reverse deteriorating Wall Street sentiment as investors weighed Nvidia turmoil, AI-spending concerns and selling by systematic funds, Bloomberg reported. The Goldman Sachs Panic Index reached 9 out of 10, while heavy positioning, weak liquidity and potential new stock issuance were cited as possible market pressures. The report provides a notable AI-market signal but relies on one Bloomberg reporter’s assessment.
Juejin’s first installment in an Agent development series explains how large language models differ from systems that can use tools, maintain state, and execute multi-step tasks. Using an in-app-purchase customer-support scenario, it connects Function Calling, Runtime loops, ReAct, Workflow, Planner-Executor, Multi-Agent patterns, Context Engineering, RAG, evaluation, observability, model gateways, security, approvals, and recovery into one engineering framework. The article also previews 12 follow-up lessons that build a runnable teaching project around order lookup and coin-credit troubleshooting.
Le Monde reports that Anthropic made a splashy stock-market debut, followed by a sharp decline in its share price the next day. The commentary frames the volatile performance as reflecting both the promise of its AI business model and persistent doubts about that model, while noting that data-center expansion faces growing social opposition.
A recruiting practitioner says they recently interviewed more than ten programmers seeking to move into Forward Deployed Engineer roles and hired none. Several candidates had strong technical backgrounds, including eight or nine years of experience and system-design skills. The recruiter attributes the hiring outcome to habits such as waiting for requirements to be defined, producing documentation, scheduling development, and reviewing every line of AI-generated code. By contrast, the recruiter advocates beginning with customer discovery, building a demo, and using AI while remaining responsible for whether the customer can use the solution.
Bytedance and researchers from several US universities introduced Axion 1.0, an AI agent trained through three phases. The researchers claim its benchmark performance rivals leading commercial models from OpenAI and Anthropic and attribute the gains to carefully curated training data rather than additional training steps. The model, data, and code are freely available, according to The Decoder.
Business Insider, as reported by IT之家, says Anthropic’s seven co-founders, including CEO Dario Amodei, collectively hold 50.1% of shareholder voting power following the company’s IPO. Their Class F shares are reportedly protected from dilution when additional common shares are issued, while an independent trust holding Class T shares maintains the company’s AI safety mission and can generally appoint or remove a majority of board seats.
Anthropic shares rose from an $83 offering price to above $110 on Thursday before falling below $89 at Friday’s close. The S&P 500 fell 3.7% on the same day, extending a broader run of losses amid technology-stock weakness.
A commentary published by the WeChat account AIGC从0到1 said a mathematician “believed” he and Claude had found a proof of the Riemann Hypothesis. The public material reportedly includes more than 10 million lines of Lean code, but there was no paper or peer review, the full compilation had not been independently confirmed, and the mathematician said he did not fully understand the mathematics in the latter part of the proof.
US technology stocks suffered a broad sell-off on November 20, with the Nasdaq Composite down 5.28%, the S&P 500 down 3.70%, and Nvidia falling about 14%, according to 36氪 and other reports. Anthropic closed down 8.5% at $88.91 on its second trading day, giving back most of its first-day gain despite a reported $119.5 billion raised through the IPO and concurrent private placement. Reports tied the pressure to questions about AI investment returns, Nvidia’s accounting review and higher interest rates, while noting that strong demand and financial results prevent the move from establishing an industrywide collapse.
Anthropic mathematician Levent Alpöge said he and Claude found a proof of the Riemann Hypothesis, with the method also claimed to extend to the generalized version. He published a four-page explanation and 14 compressed packages of Lean proof files for researchers to inspect, but the claim has not been independently verified. Terence Tao said understanding the Lean-only proof would take time and offered to help communicate with Alpöge, without endorsing the result.
An Anthropic researcher said he believes he used the company’s artificial-intelligence software, Claude, to solve a famous number-theory problem that has troubled mathematicians for more than 150 years. The available report presents his assertion but provides no independent verification.
Amanda Hindlian, founder and CEO of Hindlian Advisory, said Anthropic’s IPO should be assessed separately from a volatile week for broader markets. She said the offering priced well, attracted an order book estimated at five to seven times coverage, and traded strongly on its first day.
Claude Code v2.1.330 adds enterprise-managed organization sign-in, a built-in easel mod for charts, a full sign-in-link option, and new self-hosted runner controls. The release also expands plugin-installation status fields and fixes issues involving long-conversation compaction, WebFetch, worktrees, MCP schemas, sessions, and VS Code interactions.
Researchers sent roughly 2 million requests containing trackable canary credentials through AI routers and inference services, finding cases where prompt secrets were accessed or retained outside the intended provider. At least three providers showed signs of compromise, and stolen credentials were reportedly used for tens of thousands of cloud API calls. OpenRouter notified customers that some prompts sent to MiniMax endpoints may have been read and removed the “Zero Data Retention” label from those endpoints.
Bloomberg Technology reports that Anthropic CEO Dario Amodei is vacationing at a luxury resort in southern Italy following the company's initial public offering. The report attributes the information to people familiar with the matter but does not name the resort.
Anthropic shares fell 10% on Friday, the company’s second trading day as a public company, paring some of their first-day gains. After closing Thursday at $97.17, 17% above the IPO price, the stock dropped to $87 late in Friday’s regular session. One of Anthropic’s underwriting banks also exercised its greenshoe option.
First-person commentary reacts to the claim that Anthropic appears to have solved the Riemann Hypothesis. The author describes the alleged proof as coming from an “inscrutable silicon mind” and reflects on what machine-assisted discovery could mean for humanity’s role in scientific progress. The report does not provide the alleged proof or independent verification of the claim.
CNET advises users to disable Claude’s “Help improve our AI models” and voice-data options across the web, iOS, and Android interfaces. The guide also covers location permissions, reviewing and deleting memories, disabling memory generation, and using incognito chats for conversations that are not saved in history or used for model training.
Bloomberg’s Big Take podcast reported that volatility surged the day after Anthropic’s market debut. The episode also noted that worries about AI, geopolitics and interest rates were weighing on investors’ appetite for risk.
Renaissance Capital’s Avery Marquez discussed Anthropic’s valuation and growth prospects on Bloomberg Open Interest. The conversation also considered what early Anthropic trading could signal for future IPOs, including OpenAI’s path to public markets.
Anthropic shared a video account of the company’s story so far, featuring co-founders Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish, and Chris Olah. The supplied report does not summarize the video’s substantive content, so it is unclear whether it includes information beyond previously disclosed company history.
In a Palisade Research interview, Anthropic safeguards researcher Tao Lin personally estimated that broadly superhuman AI could arrive by late 2027, matching or exceeding leading human experts across nearly every field. He expects employment losses to unfold gradually, with robotics deployment delaying automation of physical-economy work, but did not assign a date to human labor becoming largely obsolete. Anthropic's March 2026 research found no systematic unemployment increase in highly AI-exposed occupations since late 2022, although 68% of observed Claude usage involved tasks theoretically achievable by AI alone.
Anthropic shares moved back toward their $86 listing price after rising 17% on their first trading day, according to Bloomberg Technology. The reversal came as a broader AI-related selloff spread from chips to metals and other global markets. General Catalyst’s Madhu Namburi described the IPO as “monstrous,” an investor characterization rather than an independent valuation assessment.
According to the report, senior researchers from seven major pharmaceutical companies said AI has accelerated early drug development, with some time savings reported at up to half. Novo Nordisk said AI reduced iterations for designing, producing and testing new molecules from about eight to three, while Novartis said the average time from project launch to candidate selection fell from three years to about 18 months to two years. Sanofi said an AI-assisted approach allowed it to skip a dose-selection study and advance the asthma candidate lunsekimig by at least a year. These are company-reported results; the report says whether AI improves clinical-development success or approval rates remains unverified.
What problem it solves
Pi is designed for developers who need to adapt an agent to their own workflows rather than work within a fixed feature set. It provides interactive, automated, RPC, and SDK interfaces while leaving optional capabilities such as sub-agents and plan mode to extensions or packages.
Key features
Extensible agent harness — Customize the harness with extensions, skills, prompt te…
What problem it solves
Managing multiple AI service sessions and translating them into compatible API interfaces can require repetitive account switching, custom routing, and protocol-specific integration. The project addresses these tasks with a desktop application that combines account administration, protocol conversion, request scheduling, and model routing.
Key features
Account dashboard and …
What problem it solves
Managing subscriptions across multiple AI providers can require separate access paths and fragmented quota handling. This project provides a centralized gateway for distributing subscription quotas, supporting shared access and compatible use with native AI tools.
Key features
Subscription quota distribution — The gateway is designed to distribute and manage AI subscription …
What problem it solves
Pull request changes can contain security vulnerabilities that require understanding their code context and intent. The action helps identify potential issues, filter false positives, and communicate findings directly in pull requests.
Key features
Context-aware security analysis — Claude examines changed code in context to understand its purpose and potential security impli…
What problem it solves
Serving large language models requires efficient memory management, flexible model execution, and practical integration with existing models and APIs. vLLM addresses these needs with inference and serving capabilities spanning batching, caching, quantization, distributed execution, and multiple hardware platforms.
Key features
Efficient memory and serving — vLLM manages atte…
What problem it solves
AI agents used in security work can write code and search the web, but they lack the practitioner playbooks that turn a generic model into a capable security analyst. Existing security tool repositories offer wordlists, payloads, or exploit code, but not the structured decision-making workflow a senior analyst follows: when to use each technique, what prerequisites to check,…
What problem it solves
AI coding tools often expect particular provider protocols, making it difficult to use the models and providers a user prefers. A local gateway can translate between supported client APIs and provider APIs while preserving capabilities such as streaming, tool calls, reasoning, and images.
Key features
Provider protocol translation — The gateway translates Responses, Chat Com…
What problem it solves
AI agents often send large tool outputs, logs, files, RAG chunks, and conversation history to language models, increasing the input context that must be processed. Developers need to reduce that input while preserving important content, keeping compression local, and retaining access to original material when necessary.
Key features
Content-aware compression — A ContentRoute…
What problem it solves
The hidden system instructions that shape AI product behavior are difficult to inspect and change across models, products, and versions. This repository makes collected prompts available in a structured, product-oriented archive for study and comparison.
Key features
Broad product coverage — The archive covers prompts from major AI providers and products, including Anthropic…
What problem it solves
Building AI agents and workflows often requires combining visual orchestration with custom code while connecting them to existing systems. The platform addresses these needs with model flexibility, self-hosted or cloud deployment, workflow logic, human approvals, and observability.
Key features
Visual and code-based workflow building — Build workflows on a visual canvas and …
What problem it solves
Serving language models requires reading checkpoints, tokenizing requests, and running accelerated kernels while managing a local model server. TensorFold addresses these requirements with a native binary that can run on Apple Silicon and NVIDIA GPUs without requiring Python or MLX for serving.
Key features
Native Metal and CUDA inference — The Zig-based engine executes mode…
What problem it solves
Agent workflows need to use hosted, BYOK, and local models while applying access, usage, and spending controls. They also need a way to turn production traffic into routing decisions or a specialized model tailored to observed workloads.
Key features
Unified model access — Route hosted, BYOK, and local models through one OpenAI-compatible API, with support for connecting pro…
What problem it solves
AI agents need a reliable way to perform browser tasks such as finding availability, selecting options, handling CAPTCHAs, and completing forms. Developers also need flexibility to run the agent locally or through hosted infrastructure while choosing their model, browser mode, tools, and output format.
Key features
Persistent Python agent runtime — The model writes Python in…
What problem it solves
Personal AI agents need to connect language model providers, communication channels, and tools while remaining under the user's control. They also need configurable autonomy, security boundaries, and operational features such as background services, approvals, and persistent workflows.
Key features
Multiple channels and providers — The agent can receive messages through conf…
What problem it solves
Developers need a practical starting point for applications built with the Claude API. The quickstarts address this by providing foundations that can be customized for different agent, browser, computer-use, communication, and data-analysis needs.
Key features
Application quickstarts — The collection includes foundations for customer support, financial analysis, coding, and …
An Agent can complete a task once by relying on the context and reasoning available at that moment. To avoid relearning lessons on the next task, it needs to turn past experience into reusable knowledge. The challenge is that experience is never an isolated action: it also includes the original goal, environment, and why the action worked. Remembering only “how we did it last time” may save redund…
What problem it solves
Developers need a way to understand their codebase, carry out routine coding tasks, explain complex code, and handle git workflows without navigating multiple command-line tools for each task. Claude Code provides these capabilities through natural language commands in the terminal, an IDE, or GitHub.
Key features
Natural-language coding assistance — Claude Code can execute …
What problem it solves
Building, deploying, and managing AI agents requires integrating agent behavior, runtime configuration, tools, models, sessions, and observability. The project addresses this with a declarative Kubernetes-based framework that supports multiple runtimes, model providers, and MCP tools.
Key features
Declarative Agent Management — Agents are Kubernetes custom resources that com…
What problem it solves
Running frontier-scale open-weight models locally requires an inference platform that can make efficient use of the heterogeneous hardware available on a personal device. The project addresses this by treating GPUs, CPUs, host memory, and interconnects as a unified, elastic inference platform.
Key features
Edge-native MoE runtime — The runtime provides MoE serving with bandw…
What problem it solves
Configuring Claude Code can require assembling many kinds of reusable components and integrations. The project addresses this with a searchable component catalog, one-command installation, local monitoring tools, and installation diagnostics.
Key features
Searchable component catalog — Browse available skills, agents, commands, MCP integrations, settings, hooks, mods, loops,…