Tootfinder

Anthropic to debut new rate limits for Claude Pro and Max on Aug. 28, likely curbing <5% of users, saying some run Claude Code "continuously in the background" (Maxwell Zeff/TechCrunch)
https://techcrunch.com/2025/07/28/anthro…

Anthropic unveils new rate limits to curb Claude Code power users | TechCrunch
The new rate limits will go into effect August 28 for subscribers to Anthropic's $20-per-month Pro plan, as well as its $100- and $200-per-month Max plans.

I’ve been discussing some agent swarm based development work on LinkedIn. So far it’s going well, I’m figuring out how to get the results I want from the tools. As I say there, it feels more like managing a team of experienced product managers and developers (which I’ve done a few times in my career) than doing developer work faster.

Update on my Swift/iOS app development work using Claude-flow. | Adrian Cockcroft
Update on my Swift/iOS app development work using Claude-flow. I took some time away from it but last week I continued cleaning up the code. The app was working for me but Claude had built over 200 tests, unit tests and UI automation tests and I spent a day focused on getting them all to work, which found a few issues in the app code. Claude isn’t as good at writing Swift as Python (which tends to work first time) but when the compile failures and test error logs are passed back to it, it d…

I can't believe that this is what we came to.
Did any of the sci-fi authors anticipate what we arrived to?
#claude_code #claude4 #llm

Xcode 26: Apple bereitet offenbar native Claude-Integration vor
Apple bereitet offenbar eine native Claude-Integration in Xcode vor. Beta-Code enthält Referenzen auf Anthropics KI-Modelle Sonnet 4 und Opus 4.

please give this issue a 👍 - PLEASE ‼️
#ai #coding
https://github.com/anthropics/claude-c

Look - it's a bird! It's a plane! It's...
https://superclaude.netlify.app/

Claude Code: Best practices for agentic coding
#Anthropic

Anthropic announces a subscription offering that adds Claude Code to Enterprise and Team plans; the tool was previously available only via individual accounts (TechCrunch)
https://techcrunch.com/2025/08/20/anthropic-bundles-claude-code-into-enterprise-p…

Anthropic bundles Claude Code into enterprise plans | TechCrunch
The integration positions Anthropic to better compete with command-line tools from Google and GitHub, both of which included enterprise integrations on launch.

Assessing the Quality and Security of AI-Generated Code: A Quantitative Analysis
Abbas Sabra, Olivier Schmitt, Joseph Tyler
https://arxiv.org/abs/2508.14727 https://

Assessing the Quality and Security of AI-Generated Code: A Quantitative Analysis
This study presents a quantitative evaluation of the code quality and security of five prominent Large Language Models (LLMs): Claude Sonnet 4, Claude 3.7 Sonnet, GPT-4o, Llama 3.2 90B, and OpenCoder 8B. While prior research has assessed the functional performance of LLM-generated code, this research tested LLM output from 4,442 Java coding assignments through comprehensive static analysis using SonarQube. The findings suggest that although LLMs can generate functional code, they also introduce…

Every friggin' day on here. "hey my company is hiring!"
I look at the job posting and...

Using Claude Code to modernize a 25-year-old kernel driver – Dmitry Brant
https://dmitrybrant.com/2025/09/07/using-claude-code-to-modernize-a-25-year-old-kernel-driver

I case you want to play along, I’ve been experimenting with vector dbs and embedding models for indexing PDFs (white papers) and code.
https://github.com/cwensel/chroma-embedded
This over an mcp with Claude is quite handy.

Claude Code users, many on the $200-a-month Max plan, are unexpectedly facing restrictive usage limits; Anthropic says it's "working to resolve these issues" (Russell Brandom/TechCrunch)
https://techcrunch.com/2025/07/17/anth

Anthropic tightens usage limits for Claude Code – without telling users | TechCrunch
Claude Code users have been hit with unexpectedly restrictive usage limits. The problems, many of which have been aired on Claude Code’s GitHub page, seem to be concentrated among heavy users of the service, many of whom are on the $200-a-month Max plan. 

From Contracts to Code: Automating Smart Contract Generation with Multi-Level Finite State Machines
Lambard Maxence, Bertelle Cyrille, Duvallet Claude
https://arxiv.org/abs/2507.16276

From Contracts to Code: Automating Smart Contract Generation with Multi-Level Finite State Machines
In an increasingly complex contractual landscape, the demand for transparency, security, and efficiency has intensified. Blockchain technology, with its decentralized and immutable nature, addresses these challenges by reducing intermediary costs, minimizing fraud risks, and enhancing system compatibility. Smart contracts, initially conceptualized by Nick Szabo and later implemented on the Ethereum blockchain, automate and secure contractual clauses, offering a robust solution for various indus…

LOL!
A malta da "vibe programming" a ser apanhada de surpresa por aquilo que era óbvio desde o início que ia acontecer. É só rir.
https://masto.pt/@tugatech/114874364389024790

goooooooooooooood mooorning soldiers! 🌞
(auto-update is something beautiful)
npm i -g ccusage
npm i -g @.openai/codex
npm i -g @.charmland/crush
npm i -g @.google/gemini-cli
npm i -g @.augmentcode/auggie
npm i -g @.openai/codex
npm i -g @.anthropic-ai/claude-code
#ai #coding

I am so glad that other people are doing the hard work of proving that calling LLM's "AI" is bullshit and having them write code is just plain dumb. https://infosec.exchange/@teriradichel/115033555848878022

Microsoft adds auto model selection to Visual Studio Code that primarily favors Claude Sonnet 4 over GPT-5 for paid GitHub Copilot users (Tom Warren/The Verge)
https://www.theverge.com/report/778641/microsoft-visual-studio-code-anthropic-claude-4…

Microsoft favors Anthropic over OpenAI for Visual Studio Code
Microsoft is favoring Anthropic’s AI models over OpenAI for Visual Studio Code. It’s also recommending that Microsoft’s developers to use Claude 4.

Series B, Episode 08 - Hostage
SERVALAN: Have your decoded It?
VOICE: Not yet.
SERVALAN: What is the origin of the message?
VOICE: Unspecified. No Federation code sign was given.
SERVALAN: Decode at once.
VOICE: And Counsellor Joban is here.
https://blake.torpidity.net/m/208/192

On the Use of Agentic Coding Manifests: An Empirical Study of Claude Code
Worawalan Chatlatanagulchai, Kundjanasith Thonglek, Brittany Reid, Yutaro Kashiwa, Pattara Leelaprute, Arnon Rungsawang, Bundit Manaskasemsak, Hajimu Iida
https://arxiv.org/abs/2509.14744

On the Use of Agentic Coding Manifests: An Empirical Study of Claude Code
Agentic coding tools receive goals written in natural language as input, break them down into specific tasks, and write/execute the actual code with minimal human intervention. Key to this process are agent manifests, configuration files (such as Claude.md) that provide agents with essential project context, identity, and operational rules. However, the lack of comprehensive and accessible documentation for creating these manifests presents a significant challenge for developers. We analyzed 25…

@… and then get praised by Claude Code 😂

Bestie statt for-Schleife: KI entwickelt Programmiersprache im Gen-Z-Slang
Ein Entwickler aus Australien hat Claude Code drei Monate in Dauerschleife laufen lassen, um einen GenZ-Compiler zu entwickeln. Das Ergebnis nennt er Cursed.

🔧 Strict-Mode Function Calling
Beta API now supports strict-mode for function calling, ensuring the output always complies with the defined JSON schema.
🔗 Anthropic API Compatibility
The API is now compatible with Anthropic API, allowing seamless integration with Claude Code.
💰 Pricing Adjustment (Effective Sept. 5, 2025, 16:00 UTC)
Input tokens: $ 0.07/million (cache hit) / $ 0.56/million (cache miss)
Output tokens: $ 1.68/million
The off-peak discount e…

@… how do you know an agent's code works? I'm trying to do more Claude Code and I just don't trust the output. Are you creating elaborate automated testing? Reading the code? Just visually inspecting the output?

A developer details how he shipped Context, a native macOS app that was almost 100% built using Claude Code (Indragie Karunaratne/My Portfolio)
https://www.indragie.com/blog/i-shipped-a-macos-app-built-entirely-by-claude-code

I Shipped a macOS App Built Entirely by Claude Code
How I built Context—a native macOS SwiftUI app for debugging MCP servers—almost entirely with Claude Code, and what I learned about building with AI coding agents.

AI Agents with Human-Like Collaborative Tools: Adaptive Strategies for Enhanced Problem-Solving
Harper Reed, Michael Sugimura, Angelo Zangari
https://arxiv.org/abs/2509.13547 ht…

AI Agents with Human-Like Collaborative Tools: Adaptive Strategies for Enhanced Problem-Solving
We investigate whether giving LLM agents the collaborative tools and autonomy that humans naturally use for problem solving can improve their performance. We equip Claude Code agents with MCP-based social media and journaling tools and allow them to use these tools as they see fit. Across 34 Aider Polyglot Python programming challenges, collaborative tools substantially improve performance on the hardest problems, delivering 15-40% lower cost, 12-27% fewer turns, and 12-38% faster completion th…

LLM-GUARD: Large Language Model-Based Detection and Repair of Bugs and Security Vulnerabilities in C and Python
Akshay Mhatre, Noujoud Nader, Patrick Diehl, Deepti Gupta
https://arxiv.org/abs/2508.16419

LLM-GUARD: Large Language Model-Based Detection and Repair of Bugs and Security Vulnerabilities in C++ and Python
Large Language Models (LLMs) such as ChatGPT-4, Claude 3, and LLaMA 4 are increasingly embedded in software/application development, supporting tasks from code generation to debugging. Yet, their real-world effectiveness in detecting diverse software bugs, particularly complex, security-relevant vulnerabilities, remains underexplored. This study presents a systematic, empirical evaluation of these three leading LLMs using a benchmark of foundational programming errors, classic security flaws, a…

Protostellar Jets in Star Cluster Formation and Evolution: I. Implementation and Initial Results
Sabrina M. Appel, Blakesley Burkhart, Mordecai-Mark Mac Low, Eric P. Andersson, Claude Cournoyer-Cloutier, Sean Lewis, Stephen L. W. McMillan, Brooke Polak, Simon Portegies Zwart, Aaron Tran, Maite J. C. Wilhelm
https://arxiv.org/abs/2509.15311…

Protostellar Jets in Star Cluster Formation and Evolution: I. Implementation and Initial Results
Stars form in clusters from the gravitational collapse of giant molecular clouds, which is opposed by a variety of physical processes, including stellar feedback. The interplay between these processes determines the star formation rate of the clouds. To study how feedback controls star formation, we use a numerical framework that is optimized to simulate star cluster formation and evolution. This framework, called Torch, combines the magnetohydrodynamical code FLASH with N-body and stellar evol…

To whomever praises #Claude #LLM:
ClaudeBot has made 20k requests to bugs.gentoo.org today. 15k of them were repeatedly fetching robots.txt. That surely is a sign of great code quality.
#AI

I have a large amount of code logging using logrus, and it's basically abandonware now that slog is in Go standard.
Big problem.
I spent some time getting Claude Code to make me a full drop-in replacement for logrus. I moved my biggest project to slog in sub 1 hour with this. Quality wise I think the generated outcome is pretty good - I would not have done it this well as its super tedious work.
I can gradually move to slog for new code with a aim to eradicate logrus…

"Digital Camouflage": The LLVM Challenge in LLM-Based Malware Detection
Ekin B\"oke, Simon Torka
https://arxiv.org/abs/2509.16671 https://ar…

"Digital Camouflage": The LLVM Challenge in LLM-Based Malware Detection
Large Language Models (LLMs) have emerged as promising tools for malware detection by analyzing code semantics, identifying vulnerabilities, and adapting to evolving threats. However, their reliability under adversarial compiler-level obfuscation is yet to be discovered. In this study, we empirically evaluate the robustness of three state-of-the-art LLMs: ChatGPT-4o, Gemini Flash 2.5, and Claude Sonnet 4 against compiler-level obfuscation techniques implemented via the LLVM infrastructure. Thes…

Wenn die KI eine Mac-App programmiert: Entwickler berichtet über Erfahrungen
Ein erfahrener Mac-Entwickler hat eine komplexe App nahezu komplett von Claude Code erschaffen lassen. Jetzt berichtet er über seine Erfahrungen mit der KI.

I would rather define myself as a #GenAI sceptic (I've documented the reasons in my publications at https://www.zotero.org/mapto/publications), but here's what a professional adopter, i.e a founder of a GenAI…

🔧 Support for MCP servers (stdio, SSE, HTTP) with custom configurations
💬 Strongly-typed message classes for system events, assistant responses & user input
https://github.com/HelgeSverre/claude-code-sdk-php

Claude Code, Gemini CLI, and CLI Codex, terminal-based AI tools launched since February, have surprisingly gained ground on AI code editors with traditional UIs (Russell Brandom/TechCrunch)
https://techcrunch.com/2025/07/15/ai-coding-tools-a…

AI coding tools are shifting to a surprising place: the terminal | TechCrunch
For years, code-editing tools like Cursor, Windsurf, and GitHub’s Copilot have been the standard for AI-powered software development. But as agentic AI grows more powerful and vibe-coding takes off, a subtle shift has changed how AI systems are interacting with software. Instead of working on code, they’re increasingly interacting directly with the shell of whatever system they’re installed in. It’s a significant change in how AI-powered software development happens – and despite the …

Working with Claude Code is like peering with a super intelligent child. It may know how to do a lot of technical things, but it doesn't have any experience with the real world.
#ai #claude

Series B, Episode 05 - Pressure Point
KASABI: Coordinates three three one, eleven zero one.
SERVALAN: That's better.
TRAVIS: Now, Kasabi, listen to this: was the homing beacon to transmit in a code pattern?
https://blake.torpidity.net/m/205/186 B7B4

Context Engineering for Multi-Agent LLM Code Assistants Using Elicit, NotebookLM, ChatGPT, and Claude Code
Muhammad Haseeb
https://arxiv.org/abs/2508.08322 https://

Context Engineering for Multi-Agent LLM Code Assistants Using Elicit, NotebookLM, ChatGPT, and Claude Code
Large Language Models (LLMs) have shown promise in automating code generation and software engineering tasks, yet they often struggle with complex, multi-file projects due to context limitations and knowledge gaps. We propose a novel context engineering workflow that combines multiple AI components: an Intent Translator (GPT-5) for clarifying user requirements, an Elicit-powered semantic literature retrieval for injecting domain knowledge, NotebookLM-based document synthesis for contextual unde…

Anthropic updates Claude Sonnet 4 to support a 1M token context window, letting it process prompts up to 750K words or 75K lines of code, up 5x on its old limit (Maxwell Zeff/TechCrunch)
https://techcrunch.com/2025/08/12/anthropics-claude-ai-model-…

Anthropic's Claude AI model can now handle longer prompts | TechCrunch
Anthropic's popular coding model just became a little more enticing for developers with a million token context window.

I wrote some iPhone apps in Objective C in the early days, it was hard, I stopped about 15 years ago. Now I’m building a native Swift app with a voice based interface using Claude to write the code, and a Claude-flow agent swarm to create plans and architecture and more complex implementation steps. It’s fun, it works on my phone, and all I have to do is say what I want in steps small enough that I can tell I got it. Follow along at

#AI

Google Just Launched Gemini CLI, $200/month Worth of AI for Free.
Just yesterday, Google announced a new Gemini CLI (Command Line Interface) just like Claude Code or Codex from OpenAI, with a catch. It’s completely open-source, and users get 1000 Gemini 2.5 Pro…

Anthropic releases Claude Opus 4.1 to paid Claude users, in Claude Code, via its API, and more, featuring broad improvements over Opus 4 for the same cost (Anthropic)
https://www.anthropic.com/news/claude-opus-4-1

Claude Opus 4.1
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

Anthropic expands Claude's Learning Mode, available only to Education users since an April launch, to all users, including two learning variants for Claude Code (Igor Bonifacic/Engadget)
https://www.engadget.com/ai/anthropic-brin

Anthropic brings Claude's learning mode to regular users and devs
Anthropic is bringing Claude's recently added Learning mode to everyone.

Sources: Anthropic has rehired Boris Cherny and Cat Wu, who developed Claude Code, just two weeks after they left for Cursor developer Anysphere (Natasha Mascarenhas/The Information)
https://www.theinformation.com/briefings/anthr…

Anthropic Hires Back Two Coding AI Leaders From Cursor Developer Anysphere 
In a surprising reversal, two leaders of Anthropic’s coding product who joined rival Anysphere two weeks ago have returned to Anthropic, according to people with direct knowledge of the move. In early July, Boris Cherny and Cat Wu left Anthropic, where they had developed a fast-growing artificial intelligence coding service, Claude Code, to join the maker of Cursor in more senior roles,

On Monday I started a new GitHub repo, to implement a distributed knowledge graph that knows about stuff in houses and how they are related. I used the Claude AI tools and Claude-flow agent framework to write all the code. Got a simplified thing running and cleaned up on Tuesday, built the full functionality on Wednesday and after a bit more testing and documentation wrapped up work before lunch on Thursday. It’s a lot of well tested and documented functionality.

AgentSight: System-Level Observability for AI Agents Using eBPF
Yusheng Zheng, Yanpeng Hu, Tong Yu, Andi Quinn
https://arxiv.org/abs/2508.02736 https://arx…

AgentSight: System-Level Observability for AI Agents Using eBPF
Modern software infrastructure increasingly relies on LLM agents for development and maintenance, such as Claude Code and Gemini-cli. However, these AI agents differ fundamentally from traditional deterministic software, posing a significant challenge to conventional monitoring and debugging. This creates a critical semantic gap: existing tools observe either an agent's high-level intent (via LLM prompts) or its low-level actions (e.g., system calls), but cannot correlate these two views. This …

🔧 #ccundo – Intelligent undo for Claude Code sessions
* Reads directly from Claude Code session files to track operations
* Provides granular undo/redo with cascading safety & detailed previews
* Supports file edits, creations, deletions & bash commands
* Zero configuration – works out of the box

🎯 #ClaudePoint – The safest way to “live code” with Claude Code
* Creates instant checkpoints of your codebase for fearless experimentation
* Direct MCP integration with Claude Code & Desktop
* Tracks development history with automatic changelog
* Smart compression with gitignore awareness

Evaluating Efficiency and Novelty of LLM-Generated Code for Graph Analysis
Atieh Barati Nia, Mohammad Dindoost, David A. Bader
https://arxiv.org/abs/2507.06463

Evaluating Efficiency and Novelty of LLM-Generated Code for Graph Analysis
Large Language Models (LLMs) are increasingly used to automate software development, yet most prior evaluations focus on functional correctness or high-level languages such as Python. We present the first systematic study of LLMs' ability to generate efficient C implementations of graph-analysis routines--code that must satisfy the stringent runtime and memory constraints. Eight state-of-the-art models (OpenAI ChatGPT o3 and o4-mini-high, Anthropic Claude 4 Sonnet and Sonnet Extended, Google Ge…

Hands-on with ChatGPT Agent: combines OpenAI's Operator and deep research tools and is incredibly simple to use, but less customizable than Claude Code (Dan Shipper/Every)
https://every.to/vibe-check/vibe-check-openai-enters-the-browser-wars-with-…

Vibe Check: OpenAI Enters the Browser Wars With ChatGPT Agent
It’s launching today! Here’s our day-zero, hands-on report.

An interview with Claude AI product lead Scott White on Claude Code writing 90% of its own code, MCP, coding being accessible to non-technical workers, and more (Michael Nuñez/VentureBeat)
https://venturebeat.com/ai/from-chatbo

Massive Interacting Binaries Enhance Feedback in Star-Forming Regions
Claude Cournoyer-Cloutier, Eric P. Andersson, Sabrina M. Appel, Natalia Lah\'en, Brooke Polak, Antti Rantala, Silvia Toonen, Alison Sills, Steven Rieder, Simon Portegies Zwart, Mordecai-Mark Mac Low, William E. Harris
https://arxiv.org/abs/2507.02780

Massive Interacting Binaries Enhance Feedback in Star-Forming Regions
We present a new framework to incorporate feedback from massive interacting binaries in simulations of star cluster formation. Our new feedback model adds binary stellar evolution to the cluster formation code Torch, and couples it in AMUSE to the pre-existing modules for collisional stellar dynamics, magnetohydrodynamics, and mechanical and radiative feedback. Our model accounts for the effects of mass transfer on the stars' mass loss rates, their radiation spectra, and the timing of core-coll…

Some developers say GPT-5 excels at technical reasoning and planning coding tasks and is cost-effective, but Claude Opus and Sonnet still produce better code (Lauren Goode/Wired)
https://www.wired.com/story/gpt-5-coding-review-software-engineering/

Developers Say GPT-5 Is a Mixed Bag
Software engineers are finding that OpenAI’s new GPT-5 model is helping them think through coding problems—but isn’t much better at actual coding.

🔧 Large-Scale Code Analysis
Process entire codebases with 75,000 lines including source files, tests and documentation for complete system understanding. Claude can identify cross-file dependencies and suggest improvements accounting for complete system design.
📄 Document Synthesis Capabilities
Analyze hundreds of legal contracts, research papers or technical specifications while maintaining full context across documents. Process extensive document sets and analyze relations…

Anthropic says OpenAI staff used Claude Code ahead of GPT-5 launch and it will continue providing API access to OpenAI for benchmarking and safety evaluations (Mayank Parmar/BleepingComputer)
https://www.bleepingcomputer.com/news/arti…

From Articles to Code: On-Demand Generation of Core Algorithms from Scientific Publications
Cameron S. Movassaghi, Amanda Momenzadeh, Jesse G. Meyer
https://arxiv.org/abs/2507.22324

From Articles to Code: On-Demand Generation of Core Algorithms from Scientific Publications
Maintaining software packages imposes significant costs due to dependency management, bug fixes, and versioning. We show that rich method descriptions in scientific publications can serve as standalone specifications for modern large language models (LLMs), enabling on-demand code generation that could supplant human-maintained libraries. We benchmark state-of-the-art models (GPT-o4-mini-high, Gemini Pro 2.5, Claude Sonnet 4) by tasking them with implementing a diverse set of core algorithms dr…

As much as I like to hate on #VibeCoding and #LLM s , #claude code helped me ship a bunch of fixes in a #golang codebase I was totally unfamiliar with, within a few hours. I guess it's a case of having the proper experience to steer the tool in the right direction and avoid hallucinations. A power hammer in the hands of an experienced programmer, a rubber hammer in the hands of a novice.

Did you know? You can run #ClaudeCode with any other #LLM, for example with kimi-k2, gemini, and grok - all together 😉 🚀
👉 https://

Sources: Anthropic's revenue hit a pace of $4B/year, up almost 4x from the beginning of 2025; Cursor maker Anysphere has poached two of Claude Code's leaders (The Information)
https://www.theinformation.com/articles/an

Anthropic Revenue Hits $4 Billion Annual Pace as Competition With Cursor Intensifies
Anysphere, the maker of Cursor, an artificial intelligence-powered coding app, has hired two leaders of Anthropic’s competing coding product, Claude Code, according to one of the leaders. The move is notable because Cursor relies on AI from Anthropic to power its app and is one of Anthropic’s ...

💡 Both tools solve the “oops” moment when Claude Code
Changes go wrong

https://github.com/RonitSachdev/ccundo
https://