Karthic Rao

🚀 Introducing Agent Panel at @IncredibleDevHQ: An observability platform for optimizing the control flow, performance, token usage, and correctness of LLM/AI agents! 🔗 github.com/IncredibleDevH… Built-in @rustlang, the first release of Agent Panel currently features an AI gateway that provides seamless access to 100+ LLMs across 20+ platforms, including @OpenAI GPT-4o, @GoogleAI Gemini 1.5 Pro latest, @AnthropicAI Claude 3.5, @MistralAI, @cohere, @GroqInc, @perplexity_ai, and more. With every release, future versions will inch towards different aspects of addressing the complexities of managing multi-agent systems involving numerous function calls and invocations. Key features: • API-first approach for easy integration • Multi-platform support in a single gateway • Simplified configuration and usage • Option for streaming response • Written in Rust, blazingly fast! Getting started is a breeze: 1. Clone: github.com/IncredibleDevH… 2. Build with Cargo 3. Configure your `config.yaml` 4. Run the binary Roadmap 🛣️ • Long-term centralized memory • Distributed agent state resumption • Time travel debugging & graph visualization • API function calling decision-making • Prompt optimization mode • Flexible function calling with lambda and Kubernetes support for deploying functions Get involved✨ • Star the repo ⭐ • Join Discord: discord.gg/DETRMEVqQ6 #AITools #OpenSource #Rust #AIEngineering #AIAgents #LLM @googledevs

Introducing Hyperagents: an AI system that not only improves at solving tasks, but also improves how it improves itself. The Darwin Gödel Machine (DGM) demonstrated that open-ended self-improvement is possible by iteratively generating and evaluating improved agents, yet it relies on a key assumption: that improvements in task performance (e.g., coding ability) translate into improvements in the self-improvement process itself. This alignment holds in coding, where both evaluation and modification are expressed in the same domain, but breaks down more generally. As a result, prior systems remain constrained by fixed, handcrafted meta-level procedures that do not themselves evolve. We introduce Hyperagents – self-referential agents that can modify both their task-solving behavior and the process that generates future improvements. This enables what we call metacognitive self-modification: learning not just to perform better, but to improve at improving. We instantiate this framework as DGM-Hyperagents (DGM-H), an extension of the DGM in which both task-solving behavior and the self-improvement procedure are editable and subject to evolution. Across diverse domains (coding, paper review, robotics reward design, and Olympiad-level math solution grading), hyperagents enable continuous performance improvements over time and outperform baselines without self-improvement or open-ended exploration, as well as prior self-improving systems (including DGM). DGM-H also improves the process by which new agents are generated (e.g. persistent memory, performance tracking), and these meta-level improvements transfer across domains and accumulate across runs. This work was done during my internship at Meta (@AIatMeta), in collaboration with Bingchen Zhao (@BingchenZhao), Wannan Yang (@winnieyangwn), Jakob Foerster (@j_foerst), Jeff Clune (@jeffclune), Minqi Jiang (@MinqiJiang), Sam Devlin (@smdvln), and Tatiana Shavrina (@rybolos).

Eric Schmidt says the 10x advantage is no longer execution. It is defining what counts as success. A programmer writes a spec and an evaluation function, runs it at 7pm, and wakes up to what was invented overnight. The advantage now belongs to whoever can specify the problem precisely. The rest will be automated.

Prediction came true faster than expected 😂 Durandhar effect is real… trust issues level max 😂😂

Models are getting so clever, it's almost scary.

🚨: A petri dish of human brain cells just learned to play DOOM

This week Cloudflare shamelessly cloned and released the product of one of their main competitors. I'm beginning to wonder - if AI can rewrite software from scratch without using any of the original code, do software licenses even mean anything anymore? john.onolan.org/open-source-in…

@robustus Code itself will go away in favor of just making the binary directly. The next step after that is direct, real-time pixel generation by the neural net.

It also suggests that alignment work should focus more on reward hacking and goal misgeneralization during training, and less on preventing the relentless pursuit of a goal the model was not trained on. Read the full paper: arxiv.org/pdf/2601.23045

10x engineer is an outdated idea. With AI, we’re seeing 100x engineers. I dug into @steipete OpenClaw git history: • 6,933 commits • 68 days • One developer Do the math: - 102 commits/day - One commit every 8 minutes - ~8k net new lines per day - ~550K lines of surviving production code This is the new frontier of development.

AI can make work faster, but a fear is that relying on it may make it harder to learn new skills on the job. We ran an experiment with software engineers to learn more. Coding with AI led to a decrease in mastery—but this depended on how people used it. anthropic.com/research/AI-as…

I'm keeping two Claude windows open. One making changes to the source code. The other helping me to plan future changes. This uses my time better since I spend a lot less time waiting for Claude to finish. The two Claudes are in two different directories, both have git repos of the project. The planning directory has rules that prevent source code changes. Only planning is allowed. I manually pull and push changes between the two.

Anthropic gave OpenAI such an easy lay-up to dunk like this. Competition is beautiful.

I’m spending more time thinking through harder problems which I previously never managed to get to! Previously, execution used to consume most of the time. This is freeing up more time to spend on the fun part, critical thinking to solve hard problems in novel ways!

This illustrates an aspect of AI that I hadn't thought about till now: it cuts through bureaucracy. If a big organization is paralyzed by indecision, AI doesn't care. It will happily generate a version 1. And that becomes the starting point, because there is no other version 1.

I was inspired by this so I wanted to see if Claude Code can get into my Lutron home automation system. - it found my Lutron controllers on the local wifi network - checked for open ports, connected, got some metadata and identified the devices and their firmware - searched the internet, found the pdf for my system - instructed me on what button to press to pair and get the certificates - it connected to the system and found all the home devices (lights, shades, HVAC temperature control, motion sensors etc.) - it turned on and off my kitchen lights to check that things are working (lol!) I am now vibe coding the home automation master command center, the potential is 🔥.And I'm throwing away the crappy, janky, slow Lutron iOS app I've been using so far. Insanely fun :D :D

At the Bay Area Iceberg Meetup, MinIO's @hackintoshrao shared a big milestone: MinIO has launched AIStor Tables, bringing native Apache Iceberg support directly into AIStor to unify enterprise data for AI. See Karthic & @BrennaBuuck's full talk here: youtu.be/Rulhj-EBWko?si… #AI #Iceberg #Catalog

MinIO's @BrennaBuuck & @hackintoshrao take the stage at the October 2025 Bay Area @ApacheIceberg Meetup, exploring what it takes to make Iceberg work for AI. youtu.be/Rulhj-EBWko?si… #ApacheIceberg #AI

@koolhead17 That’s why you own 50%of the company🤪

As a founder, you are last man standing. The sooner you realise this, running the journey becomes easier.

From Data Swamps to Reliable Data Systems: How Iceberg Brought 40 Years of Database Wisdom to Data Lakes. blog.min.io/how-iceberg-br… @hackintoshrao

We are very excited to announce the launch of MinIO Academy, a centralized education hub that empowers IT professionals with the skills and training needed to master MinIO AIStor. Learn more about all that #MinIOAcademy has to offer here: min.io/press/minio-la…