Mihai Chindriș 🦋

GEPA optimizes LLMs without costly reinforcement learning Moving beyond the slow, costly trial-and-error of RL, GEPA teaches AI systems to learn and improve using natural language.

GEPA optimizes LLMs without costly reinforcement learning venturebeat.com/ai/gepa-opti... via @venturebeat.com

Stop what you are doing and try out GEPA now!

"GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning" presents such elegant ideas by a collection of amazing researchers!

Here is a tldr of how it works:

Claude Code costs up to $200 a month. Goose does the same thing for free. Goose, Block’s open-source AI coding agent, is emerging as a free alternative to Anthropic’s Claude Code, as developers weigh offline control, rate limits, and the rising cost of AI coding tools.

venturebeat.com/infrastructu...

Break Free From AI Lock-In: GitHub Models + cagent | Docker See how Docker cagent integrates with GitHub Models to build and ship multi-agent apps without vendor lock-in.

www.docker.com/blog/configu...

A Day in Gas Town Have you ever thought

www.dolthub.com/blog/2026-01...

Effective harnesses for long-running agents Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

www.anthropic.com/engineering/...

Introducing Gemini 3.1 Pro across the developer ecosystem. It's a smarter model for your most complex coding problems.

Start building with Gemini 3.1 Pro today via the Gemini API in
Google AI Studio, Vertex AI, Gemini CLI, Android Studio and our agentic development platform, Google Antigravity.

Ralph Wiggum as a "software engineer" How Ralph Wiggum went from 'The Simpsons' to the biggest name in AI right now - Venture Beat 😎Here's a cool little field report from a Y Combinator hackathon event where they put Ralph Wiggum to the ...

ghuntley.com/ralph/

11 Tips For AI Coding With Ralph Wiggum by @mattpocock.com

www.aihero.dev/tips-for-ai-...

DEV is joining the @mlhacks.bsky.social family!

Watch Ben, Jess, and Peter talk about the acquisition, and why we're excited for this new chapter of the community.

#MLH #DEVCommunity

Ralph Wiggum: Autonomous Loops for Claude Code The official Claude Code plugin that lets agents work autonomously for hours. When to use it, when not to, and the philosophy behind letting AI fail repeatedly until it succeeds.

Claude Code's official plugin marketplace includes ralph-wiggum: autonomous development loops that let agents work for hours without intervention.

The philosophy: "Better to fail predictably than succeed unpredictably." Prompt engineering becomes the skill.

Deep dive: paddo.dev/blog/ralph-w...

AGENTS.md AGENTS.md is a simple, open format for guiding coding agents. Think of it as a README for agents.

www.aihero.dev/my-agents-md...

exa.ai/blog/people-...

Sentry acquires XcodeBuildMCP Sentry acquires XcodeBuildMCP, an open source MCP server that gives AI agents the ability to build, test, and debug native iOS and macOS apps autonomously.

Sentry acquires XcodeBuildMCP blog.sentry.io/sentry-acqui...

How Ralph Wiggum went from 'The Simpsons' to the biggest name in AI right now Credit: VentureBeat made with Nano Banana Pro on Fal.ai

venturebeat.com/technology/h...

Gemini 3 Deep Think: Advancing science, research and engineering We’re releasing a major upgrade to Gemini 3 Deep Think, our specialized reasoning mode.

blog.google/innovation-a...

Ollama vs Llama.cpp: Complete Comparison Guide - 피카부랩스 블로그 Discover the concepts and differences between Llama.cpp and Ollama, local LLM execution tools, explained at a beginner-friendly level. Understand their relationship, pros and cons, and how to choose t...

peekaboolabs.ai/blog/ollama-...

z.ai/blog/glm-5

www.cmu.edu/tepper/news/...

Trust in the Age of Acceleration How to build and preserve social trust amid AI acceleration and transformative technological change.

www.aicollective.com/trust

The Dawn of the Renaissance Developer Tools change, but the fundamentals endure. As generative AI reshapes how we build software, a familiar trope has re-emerged, the narrative that developers will become obsolete. But if history has taug...

thekernel.news/articles/daw...

Dexa aims to get more out of podcasts with AI-powered search | TechCrunch If you listen to a lot of podcasts, there is a chance you might remember funny tidbits and are wondering… "Wait, who talked about eating fries with

techcrunch.com/2024/02/05/d...

Why Sourcegraph and Amp Are Becoming Independent Companies | Sourcegraph Blog Sourcegraph and Amp are becoming two separate companies, with Dan Adler stepping up as CEO of the code search business while co-founders Quinn Slack and Beyang Liu launch Amp Inc. to focus on frontier...

sourcegraph.com/blog/why-sou...

Join the GitHub Copilot CLI Challenge! Win GitHub Universe Tickets, Copilot Pro+ Subscriptions and $1,000 in Cash 💸 We're excited to announce our newest challenge with GitHub! Running through February 15, the GitHub...

The Github Copilot CLI Challenge ends Sunday!

💰 $1,000 USD
🎟️ Ticket to GitHub Universe
➕ GitHub Copilot Pro+ Subscription
🏆 Exclusive DEV Badge

dev.to/devteam/join...

Tech predictions for 2026 and beyond We’ve caught glimpses of a future that values autonomy, empathy, and individual expertise. Where interdisciplinary cooperation influences discovery and creation at an unrelenting pace. In the coming y...

www.allthingsdistributed.com/2025/11/tech...

Introducing Model Council Today we are launching Model Council, a multi-model research feature that brings several models together for one answer.

www.perplexity.ai/hub/blog/int...

GitHub - Mihai-Codes/opencode-puter-auth: OpenCode plugin for Claude Opus 4.5, GPT-5, Gemini via Puter.com - 500+ AI models with automatic fallback. No API keys needed. OpenCode plugin for Claude Opus 4.5, GPT-5, Gemini via Puter.com - 500+ AI models with automatic fallback. No API keys needed. - Mihai-Codes/opencode-puter-auth

Kept hitting 403 errors trying to access powerful AI models in #OpenCode.

Found the fix: @heyputer.bsky.social's user-app-token flow.

Now it's a plugin - Opus 4.5, GPT-5.2 & multiple other models. One login, no API keys.

npm i opencode-puter-auth

github.com/Mihai-Codes/opencode-puter-auth

Another Year, Another $750,000 to Open Source Maintainers Sentry’s 2025 open source report: $750k committed, $4.5M paid by Pledge members, and why paying maintainers is no longer optional.

Another year, another $750,000 to Open Source maintainers blog.sentry.io/another-year...

Mozilla’s new CEO is doubling down on an AI future for Firefox Anthony Enzor-DeMeo is taking charge of one of the internet’s most important — and most complicated — companies.

Mozilla’s new CEO is doubling down on an AI future for Firefox www.theverge.com/tech/845216/...

Gemini 3 Flash is live. ⚡

It outperforms 2.5 Pro while being 3x faster at a fraction of the cost. The model is now rolling out to:

🔹 Gemini API in Google AI Studio
🔹 Google Antigravity
🔹 Gemini CLI
🔹 Android Studio
🔹 Vertex AI

Get started → goo.gle/3flash