All tags
LLM
Posts about large language models and the decisions around them: which model to use for a given task, how to pair an expensive planner with a cheap executor, and why leaderboard rankings rarely survive contact with your own data. Prices and context limits move every few weeks, so I try to write down the parts that stay true. Vendor specific coverage lives under OpenAI, Anthropic and Google AI.
12 posts
How AI Agents Handle Unstructured Data from Customer Attachments
Extracting data from customer PDFs is harder than it looks. Here are 5 layers every production AI agent needs to handle unstructured documents reliably.
Should We Be Afraid of Anthropic's Mythos?
Anthropic's Mythos model launched with restricted access and cybersecurity headlines. But is the real threat a closed model few can access - or the open-source tools already in the wrong hands?
Top AI GitHub Repos Right Now - What They Tell Us About the State of AI
The 12 most starred AI GitHub repos tell a clear story about where the AI industry is heading and what it means for businesses building with AI.
The Future of AI Startups: What Y Combinator Is Betting On in 2026
Y Combinator’s 2026 Spring startup requests reveal where AI is heading - from AI-native hedge funds to product decision agents.
I used AI as a tax advisor
I used AI while preparing yearly stock market tax documents in Poland. Here are 7 practical rules to use LLMs for tax research safely without treating them as legal advice.
AI in Web Scraping and Web Crawling
After 8+ years in web scraping, learn where AI and LLMs genuinely help — from writing selectors and fallback parsing to deduplication and self-healing scrapers — and what to avoid.
Which AI model is best?. Why it is wrong question to ask?
Stop asking which AI model is best. Learn why LLM comparisons are tricky, and discover practical methods like benchmarks, Chatbot Arena, and custom testing to find the right model for your work.
Plan with Opus 4.5, execute with cheap models - AI costs strategy
Learn how to cut AI costs by up to 25x using a simple two-step strategy: plan complex tasks with premium models like Opus 4.5, then execute with cheaper LLMs. Practical tips for developers and businesses building cost-effective AI systems.
Will AI Replace Developers in 2026?
Despite incredible advances in AI coding tools like Opus 4.5 and agentic workflows, AI won't replace developers. Learn why skilled operators are still essential and how to leverage AI effectively.
Perplexity vs Google AI: Which AI Search Tool Wins in 2026?
Compare Perplexity and Google AI search tools. Explore UX, source quality, features, and output for the best AI-powered search experience.
OpenAI DevDay 2025: 5 Game-Changing AI Updates You Need to Know
Discover the top 6 OpenAI DevDay 2025 announcements: AppsSDK, AgentKit, Sora 2 API, Codex workflows, GPT-5 Pro, and more. Stay ahead with AI innovations!
OpenRouter: Access 400+ LLMs with Single API
Stop managing multiple AI APIs. OpenRouter gives you GPT-4, Claude, Llama & 400+ models through one endpoint. Complete Python guide with code examples.