https://sakshamgoyal.com/blog Writing about AI, engineering, and building products. en-us https://sakshamgoyal.com/blog/gemini-4-argon https://sakshamgoyal.com/blog/gemini-4-argon Thu, 01 Oct 2026 00:00:00 GMT Gemini 4 Argon allows up to 1M output tokens. Google's coding examples, benchmark results, prices, and rollout explained, including the limits of each claim. ai google models agents https://sakshamgoyal.com/blog/openai-devday-2026 https://sakshamgoyal.com/blog/openai-devday-2026 Wed, 30 Sep 2026 00:00:00 GMT OpenAI DevDay 2026 brought a Sol upgrade, browser access for Agents API, Dots, and code security tools. How five announcements work, with examples. ai openai agents https://sakshamgoyal.com/blog/jev-system-one-model https://sakshamgoyal.com/blog/jev-system-one-model Wed, 23 Sep 2026 00:00:00 GMT How Jev classifies and scores text, what its probabilities mean, and how it compares with LLMs on accuracy, speed, and cost. ai llm agents https://sakshamgoyal.com/blog/agent-patterns-2026 https://sakshamgoyal.com/blog/agent-patterns-2026 Sat, 19 Sep 2026 00:00:00 GMT Eight ways to organize AI agents, from one agent with tools to teams. How they divide work, share information, and handle failures. agents agent-patterns multi-agent orchestration llm tutorial explainer https://sakshamgoyal.com/blog/astra https://sakshamgoyal.com/blog/astra Mon, 24 Aug 2026 00:00:00 GMT GPT-6 Astra handles coding, research, and browser tasks. How its tools, saved context, and checks affect its results and the cost of a task. ai openai agents research https://sakshamgoyal.com/blog/kitesurf https://sakshamgoyal.com/blog/kitesurf Mon, 24 Aug 2026 00:00:00 GMT Cloudflare's Kitesurf runs a browser on Workers. Why it used less CPU and memory but took longer than Chromium in Cloudflare's tests. ai cloudflare agents infrastructure https://sakshamgoyal.com/blog/orchestrator-worker-agents https://sakshamgoyal.com/blog/orchestrator-worker-agents Sat, 15 Aug 2026 00:00:00 GMT A lead agent splits a task among workers, then combines their results. When working in parallel helps and why shared files or decisions can cause problems. agents agent-patterns multi-agent orchestration llm tutorial explainer https://sakshamgoyal.com/blog/gemma4 https://sakshamgoyal.com/blog/gemma4 Fri, 03 Apr 2026 00:00:00 GMT Gemma 4 launched with four downloadable models. How they differ, what hardware they need, and what to check before running them locally. ai open-source on-device google models https://sakshamgoyal.com/blog/claude-code-source-leak https://sakshamgoyal.com/blog/claude-code-source-leak Wed, 01 Apr 2026 00:00:00 GMT A debugging file exposed readable Claude Code source. What the leak revealed about the agent, and why experimental features do not confirm future releases. ai anthropic security engineering https://sakshamgoyal.com/blog/arc-agi https://sakshamgoyal.com/blog/arc-agi Wed, 18 Mar 2026 00:00:00 GMT ARC-AGI-3 tests how efficiently agents learn unfamiliar games. How its scoring works, what the March 2026 paper reported, and why preview scores differ. ai research benchmarks reasoning agi https://sakshamgoyal.com/blog/event-driven-agents https://sakshamgoyal.com/blog/event-driven-agents Mon, 10 Nov 2025 00:00:00 GMT An event-driven agent starts from a message, alert, or schedule. How to handle retries, limit its tools, and protect private data from malicious input. agents agent-patterns llm tutorial explainer https://sakshamgoyal.com/blog/evaluator-optimizer-agents https://sakshamgoyal.com/blog/evaluator-optimizer-agents Tue, 05 Aug 2025 00:00:00 GMT One step writes a draft; another checks it and flags what needs fixing. How this loop works, when it helps, and when to stop. agents agent-patterns multi-agent llm tutorial explainer https://sakshamgoyal.com/blog/alpha-evolve https://sakshamgoyal.com/blog/alpha-evolve Sat, 02 Aug 2025 00:00:00 GMT AlphaEvolve suggests code changes, tests them, and keeps the improvements. How DeepMind used this loop to find better algorithms and speed up its systems. ai deepmind research agents https://sakshamgoyal.com/blog/chatgpt-agent https://sakshamgoyal.com/blog/chatgpt-agent Thu, 24 Jul 2025 00:00:00 GMT ChatGPT Agent can browse, run code, and create files in a virtual computer. How its tools work together, what a completed task looks like, and what to check. ai openai agents https://sakshamgoyal.com/blog/single-agent-tools https://sakshamgoyal.com/blog/single-agent-tools Sat, 05 Jul 2025 00:00:00 GMT One agent chooses a tool, checks the result, and decides what to do next. How the ReAct loop works and where to set limits. agents agent-patterns llm tutorial explainer https://sakshamgoyal.com/blog/agent-handoff-pattern https://sakshamgoyal.com/blog/agent-handoff-pattern Tue, 25 Mar 2025 00:00:00 GMT An agent hands the conversation to a specialist who replies directly. How to pass the right details and avoid losing track of the request. agents agent-patterns multi-agent llm tutorial explainer https://sakshamgoyal.com/blog/agent-router-pattern https://sakshamgoyal.com/blog/agent-router-pattern Fri, 20 Dec 2024 00:00:00 GMT A router sends a request to the right specialist or model. How to choose a route, handle unclear requests, and check whether the user gets an answer. agents agent-patterns multi-agent llm tutorial explainer https://sakshamgoyal.com/blog/hierarchical-agent-teams https://sakshamgoyal.com/blog/hierarchical-agent-teams Tue, 10 Sep 2024 00:00:00 GMT A manager sends work to team leads, who assign it to specialist agents. How this extra layer works and when it helps. agents agent-patterns multi-agent orchestration llm tutorial explainer https://sakshamgoyal.com/blog/supervisor-worker-agents https://sakshamgoyal.com/blog/supervisor-worker-agents Mon, 15 Jul 2024 00:00:00 GMT A supervisor assigns tasks to specialist agents, checks their results, and decides what happens next. How the team works and where it can get stuck. agents agent-patterns multi-agent orchestration llm tutorial explainer