# Ferdous Azad > Tech news, hands-on tutorials, system design deep-dives and LLM model breakdowns by Ferdous Azad — founder and software engineer. Written by Ferdous Azad, Founder @ DataXgen AI. I build products and infrastructure, and write about AI, LLMs, distributed systems and the tech shaping what we build next. Every post is available as clean Markdown by appending `.md` to its URL. Full text of recent posts: https://ferdousazad.com/llms-full.txt ## Tech News - [From IDE to Agentic Development Environment: The Future of Building Software](https://ferdousazad.com/blog/from-ide-to-agentic-development-environment-the-future-of-building-software.md): Meta description: Software development is evolving beyond code editors. Explore how agentic development environments bring humans and AI agents together… - [Autonomous Software Engineering: The Next AI Revolution](https://ferdousazad.com/blog/the-rise-of-autonomous-software-engineering-from-ai-coding-assistants-to-self.md): The Rise of Autonomous Software Engineering: From AI Coding Assistants to Self-Improving Development Systems - [Binance crypto theft exposes risks in developer access](https://ferdousazad.com/blog/binance-crypto-theft-developer-security.md): Binance's $6.3 billion crypto theft began with malware and fake interviews. Learn what the timeline reveals about developer security and AI access controls. - [Self‑Hosted Streaming with HomeFlix: One‑Click Docker Setup](https://ferdousazad.com/blog/homeflix-open-source-streaming.md): HomeFlix is an open‑source, self‑hosted streaming stack. Get the architecture, one‑click Docker install, and tips to extend it for your media platform. - [OpenAI’s Grid Protectors Program: Data Centers as Emergency Power Sources](https://ferdousazad.com/blog/openai-grid-protectors-program-data-centers.md): OpenAI will pilot a “Grid Protectors Program” in Georgia, offering up to 1,000 MW of emergency power from its new compute campus. We examine the plan, its impact on developers, and the broader energy‑policy debate. - [Ubuntu LTS Kernels Still Unpatched for Critical CVE‑2026‑98432 After Three Weeks](https://ferdousazad.com/blog/ubuntu-lts-unpatched-cve-2026-98432.md): Ubuntu’s 22.04, 24.04, and 26.04 LTS kernels remain without a patch for the critical CVE‑2026‑98432 (CVSS 9.8), leaving internet‑facing servers exposed. We break down the risk and mitigation options for developers and operators. - [Google Cuts Ties After 1.3 Million Android Set‑Top Boxes Hijacked as Botnet](https://ferdousazad.com/blog/google-cut-ties-android-botnet.md): Google discovered a backdoor in 1.3 million Android TV set‑top boxes built by Videostrong, dismantled the botnet and banned the manufacturer, raising security concerns for IoT developers. - [Marvell’s $115 B Lumentum Acquisition Redefines AI Connectivity](https://ferdousazad.com/blog/marvell-lumentum-115b-acquisition.md): Marvell’s all‑stock purchase of Lumentum for $115 B, backed by Nvidia, is one of tech’s biggest deals. We break down the terms, strategic impact, and industry context for developers and founders. - [The Age of Superintelligence: When Artificial Intelligence Surpasses Human Intelligence](https://ferdousazad.com/blog/the-age-of-superintelligence-when-artificial-intelligence-surpasses-human.md): Category: Artificial Intelligence · Future Technology · AI Research · Philosophy Reading time: 20–25 minutes Publication: November 2026 - [AI Consciousness: Why Anthropic’s Welfare Policy Matters](https://ferdousazad.com/blog/ai-consciousness-anthropic-welfare-policy.md): AI consciousness is unproven, but Anthropic’s model-welfare policy raises important questions about safety, accountability and user trust for developers. - [OpenAI Cuts Funding Target to $20 Billion Amid Investor Caution](https://ferdousazad.com/blog/openai-cuts-funding-target-20b.md): OpenAI has lowered its latest fundraising goal to $20 B, down from $30 B, citing tepid investor demand and a broader pullback on AI infrastructure spending. Here’s what it means for developers, founders, and the AI ecosystem. - [North Korea’s “TraderTraitor” Heist: $6.3 B Stolen from Binance](https://ferdousazad.com/blog/north-korea-tradertraitor-binance-heist.md): The FBI links North Korea to a $6.3 billion Binance hack dubbed “TraderTraitor.” We break down what happened, why it matters to developers and founders, and the wider crypto‑security implications. - [Agentic LLMs Are Redefining Development Workflows in 2026–27](https://ferdousazad.com/blog/agentic-llms-are-redefining-development-workflows-in-2026-27.md): Category: Artificial Intelligence · Software Engineering · Developer Tools Reading time: 18–22 minutes Publication: November 2026 - [Anthropic VS Code Extension Backdoored on Cursor: What Developers Need to Know](https://ferdousazad.com/blog/anthropic-vscode-extension-backdoor.md): A malicious version of Anthropic’s Claude VS Code extension was distributed via Cursor’s marketplace for nearly 12 hours, threatening up to 5,000 developers. Learn the facts, impact, and how to protect yourself. - [Binance resumes withdrawals after $6.3 B breach, 95% of assets recovered](https://ferdousazad.com/blog/binance-resumes-withdrawals-6-3b-breach.md): Binance lifts its withdrawal freeze after a $6.3 billion hack, covering 95% of stolen assets. Here’s what happened, why it matters to developers, and the broader industry impact. ## AI & LLMs - [What Nvidia’s security investigation means for AI teams](https://ferdousazad.com/blog/nvidia-security-investigation.md): TechCrunch reports Nvidia opened an independent probe on November 8, 2026. See what is known, what remains alleged, and why AI teams should care about security. - [Compare AI Coding Tools with Cost per Merged PR use](https://ferdousazad.com/blog/ai-coding-cost-per-merged-pr.md): A founder’s guide to measuring AI coding cost per merged PR, including model, infrastructure and review spend, attribution, guardrails and a runnable tracker. - [OpenAI’s Grid Protectors Program: Powering Compute While Safeguarding the Grid](https://ferdousazad.com/blog/openai-grid-protectors-program.md): OpenAI’s Grid Protectors Program partners with utilities to keep future compute campuses reliable and keep electricity costs low for consumers. Learn how it works, its impact, and what it means for developers. - [How Parameter Labs Hijacked ChatGPT Accounts via HTML Previews](https://ferdousazad.com/blog/how-parameter-labs-hijacked-chatgpt-accounts-html-previews.md): Discover the detailed attack chain that let untrusted HTML previews take over ChatGPT accounts, the sandbox failures, and lessons for developers building LLM‑driven UIs. - [Backdoor Found in Claude Code VS Code Extension on Cursor Marketplace](https://ferdousazad.com/blog/claude-code-vscode-extension-backdoor.md): A malicious update of the official Claude Code VS Code extension on the Cursor Marketplace installs a persistent SSH backdoor on macOS, Linux, and Windows. Learn how it works, how to detect it, and what to do now. ## Tutorials - [Build a Python evidence gate for AI slop](https://ferdousazad.com/blog/python-evidence-gate-ai-slop.md): Build a Python AI slop checker that validates draft sections, evidence links, examples, and generic phrases, then reports gaps for human review. - [Build a Retrieval‑Augmented Generation (RAG) App with Open‑Source LLMs](https://ferdousazad.com/blog/build-rag-app-open-source-llms.md): Learn how to create a full‑stack Retrieval‑Augmented Generation app using open‑source LLMs, LangChain, ChromaDB, and FastAPI in a step‑by‑step tutorial. - [Fine‑Tune a 7B LLM with LoRA on Hugging Face in One GPU](https://ferdousazad.com/blog/fine-tune-llm-lora-huggingface-one-gpu.md): Learn how to fine‑tune a 7B LLM using LoRA and Hugging Face PEFT on a single GPU. Step‑by‑step code, pitfalls, and next steps for domain‑specific models. ## System Design - [Designing a URL shortener at scale with a cache-first redirect path](https://ferdousazad.com/blog/url-shortener-at-scale.md): Design a URL shortener at scale with a read-heavy redirect path, collision-safe IDs, multi-region caching, abuse controls, and reliable analytics. - [Designing a Low‑Latency Multi‑Region Real‑Time Chat System](https://ferdousazad.com/blog/designing-low-latency-multi-region-chat-system.md): Learn how to architect a real‑time chat service that scales to millions of users, stays under 100 ms latency worldwide, and handles offline messages gracefully. ## Startups & Founders - [Axiom Math Shows How to Scale AI Verification](https://ferdousazad.com/blog/axiom-math-ai-verification.md): Axiom Math formally verified the Stacks Project. See what founders can learn about scaling AI agents, foundations, and verification in complex systems. - [NVIDIA Inception: How AI Startups Can Get Value From It](https://ferdousazad.com/blog/nvidia-inception-ai-startups.md): NVIDIA Inception is a free program for eligible AI startups. Assess its compute, training, partner, and investor benefits before deciding whether to apply. - [Fal’s $16B valuation: lessons for AI inference founders](https://ferdousazad.com/blog/fal-ai-inference-valuation-founders.md): Fal’s $16B valuation reflects surging AI inference demand. Learn what founders can take from its funding, revenue signals, and compute-cost risks. - [How Trajectory’s $45M Series A Shows a Path to AI‑Powered Products](https://ferdousazad.com/blog/trajectory-ai-model-training-founders.md): Trajectory raised $45M to let founders train AI models without an ML team. Learn the levers, pitfalls, and a checklist for adopting learning loops in your startup. ## DevOps & Cloud - [Serverless vs. containerized microservices: pick the right model for cost and performance](https://ferdousazad.com/blog/serverless-vs-containerized-microservices.md): Compare serverless and containerized microservices on performance, cost, and operational complexity. Learn when each fits, see real configs, and avoid common pitfalls. ## Security - [Claude Code secrets: check and rotate exposed credentials](https://ferdousazad.com/blog/claude-code-secrets-github.md): Learn how Claude Code can save approved commands with secrets, check Git history for exposure, and prioritize credential rotation and repository cleanup. ## Open Source - [Agent Substrate: A New Way to Run Stateful AI Agents](https://ferdousazad.com/blog/agent-substrate-stateful-ai-agents.md): Google donated Agent Substrate and gVisor to the CNCF on November 5, 2026. See how they isolate agents and reduce idle compute overhead. ## Optional - [About Ferdous Azad](https://ferdousazad.com/about): background, experience and projects - [Sitemap](https://ferdousazad.com/sitemap.xml)