Di Guardia
-
Inside an LLM: an interactive guide to tokenization, attention, and generation
You ask ChatGPT a question. It answers. In between those two events, something genuinely strange happens, and almost nobody explains it. Your words get chopped up, turned into numbers, multiplied by giant matrices, voted on, and reassembled, one token at…
-
The AI Price War Is Bifurcating the Stock Market
The Shock That Already Happened (And the Recovery That Pretended It Didn’t) On January 27, 2025, Nvidia lost $600 billion in market cap in a single day. A 17% crash on one stock. Broadcom dropped 17%. ASML dropped 7%. The…
-
GPT-5.6 Sol Is Here. The US Government Decides Who Gets to Use It.
OpenAI just shipped GPT-5.6 Sol, its most powerful model yet, running at 750 tokens/sec on Cerebras. The catch: the US federal government decides who gets to use it. Welcome to the gatekeeper era of frontier AI.
-
The $285 Billion Miscalculation: How Chinese AI Is About to Break the American Model
The Gap That $285 Billion Couldn’t Buy Here’s a number that should make every OpenAI and Anthropic investor lose sleep: the United States spent 23 times more than China on private AI investment in 2025. $285.9 billion versus $12.4 billion.…
-
GLM 5.2: The First Frontier AI Model You Can Actually Download
For about three years, the same sentence kept appearing in every AI conversation, like a tic. “Yes, the open models are catching up, but they’re still six to twelve months behind the closed frontier.” It became background noise. A polite…
-
SpaceX Buys Cursor for $60 Billion: The Deal That Reshuffles the Entire AI Coding Tools Market
The headline is real, even if it reads like a Mad Libs entry from 2021: SpaceX is paying $60 billion in stock for Cursor. Not for a satellite constellation, not for a hypersonic engine company, but for a Visual Studio…
-
GPT-5.5 Is Here. The Benchmarks Look Great. Hacker News Still Doesn’t Trust It.
GPT-5.5 is here, and OpenAI wants you to see it as the next serious step toward agentic work on a computer. The launch page pushes hard on one message: better coding, better knowledge work, better scientific research, with GPT-5.4-level latency…
-
Gemma 4: How Google Fit Frontier AI Into 31 Billion Parameters
Google DeepMind dropped Gemma 4 on April 2, 2026. Four model sizes, full multimodal support, Apache 2.0 license, and benchmark scores that make models 20x its size nervous. The smallest variant runs on a Raspberry Pi 5. The largest, a…
-
I Read All 512,000 Lines of Claude Code’s Leaked Source Code
A few weeks ago, I reverse-engineered Cursor by pointing a proxy at it and capturing every API call. That took some effort. This time, Anthropic did the hard part for me. On March 31, 2026, developer Chaofan Shou noticed something…
-
This Week on GitHub: 10 AI Repos That Gained 100,000+ Stars Combined
Two weeks ago I wrote that every hot GitHub repo was an agent. This week? It’s worse. Or better, depending on where you sit. I pulled the full GitHub Trending data for daily, weekly, and monthly rankings on March 26,…