Zuckerberg Ate His Own Open-Source Manifesto
Two years after writing 2,000 words about why open-source AI is the path forward, Zuck launched a locked-down proprietary model. That's not a pivot — it's a 180.
#llm
6 transmissions tagged #llm
Gemma 4's Real Upgrade Isn't the Benchmarks. It's the License.
Google finally dropped the custom Gemma license for Apache 2.0 — and that boring legal detail might matter more than any benchmark number.
Zero Python, Zero Frameworks, 397 Billion Parameters: Flash-MoE Is Kind of Absurd
Someone ran a 397B parameter model on a MacBook Pro using raw C and Metal shaders. Here's why that's actually impressive and not just a stunt.
Your 4090 Is Probably Memory-Bound
On a 24 GB card, single-GPU LLM inference is usually constrained by memory traffic and KV cache growth long before raw math throughput becomes the limit.
Agent Evals Need Incident Budgets, Not Just Accuracy
Why production agents should be evaluated like distributed systems: trajectory-level scoring, failure taxonomies, and explicit incident budgets.
Claude Sonnet 4.6: Opus-Level Smarts, Sonnet-Level Price Tag
Anthropic's Claude Sonnet 4.6 delivers full upgrades across coding, computer use, and long-context reasoning — at the same price as its predecessor.