Meta Llama 5 open weights land with 405B MoE option
The release includes a dense 70B and a sparse Mixture-of-Experts tier.
Why it matters Open 405B MoE weights keep large-model competition from locking entirely behind closed APIs.
Daily Read Book
Google Quantum AI claims 100 logical qubits in error-corrected demo; Meta Llama 5 open weights land with 405B MoE option
31 Aug 2026
A calm edition of the prior 24 hours — 2 briefings in chapters, with links back to sources and the live feed — about 2 min to read.
AI Chapter
Emerging Chapter
Google Quantum AI claims 100 logical qubits in error-corrected demo
Surface-code experiments cut error rates below break-even on a Willow-class processor.
Why it matters Logical qubits past break-even are the error-correction milestone chemistry sims still need more of.