Latest

TL;DR - Muse Spark 1.1 is a closed-weight AI model from Meta. You access it through the Meta Model API at $1.25 per million input and $4.25 per million output tokens. - New devs get $20 in starter credits before billing kicks in. US-only during public

Key Takeaways - Noma Labs dropped details on GitLost, a prompt injection that tricks GitHub's preview Agentic Workflows into dumping private repo data through public issue comments. - Attacker needed no coding skills or special access — The Register confirmed this. - The vulnerability cannot be completely fixed in

Key Takeaways - GPT-Live-1 handles simultaneous audio. Listens while talking, no walkie-talkie gaps - Realtime API updates add SIP telephony, MCP tool connections, image input, and two new voices (Cedar, Marin) - Mid-sentence corrections actually register now. You redirect without restarting the whole exchange - Gemini Live shipped full-duplex first.

Key Takeaways - DeepSeek-V4-Pro is performing at a competitive level with other leading models in coding benchmarks. - Per-token pricing runs significantly lower than equivalent frontier models, according to this MindStudio breakdown. - Open weights. Recently released. Grab the weights, run your own inference, skip the metered API if that&

TL;DR - A limited number of silent concepts fire simultaneously inside a private internal zone that Anthropic calls J-space, and the model never writes most of them down. - J-space self-organized during training. Nobody coded it. Takes up a small portion of Claude's activation capacity. - Anthropic&

Key Takeaways - GLM-5.2 delivers near-Claude Opus coding quality at roughly 85% lower cost per token, per MindStudio's analysis. - Artificial Analysis ranks it the #1 open-weight model on their Intelligence Index, with a weighted cost per task of $0.42 versus $0.83 for GPT-5.5

TL;DR - Alibaba's internal ban on every Anthropic product hits July 10, 2026. Claude Code got slapped with a "high-risk" label over what the company calls "back-door risks" following a "comprehensive evaluation." - Anthropic fired first, telling U.S. senators that