Open AI infrastructure releases are increasingly targeted at specific production bottlenecks. Mistral released a 3B policy-adaptive multimodal safety model deployable on one 16GB GPU, while Cursor open-sourced an NVL72 MoE training kernel that increased end-to-end throughput 41% in its stack. Both performance claims are vendor-reported, and Cursor's gain is hardware- and workload-specific.
No must-reads have been published yet.
Service-firm AI use rose to 40% from 25% and manufacturing use to 26% from 16%, yet surveyed firms reported very few AI-driven layoffs and some hiring for AI skills.
Meta has released Muse Spark 1.2. It's their third release in four months and scores 54 on the Artificial Analysis Intelligence Index, significantly improving agentic knowledge work capabilities over prior releases and putting Meta next to SpaceXAI in a tie for third place…
Meta has released Muse Spark 1.2. It's their third release in four months and scores 54 on the Artificial Analysis Intelligence Index, significantly improving agentic knowledge work capabilities over prior releases and putting Meta next to SpaceXAI in a tie for third place amongst US labs Muse Spark 1.2 (xhigh) lands at 54, up 3 points from Muse Spark 1.1 (51) and 11 points from Muse Spark 1.0 (43, April). It enters effectively tied with GPT-5.5 (xhigh, 55) and Grok 4.5 (high, 54), narrowly behind current frontier models Claude Opus 5 (max, 61), Claude Fable 5 (max w/ fallback, 60), GPT-5.6 Sol (max, 59), and Kimi K3 (max, 57) Congratulations to @AIatMeta, @finkd, and @alexandr_wang on the release! Key Takeaways: ➤ Muse Spark 1.2 gets closer to the frontier on agentic knowledge work. At Muse Spark 1.1's launch, we noted agentic knowledge work as its clearest gap; Muse Spark 1.2's gains help to close this. Its GDPval-AA v2 Elo rose 260 points to 1631, #5 among all models we have benchmarked and ahead of Claude Opus 4.8 (max, 1588). Terminal-Bench 2.1 gained 2 points (78% to 80%), and Tau3-Bench Banking rose 2 points (25% to 27%) ➤ Among the most cost-efficient models at its intelligence level. Muse Spark 1.2 costs $0.40 per Intelligence Index task at Meta's unchanged $1.25/$4.25 per 1M token pricing, with only Grok 4.5 (high, $0.37) and GPT-5.6 Sol (medium, $0.39) cheaper in its intelligence cluster - GPT-5.6 Terra (max, $0.51), Kimi K3 (max, $0.86), and GPT-5.5 (xhigh, $1.18) all cost more per task. The cost increase over Muse Spark 1.1 ($0.29 per task) is driven by increased token usage per Intelligence Index task ➤ AA-Omniscience abstention rate increases. The score rose from 18 to 22 as the hallucination rate fell 10 points (38% to 28%) and the attempt rate dropped from 82% to 67%. This heavy abstention (not answering questions when unsure) now drives both the low hallucination rate and a lower accuracy (41% to 38%) ➤ Scientific Reasoning results remain largely unchanged. CritPt notably gained 3 points (15% to 18%), while SciCode fell 2 points (58% to 56%), and Humanity's Last Exam fell 1 point (45% to 44%) Other model details: ➤ Context window: 1M tokens, unchanged from Muse Spark 1.1 ➤ Pricing: unchanged from Muse Spark 1.1: $1.25/$4.25 per 1M input/output tokens, with cache hits discounted to $0.15 per 1M ➤ Availability: Meta's first-party API at launch
The benchmark index rose to 54 from 51, led by a 260-point GDPval-AA gain. Its hallucination rate fell to 28% as the attempt rate dropped to 67%, so lower error came with more abstention.
70 posts · 88% confidence
AI infrastructure demand is now visible in revenue and EBITDA, while SpaceX's capital intensity keeps returns unresolved. AMD's Q2 data-center revenue rose 107% to $6.7B; SpaceX's AI revenue rose 247% to $2.6B, with $1.1B in adjusted EBITDA. SpaceX's segment still posted a $1.3B operating loss after $15.8B of quarterly AI capex, leaving return-on-capital risk unaddressed.
The energy shock has become a transmission problem as much as a supply shock. BIS says persistence, importer status, energy intensity and second-round effects now determine policy, while EIA's preliminary week-ended July 31 data show crude stocks up 2.5 million barrels but gasoline and distillate down 1.6 million and 3.5 million, still 7% and 12% below five-year averages. One volatile U.S. week does not establish persistence.
No must-reads have been published yet.
The index rose to -0.5063 for the week ended July 31 from -0.8264, its August 5 vintage. The move signals less benign conditions, while a value below zero still denotes below-average stress.
514 posts · 90% confidence