Claude Sonnet 5.5

• <p><strong><a href="https://www.anthropic.com/claude-sonnet-5-5">Claude Sonnet 5.5</a></strong></p> New Sonnet model from Anthropic today. They say it "runs 30%+ faster, and costs up to 30% less for most work" - it's priced the same as Sonnet 5 but appears to beat it on every benchmark, and should b

Holo4: powering generalist computer-use agents

15 Best AI Platforms for Content Creators in 2026

• 15 AI platforms compared for content creators in 2026: Higgsfield, Canva, Descript, HeyGen, and more. Pricing, features, and where each one falls short.

OpenAI halts frontier-model training amid string of agent misalignment incidents

Claude quietly moved from assistant to autonomous operator by breaking a physics record for under $2k, while a free tool now clones voices in 646 languages on your own machine.

• Claude solved a nine-loop particle physics calculation mostly unsupervised, beating the previous eight-loop record, running on 96 CPUs for a week at $1-2k total cost
• VoiceStudio, a free open-source tool with 37k GitHub stars, clones voices and dubs video into 646 languages locally (vs ElevenLabs' 32 languages, cloud-based, per-character pricing)
• Anthropic launched a plugin portal for Claude as MCP (Model Context Protocol) usage jumped 110x, signaling rapid ecosystem buildout
• A new detector identifies AI-written blog posts with 98% accuracy based on structural patterns alone, meaning paraphrasing tools won't evade it
• A survey of 600 scientists found AI saves them nearly 7 hours per week, quantifying real productivity gains
• DuoNeural released a 9B parameter open cybersecurity model built specifically for terminal use and agentic tool-calling

2026 in LLMs (so far)

• <p>On Friday I gave the closing keynote at the <a href="https://www.wearedevelopers.com/world-congress-north-america">WeAreDevelopers World Congress North America</a> in San Jose. I tied together the key trends from the past year into a chronological exploration of everything that happened in 2026.

TypeSafe AI's Jev launches a new 'System One' model category for fast, cheap structured decisions—triggering rapid open-source alternatives like Laya and Stanford/Nvidia's CLM-8B.

• TypeSafe AI released Jev on Sept 15: a non-generative model returning typed decisions (e.g. 'billing' vs text) in 70-500ms at $0.042/M input tokens with no output-token cost.
• Jev claims 40-200x lower latency and 444.6x cost reduction vs frontier LLMs for classification/routing/scoring tasks.
• Open-source alternative Laya (421M params, ModernBERT-based) offers a Jev-compatible API and ~33ms inference on a Tesla T4.
• Stanford/Nvidia's CLM-8B uses contrastive embeddings instead of classification heads, running up to 9x faster than Jev in zero-shot tests but sometimes with lower accuracy (CLM: 81.6%/87.6% vs Jev: 71.1%/83.1% on coding-judge benchmarks).
• Emerging pattern: 'Jev-as-judge' replaces expensive LLM-as-judge—generative models produce candidates, cheap decision models rank/route them, escalating only low-confidence cases to LLMs or humans.
• Trained via RLCD (Reinforcement Learning for Calibrated Decisions) so confidence scores are meaningful—0.52 vs 0.998 can trigger different handling logic.

Qualcomm's new chip can run 30B-parameter MoE models on phones, but agentic AI still hits connectivity and battery walls, while Anthropic's job-loss forecasts draw pushback from economists.

• Qualcomm's Snapdragon 8 Elite Extreme Gen 6 runs 30B-parameter Mixture-of-Experts models on-device by swapping 'experts' between flash storage and RAM, a 7x jump from the ~4B models on Apple's A20 Pro
• Qualcomm VP Vinesh Sukumar says raw chip power isn't the bottleneck for agentic phone assistants anymore — connectivity, battery life, and orchestration across devices are
• Anthropic's economic model predicts up to 13.5% of knowledge workers displaced by 2030 in its 'extreme' AI scenario, but economist Julius Probst calls the GDP-consumption assumptions shaky given wealth concentration
• MoE architecture (used by DeepSeek V4, Kimi K2.6, Qwen3.6-35B-A3B) is becoming the standard path to running larger effective models on constrained hardware
• Runway shipped an MCP integration letting Claude Opus 5.5 directly trigger Gen-4.5, Seedance 2.5, GPT Image 2, and Kling for video/image generation — direct relevance for creative pipelines
• Microsoft launched a Copilot 'superapp' merging chat, coding, and agents into one interface; Meituan released LongCat-2.5-Preview, a 1.6T-parameter model with 1M token context

Experts worry about Nvidia's AI chip sales in China and influence over Trump

CrowdStrike's product chief says AI agents don't just attack faster, they chain tactics into novel attack patterns that break human-reasoning-based defenses.

• AJ Shipley (CrowdStrike CPO) argues AI agents 'reason differently' than humans, combining known TTPs into unfamiliar attack chains that 20-30 year old security architectures weren't built to detect
• Entry-level SOC jobs are shifting up-stack: AI now handles tier-1 alert triage and initial evidence-gathering, so 'entry-level' will mean tier-2/tier-3 work instead
• Shipley frames this as a productivity multiplier, not headcount replacement — same analyst count expected to do 10x-100x more work with AI tooling
• CrowdStrike's proposed fix for AI non-determinism: ensemble of models 'voting' on outcomes plus input-controlling harnesses, similar to old multi-engine malware sandboxing
• 3-5 year bull case: 'red models' auto-discover vulnerabilities, 'blue models' auto-remediate them across endpoint/cloud/SaaS/network before exploitation
• Sponsor segment (Coder) pushes concept of a dedicated 'AI Operating Layer' sitting outside the app/data/compute stack for governance and visibility