Anthropic ships Claude Opus 4.7 with 1M context window
1M-token context arrives in the flagship tier, closing the spec gap with Gemini and shifting the long-context economics for agentic workflows.
Daily roundup
Models, tools, enterprise AI, research, policy, product. Distilled to what changes positioning.
Sunday, May 17, 2026
Saturday, May 16, 2026
OpenAI Codex lands Microsoft, Stripe as enterprise design partners
Coding-agent product moves from limited preview to named enterprise rollouts, with usage-based pricing and SOC 2 attestation as the procurement unlock.
Friday, May 15, 2026
EU AI Act foundation-model rules finalized, compliance window opens
Final implementing acts published; 12-month compliance window starts for general-purpose model providers. Training-data disclosure is the binding constraint.
Thursday, May 14, 2026
Cursor acquires Windsurf in $2.1B all-stock deal
The two leading AI-native IDEs combine, consolidating the coding-tool layer that VS Code's Copilot integration left to startups.
Tuesday, May 12, 2026
Meta releases Llama 4.5 open weights with 600B sparse MoE
First open-weight release to land near flagship-tier evals; 17B active parameters keep single-node inference viable on H200 boxes.
Monday, May 11, 2026
The White House AI executive orders: what changed, what didn't, what's still moving
The Biden administration's October 2023 AI executive order set a broad framework; subsequent actions and the administration transition have selectively advanced, stalled, or reversed different elements.
Sunday, May 10, 2026
Salesforce Agentforce 2.0 hits GA with usage-based pricing
ServiceNow and HSBC named as launch customers. Per-resolved-conversation billing is the headline; the per-agent floor is what enterprises will actually negotiate on.
Saturday, May 9, 2026
GitHub Copilot vs. internal fine-tuned models in large engineering orgs
Large engineering organizations are increasingly running side-by-side evaluations of GitHub Copilot against internally fine-tuned code models. The results depend heavily on codebase characteristics.
Friday, May 8, 2026
Vercel AI SDK 5.0 ships first-class agent loop primitive
The framework that owns the React-side LLM integration adds a server-side agent runtime. Removes a category of glue code most teams were writing by hand.
OpenAI's safety board restructuring and what it signals about frontier lab governance
OpenAI's board has been restructured twice in 18 months. The pattern reveals something about how frontier AI labs are resolving the tension between governance and commercial velocity.
Wednesday, May 6, 2026
Why 70% of enterprise RAG pilots don't make it to production
RAG system failure in enterprise pilots follows predictable patterns. The problems are rarely the retrieval architecture -- they're organizational and operational.
Tuesday, May 5, 2026
California signs AB-2013 training-data disclosure into law
First US state to mandate training-data summaries for generative AI products sold to California customers. Effective January 2027; the disclosure template is the binding detail.
NIST AI Risk Management Framework 1.1 update and enterprise compliance implications
NIST's AI RMF 1.1 release adds specificity on generative AI risk categories, giving enterprise compliance teams their most concrete US-government guidance to date.
Sunday, May 3, 2026
DeepMind paper sets new bar on long-horizon agent evaluation
New benchmark stresses multi-day task completion with realistic tool calls. Frontier models clear under 30% , the gap between demo and production has a number now.
Saturday, May 2, 2026
Goldman Sachs' internal LLM deployment and what it tells us about enterprise AI ops
Details of Goldman's internal AI stack have surfaced through job postings, conference talks, and industry reporting. The architecture choices reflect constraints specific to regulated financial services.
Friday, May 1, 2026
The EU AI Act's high-risk classification: what it means for US companies selling into Europe
The AI Act's Annex III high-risk categories have specific compliance obligations that many US AI companies are not operationally ready to meet. The 2-year compliance window is shorter than it sounds.
Monday, Apr 27, 2026
Klarna's AI customer service deployment: the real numbers 18 months in
Klarna's widely-cited claim of replacing 700 customer service agents with AI is more nuanced in practice. The actual deployment tells a different story about enterprise AI in customer operations.
Sunday, Apr 19, 2026
The rise of AI gateway middleware: what Portkey, Helicone, and Braintrust are solving
A new middleware category is crystallizing around the need to manage, observe, and optimize traffic across multiple LLM providers. These products solve real production problems that individual API integrations cannot.
Saturday, Apr 11, 2026
Hugging Face's Inference API pricing changes and the open-source model hosting market
Hugging Face restructured its Inference API pricing to reflect actual GPU costs, ending the free-tier economics that subsidized most experimental deployments. The change signals a maturation of the open-source model hosting market.
Saturday, Apr 4, 2026
Cursor AI crosses 1M developers: what the adoption curve says about AI-native tooling
Cursor's growth from zero to 1M developers in under 18 months is one of the fastest adoption curves in developer tooling history. The reasons say something specific about where AI-native products win.
Friday, Mar 27, 2026
LangChain vs LlamaIndex in 2025: which RAG framework is actually winning
After two years of parallel growth, LangChain and LlamaIndex have diverged meaningfully in where they're winning: LangChain for agent orchestration, LlamaIndex for structured retrieval pipelines.
Thursday, Mar 19, 2026
Anthropic's Constitutional AI methodology update and what changed in Claude's training
Anthropic published updated methodology notes on Constitutional AI, detailing how the framework evolved from v1's static principles to a more dynamic, model-assisted approach.
Monday, Mar 9, 2026
Gemini 1.5 Pro's 1M token context window: what works in production
Six months into Gemini 1.5 Pro's 1M context availability, the use cases that justify the cost premium are narrower than the launch coverage suggested.
Tuesday, Feb 24, 2026
Meta's Llama 3 release is reshaping enterprise fine-tuning economics
Llama 3's permissive commercial license and strong base performance have made it the de facto starting point for enterprise fine-tuning programs, with significant downstream effects on model hosting and tooling.
Tuesday, Feb 17, 2026
GPT-4o mini price cuts signal the next phase of inference cost compression
OpenAI cut GPT-4o mini input pricing by 60%, continuing the pattern of aggressive small-model repricing that is reshaping application-layer economics.
No briefs in this topic yet.