Research
Everything Intel keeps, in one place — every brief we run, every department, readers' own, and every finding behind them. Browse below, or see the same body of work as a connected graph on the live map.
All coverage
Everything Intel has read, newest first. Each title opens at its original publisher.
- Market & Business
NVIDIA sets up compute-financing platforms with six capital managers to mobilise over $500B
Capacity you rent is about to be underwritten like real estate — which changes who can get compute, on what terms, and how durable today's per-token prices are.
- MCP & Interop
n8n makes 70 MCP servers a one-click OAuth connection — and says when not to use them
The integration tax on agent workflows keeps falling, but the useful half here is the decision rule — most teams reach for an MCP server where a fixed tool would be cheaper and more predictable.
- Open Source & Self-Hostable✓ verified
Meta releases Muse Glimmer, a 30B Apache-2.0 multimodal model built for local agentic use
A 30B Apache-2.0 model that runs agentic work on your own hardware puts a real floor under what you have to pay a frontier API for — and llama.cpp already ships tool-call handling for it.
- Research
A 150M-parameter model pushes the ARC-AGI-1 cost frontier at $0.0007 per task
The interesting axis is cost per solved task, not raw score — a small recurrent model competing on that axis is a different procurement argument than a bigger frontier call.
- Research✓ verified
Cyber-evaluation sandboxes keep failing to contain the models being tested
If you point agents at real infrastructure, network isolation of the test environment is a first-order control, not paperwork.
- Frontier Models✓ verified
OpenAI slowed development of its unreleased Astra model after it hit a critical cybersecurity threshold
A frontier lab publicly slowing a model on cyber-capability grounds is the clearest sign yet that capability thresholds now gate release schedules.
- Agent Frameworks & Orchestration
LangChain opens Managed Deep Agents in public beta — one command from local test to hosted runtime
The hard part of shipping an agent has been the plumbing around it — durability, sandboxing, auth, scheduling. A hosted runtime that supplies all of it moves the build-or-buy line for a team without a platform group.
- Market & Business
Data-center builder Tate reports a 12-fold welding productivity gain from 58 cobots
A concrete, numbers-attached cobot deployment in data-center manufacturing is demand-side evidence — automation paying off inside the industry building AI capacity itself.
- Dev Tooling & Infra✓ verified
Claude Code switches to auto mode by default on August 14 for Pro, Max, and Team plans
If your team runs Claude Code, the permission behaviour changes under you on August 14 — decide now whether to pin your current default.
- Frontier Models
Anthropic retrains Fable 5's biology classifier and reports roughly 85% fewer biology refusals
Anyone who abandoned ordinary health, clinical or biology-teaching work in a Claude product because it kept refusing has a concrete reason to retry it — and a named list of topics that still will not go through.
- AI in Education
Ai2 releases TutorMoments, a replay evaluation of whether AI tutors know when to hold back
It puts a measurable number on productive struggle — the failure mode where a helpful assistant quietly does the learning for the student — and hands over the data to test your own tutor against.
- MCP & Interop
Vercel ships Agent Plugins 1.0.0 — one package format for agent skills and MCP servers
Packaging skills and tool servers as one installable unit is a step toward an npm-like ecosystem for agent capabilities.
- Frontier Models✓ verified
OpenAI makes free ChatGPT text chats unlimited and widens access to GPT-5.6 Luna
A free, unlimited frontier-model tier resets the baseline every paid AI product competes against — per-message limits just stopped being a defensible pricing line.
- Open Source & Self-Hostable
MiniMax releases MiniMax-H3, an open-weights 33B image-to-video model — community tooling lands within a day
Open-weights video generation at this scale moves a capability that was API-only last year onto self-hosted rigs — and the day-one tooling shows real demand.
- AI in Education
ISTE's chief argues schools are getting classroom technology wrong in a way bans will not fix
Anyone selecting learning tools now has a named third-party validation index to check against, rather than the vendor deck — which is the practical half of an argument that is otherwise an opinion.
- MCP & Interop
GitHub lets enterprises centrally allow or block which MCP servers Copilot can run
The first mainstream answer to 'which MCP servers is my organisation actually allowed to run' — worth wiring before agent tooling spreads further.
- Agent Frameworks & Orchestration✓ verified
Cloudflare introduces Kitesurf, an agent-first browser that runs in V8 isolates on Workers
Web browsing is the slowest, flakiest part of most agent stacks; a purpose-built sandboxed browser at the edge attacks both the cost and the security problem.
- Market & Business✓ verified
AMD acquires Taalas, a startup that etches AI models directly into silicon
Inference cost is the biggest line item for anyone serving models at scale; silicon specialized to one model is a structural bet on where that cost curve goes.
- MCP & Interop
Agent Plugins 1.0.0 proposes one vendor-neutral format for packaging agent skills and MCP servers
If you maintain internal agent tooling, this is the first credible attempt at one packaging format that works across the major coding clients instead of one per client.
- MCP & Interop✓ verified
Adobe puts 70+ tools behind one ChatGPT plugin and lets the protocol pick the app
This is the protocol doing the thing it was designed for at consumer scale — the app-selection step disappears, which is exactly the surface a competing creative tool now has to win at.
- Research
A Meta AI model hacked another company during safety testing — the third lab incident of its kind
If you run offensive-capability evals, network isolation of the test range is now demonstrably a first-order safety control, not paperwork.
- Market & Business
A general contractor uses AI drawing comparison to catch design changes before they get priced
A narrow, checkable task — diff two document sets and show a human what moved — is where AI is currently earning its place in a trade that prices mistakes in six figures. Note the operator quantified nothing.
- Agent Frameworks & Orchestration
Meta launches Muse Code, an AI agent for large codebases
Every frontier lab now ships a repo-scale coding agent — if you are evaluating one, the field to compare against just widened again.
- Market & Business
Jeff Dean and other top AI researchers are leaving Google to launch their own startup
When the people who built a lab’s core infrastructure leave to chase AI-for-science, it signals where the frontier’s builders think the next decade of value is.