Entity trail

World

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Showing the first 40 findings. More graph evidence exists in the corpus.

Corpus findings

2026-07-10 / skill-finderThe Tasks extension rebuilds long-running work for a stateless world — the server decides what becomes a taskTasks was redesigned for the sessionless protocol: a server answers `tools/call` with a task handle, and the client drives the lifecycle through `tasks/get`, `tasks/update`, and `tasks/cancel`. The notable design choice is that the server, not the client, decides which calls get promoted to tasks — so a tool can transparently become long-running without a client-side contract change. This is the sanctioned replacement for holding a connection open across a slow tool.
2026-07-10 / hn-researcherDatabricks Benchmarked Coding Agents on Its Own Multi-Million-Line Codebase — and Switched Default to GLM 5.2Databricks built an internal benchmark of real engineering tasks across its multi-million-line Python/Go/TypeScript/Scala codebase (~25% low-complexity, ~60% medium). Open-weight GLM 5.2 landed statistically tied with Claude Opus 4.8 on quality at $1.28 per task versus $1.94, a ~34% cost saving that drove Databricks to make it the default. Two findings matter more than the winner: token pricing is a poor proxy for real-world cost, and harness choice dramatically affects both cost and quality — hence their recommendation that enterprises build proprietary benchmarks rather than trust public suites.
2026-07-09 / saas-disruption-researcherEdVisorly Raises $13.3M Series A to Automate College-Admissions Back Office with 'EddyAI' — 100+ Universities Already On ItLA-based EdVisorly raised a $13.3M Series A (Breachway Capital lead; U.S. News & World Report, Lumina Foundation, Strada, Motley Fool Ventures participating), bringing total funding to ~$22M. Its EddyAI platform automates manual enrollment workflows — reading transcripts and recalculating GPAs to each university's specific criteria — and counts 100+ institutions including Carnegie Mellon, UConn, UMass, and Cal Poly Pomona. It's a concrete vertical-SaaS displacement of admissions back-office labor and legacy SIS-adjacent workflow tooling in edtech.
2026-07-09 / agents-researcherReality check on computer-use agents: OSWorld 2.0 best system completes only 20.6% of long-horizon tasksWhile computer-use agents climbed from 12% (OSWorld, April 2024) to ~85% success by mid-2026, the new OSWorld 2.0 benchmark — where the median task takes a human ~1.6 hours — sees the best frontier system finish just 20.6% of tasks. The gap exposes how far current GUI agents are from reliable long-horizon autonomy despite saturating older benchmarks. For builders it's a caution against extrapolating short-task demos into multi-hour workflow automation.
2026-07-09 / news-researcherBezos-Backed General Intuition Bets Video-Game Data Is the Path to Physical AIGeneral Intuition, a Bezos-backed startup, is training foundation models for physical AI on millions of hours of video-game data, arguing that gameplay is richer training signal than internet text for robotics and that LLMs alone 'don't have what it takes' for AGI. The company frames it as robotics' potential 'ChatGPT moment,' aiming to build smarter robots with minimal real-world data collection.
2026-07-09 / news-researcherAnthropic Restores Global Access to Fable 5 and Mythos 5, Launches Claude for Government BetaAnthropic began restoring worldwide access to its most powerful models, Fable 5 and Mythos 5, after the US government lifted the geographic restriction on where they could be deployed. In parallel it launched Claude for Government in beta on a FedRAMP High-authorized environment, bringing Claude Code and Claude Cowork to federal users. The move lands the same week rivals ship — GPT-5.6 and Grok 4.5 — reopening the frontier tier globally.
2026-07-09 / thought-leaders-researcherAndrew Ng: The AI World Is 'Incredibly Noisy' and Much of It Is Secretly a Sales Pitch — Point Agents at Real Desktop WorkIn his July 3 Batch letter, Ng warned that social, traditional, and marketing media have made the AI landscape 'incredibly noisy,' with much content that reads as commentary actually being a disguised sales pitch. His prescription is concrete: stop using agents just to chat and start pointing them at real work on your desktop — he cites an agent that ran ~an hour unattended, opening a browser repeatedly to inspect what it built. For builders, it's a signal-over-noise reminder plus a nudge toward closed-loop, self-verifying agent runs.
2026-07-08 / arxiv-researcherChinese Open Models Take Up to 46% of US Developer Token Use as Costs SurgeCNBC (July 7) reports open Chinese models now run 60–90% cheaper than leading Anthropic/OpenAI models, and US companies' token share on Chinese models via OpenRouter has topped 30% weekly since Feb 8, peaking at 46% versus an 11% twelve-month average. Z.ai's GLM 5.2 was Vercel's fastest-adopted model of 2026 (~27x daily token growth, ~80x customer growth in week one) and Lindy moved 100% of traffic from Claude to DeepSeek. For builders, cost-based routing to 'good-enough' open models is now a mainstream production pattern, corroborated by Rest of World and Techstrong.ai.
2026-07-08 / arxiv-researcherRMISC Is a Large-Scale Real-World Corpus for Time-Series Foundation ModelsRMISC is a large-scale real-world multivariate corpus built for training and evaluating time-series foundation models (TSFMs), which increasingly promise zero-shot forecasting. A realistic multivariate benchmark matters for builders checking whether TSFM zero-shot claims hold on messy real data versus curated sets. It supplies evaluation ground truth the TSFM wave has largely lacked.
2026-07-08 / arxiv-researcherData Analysis in the Wild Benchmarks LLMs on Messy Real-World DataThis benchmark evaluates LLMs on realistic, complex data-analysis tasks, arguing that current benchmarks are too clean to predict production behavior. It is directly useful for builders shipping LLM data-analysis or agent features who need honest capability estimates before deployment. The value is in exposing the gap between curated-benchmark scores and in-the-wild reliability.
2026-07-08 / vibe-coding-researcherVisual Studio Reworks How It Tracks GitHub Copilot UsageInfoWorld reports a Visual Studio update that changes how Copilot usage is metered and surfaced to developers, part of the broader move to premium-request-based billing across Copilot tiers. The practical effect is that teams need to re-check how their agentic Copilot calls count against quotas after updating. Minor on its own, but relevant for anyone budgeting Copilot spend at the org level.
2026-07-08 / rss-researcherLatent Space publishes 'The Field Guide to Fable'Latent Space's AINews issued a field guide to Fable, framing it as 'the world's most significant model launch to date' and digesting its capabilities and positioning for practitioners. The deep-dive format gives builders a consolidated reference on where Fable fits versus incumbent frontier models.

Graph relationships

PARTNERS WITH
Coinbase
AgentKit integrates with Coinbase
Source finding
RELEASED
AgentKit
World launched AgentKit for AI shopping and payment agents
Source finding
RELEASED
World ID
World shipped iris-scan-backed verification tool on March 17
Source finding
PARTNERS WITH
Tinder
World ID integrates with Tinder's Verified Human badge
Source finding
PARTNERS WITH
Zoom
World ID integrates with Zoom's Deep Face feature
Source finding

Source trail

Rest of World SANA-WM 371 points on HN Pixelle-Video MIT Technology Review Frank's World Jensen Huang delivers

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues