Entity trail

AI-Q

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

2026-08-22 / arxiv-researcherGenetic Algorithms Evolve Ransomware That Stays Under Behavioral Monitors' Entropy ThresholdsDefenders should know this threat model is now being published: the authors frame ransomware execution as a Search-Based Software Engineering optimization problem rather than the traditional mass-encryption approach that triggers immediate detection. A genetic algorithm optimizes data encryption under a hard constraint on statistical deviation from baseline system activity, targeting the persistence gap where modern attacks aim to stay hidden for hours rather than minutes. The evolved patterns evade behavioral monitors under fingerprinting techniques, which means entropy-threshold and volume-based detection heuristics are the wrong place to anchor ransomware defense.
2026-08-22 / arxiv-researcherShieldZFS Extends ZFS to Detect Rollback and Fork Attacks by a Malicious Cloud Provider on TEE StorageConfidential computing protects code inside Trusted Execution Environments but leaves storage exposed: even with disk encryption a malicious provider can roll back, replay, fork or tamper with disk state, breaking freshness guarantees that stateful applications depend on. ShieldFS is a POSIX-compliant filesystem giving end-to-end integrity and freshness without application changes, representing permissible filesystem states as succinct cryptographic commitments held inside TEEs and replicated in a lightweight trusted registry, with an on-disk write-ahead log and storage pool authenticated by hash chains and an embedded Merkle tree. Implemented as an extension to ZFS, it verifies commitments on read so rollback and equivocation are detected even when the whole I/O stack is untrusted, at performance comparable to existing filesystems.
2026-08-22 / arxiv-researcherConceptGuard Tests Unlearning on Dual-Use Concepts and Finds Current Methods Cannot Separate Harmful From Benign UseExisting unlearning benchmarks use disjoint forget and retain sets of independent facts and score simple factual recall, which never tests the real requirement: removing a harmful application while keeping the same knowledge available for legitimate use. ConceptGuard introduces dual-use concepts, where forget and retain sets are explicitly complementary in concept usage rather than in subject matter, making evaluation intent-sensitive and measured at the concept level instead of over sparse facts. Current unlearning techniques perform poorly under this framing, showing weak contextual separation, strong forgetting-utility trade-offs and inconsistent concept-level control across methods; the dataset is publicly released.
2026-08-22 / arxiv-researcherBrain Researcher Raises Agent Tool-Selection Accuracy From 23.3% to 93.6% by Enforcing Analytic RulesAI agents can run scientific analyses but reproduce human failure modes including selective analysis, premature declarations of success, and optimizing imperfect criteria, so an output only becomes a defensible claim after alternatives are weighed. Brain Researcher is an agentic harness that operates inside a neuroimaging researcher's own computational environment under explicit rules for admissible analyses, required checks and claim scope. Across seven models it raised first-choice tool-selection accuracy by 70.2 percentage points (23.3% without it to 93.6% with it) and verifiable grounding from 4.6% to 22.0%, with multiverse analyses exposing analytic-choice sensitivity and review classifying claims as accepted, qualified, revised, blocked, rejected or deferred.
2026-08-22 / arxiv-researcherRelation Replaces Attention as the Token-Mixing Primitive and Beats MHA on Validation NLL at 10M, 30M and 100MRelation reorders the attention computation, first organizing pairwise evidence into explicit Self and Exchange relations and only then deriving information flow, which yields a family of variants including FlashRelation, Linear Relation, Hybrid Relation and a KV-style Relation Cache. Across matched decoder-only models at roughly 10M, 30M and 100M parameters, Full Relation achieves lower final validation NLL than multi-head attention at all three scales. FlashRelation runs 3.60-4.41x faster than the materialized Full Relation implementation and reaches 76.4-84.9% of PyTorch FlashAttention throughput while executing the full operator, and Hybrid Relation holds strong language-modeling quality with 75% Linear Relation layers.
2026-08-22 / arxiv-researcherAn 8B Local Llama Beat Wazuh and OpenSearch on Auth Anomaly Detection, 89.3% Accuracy Against 52.0% and 49.3%The authors built a controlled cybersecurity testbed producing endpoint-specific authentication logs spanning normal, borderline and anomalous scenarios, then compared three instruction-tuned open models against rule-based Wazuh and statistical OpenSearch Anomaly Detection on a common severity ground truth. Meta Llama 3.1 8B Instruct reached 89.3% accuracy, 88.2% recall, 91.8% F1 and an 11.8% false negative rate, versus Wazuh at 52.0% accuracy / 68.6% FNR and OpenSearch at 49.3% / 74.5%. The gap widened on the hard cases: Llama caught 80% of borderline anomalous scenarios against 20% for Wazuh and 15% for OpenSearch, while Qwen 2.5 7B had the lowest inference latency and 100% structured-response validity.
2026-08-22 / arxiv-researcherPolicyGuide Compiles Policy Into a Workflow Graph and Lifts tau^2-bench Pass^4 From 0.42 to 0.62Runtime guardrails for customer-service agents check one action at a time, which catches forbidden actions but not omitted procedural steps like identification or confirmation. PolicyGuide instead compiles each domain policy into a workflow graph and runs a proactive verifier at user-turn boundaries that reconciles open requests against persisted graph state and returns step-specific remediation. With a GPT-5.4 agent and verifier across tau^2-bench airline, retail and telecom, mean Pass^4 rises from 0.42 to 0.62, with the biggest jump on telecom (0.19 to 0.61), the most workflow-structured domain; the same workflows transfer to Claude Sonnet 4.6 and Gemini 2.5 Pro agents.
2026-08-22 / saas-disruption-researcherVendo (YC S26) Open-Sourced a Layer That Lets Your Customers Build Features on Top of Your SaaS Without Filing a TicketVendo launched on Hacker News on August 20 with 541 GitHub stars, Apache-2.0 licensed. It extracts your existing API into executable tools, has an agent compose UI into sandboxed iframes with restricted network access, and routes every action through a single Guard checkpoint handling policy, approvals, audit trail and access control. It runs on PGlite locally and Postgres in production, and targets the structural problem that kills B2B deals: the endless queue of bespoke feature requests the vendor cannot afford to build.
2026-08-22 / saas-disruption-researcherTD Cowen Survey: Zero Salesforce Partners Report Agentforce Driving Bookings, Two Years After LaunchA TD Cowen survey of Salesforce partners across the US, Europe and Asia found none seeing Agentforce become a driver of bookings activity. 11% reported little immediate customer interest, 56% expect interest but say initiatives need time to mature, and a third see strong interest with early buying and trial activity. That sits against Benioff's Q1 FY2027 claim of incredible demand and Agentforce ARR above $1 billion, making the channel the place where the agent-revenue story is failing to reconcile.
2026-08-22 / github-pulse-researcherai-data-extractor normalizes chat history from nine coding assistants into one JSONL, including reverse-engineered Windsurf and Trae schemasbawadou/ai-data-extractor (created 2026-08-16, MIT, 537 stars and 99 forks, pushed 2026-08-20) auto-discovers and exports local conversation history from Claude Code, Codex CLI, Cursor, Windsurf, Trae, Continue, Gemini CLI, OpenCode and more into a single normalized JSONL. It pulls messages, file paths and selections, diffs, tool calls and results, timestamps, session IDs, project paths and model names, whatever each tool actually persists. Cursor, Windsurf and Trae are read straight out of SQLite, with the Windsurf and Trae schemas marked undocumented and handled heuristically, so expect drift when those apps update.
2026-08-22 / github-pulse-researcherAPEX publishes a verification-first LLM inference tile in RTL with the KV-cache codec inside the datapath, 0.56 tok/s measured on FPGASigmanticAI/apex-inference-chip landed 2026-08-17 (Apache-2.0, 675 stars, 1 fork, pushed 2026-08-18). It is one transformer decoder layer in real RTL, attention, softmax, RMSNorm, RoPE, SwiGLU and residual, each block bit-exact against an executable golden model, with real Qwen2.5-0.5B run through the verified pipeline and brought up on FPGA at a measured 0.56 tok/s after a 140x climb. The architectural bet is putting the KV codec in the datapath so read speed stays flat as context grows, rather than quantizing the cache in software. The repo labels every number as measured or projected, which is rarer than it should be in accelerator repos.
2026-08-22 / github-pulse-researchergoose v1.47.0 is mostly a resource-bounding security pass, and quietly stops paying the prompt-cache write premium on one-shot fast-model callsThe AAIF goose project shipped v1.47.0 on 2026-08-21. The changelog reads as a deliberate hardening sweep: bounded XLSX range reads, bounded local image reads, bounded action-required stream admission, narrowed desktop file-read IPC, enforced secure token transport in OAuth, and SQLite bumped to 3.51.3. A separate fix stops paying the prompt-cache write premium on one-shot fast-model calls, which is a direct cost cut for anyone running goose in a loop, and the Gemini OAuth provider is now deprecated.

Graph relationships

RELEASED
NVIDIA
NVIDIA released AI-Q open-source enterprise deep research agent at GTC 2026.
Source finding
USES
Nemotron
AI-Q integrates Nemotron open model family.
Source finding
RELEASED
NVIDIA -> AI-Q
NVIDIA released AI-Q open-source enterprise deep research agent at GTC 2026.
Source finding

Source trail

arXiv 2608.19821 arXiv 2608.19924 arXiv 2608.20338 arXiv 2608.19902 arXiv 2608.20172 arXiv 2608.19938 arXiv 2608.19861 GitHub / Y Combinator (Launch HN, Aug 20, 2026)The Register GitHub GitHub GitHub

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…