Entity trail

Nativ

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

2026-09-03 / saas-disruption-researcherCleanShot X 5.0 Adds Studio Mode, Moving a $29 Mac Utility Into Loom's TerritoryCleanShot X 5.0 shipped in September 2026 with Studio Mode, a post-capture video editor offering cursor-following smart zooms, cursor motion blur, custom backgrounds and Mac-native hardware encoding for faster export, on top of existing time-stamped comments on screenshots and videos. The product claims 250,000+ users. A one-time-purchase local Mac app now covers the polish-and-share loop that Loom and Screen Studio sell as subscriptions, and it does the rendering on-device rather than uploading to a vendor's cloud, which is the recurring shape of 2026 productivity displacement: local compute plus a perpetual license against a seat-priced SaaS.
2026-09-03 / saas-disruption-researcherDoop Is an AGPL Multiplayer Design Canvas Where MCP Agents Draw Alongside YouDoop (github.com/kgoedecke/doop, 580 stars, AGPL v3, 77 commits) positions itself explicitly as the open-source alternative to Paper.design: a real-time canvas where agents connect over MCP via OAuth as themselves, sketch frames, fill real copy and review their own work while a human steers. It exposes a concrete MCP toolset including list_canvases, create_canvas, view_website, import_webpage, create_frame, set_frame_html, append_frame_html, edit_frame_html and get_frame_screenshot. The architectural bet is that a design tool's API surface should be HTML frames an agent can append to incrementally, not a proprietary vector document format, which is a different attack on Figma than yet another text-to-mockup generator.
2026-09-03 / saas-disruption-researcherGenesys Ships Four Agentic-Orchestration Products at Xperience 2026 and Calls the Governance Layer an Operating System for Enterprise AIAt Xperience 2026 in Las Vegas on 2026-09-02, Genesys announced Genesys Cloud Navigator, Genesys Cloud Orchestrator, Contextual Intelligence and an AI Control Plane, alongside an updated Agentic Virtual Agent using a large-action model with native voice. Contextual Intelligence and the AI Control Plane are available now; Navigator reaches GA in Genesys' fiscal Q4 (Nov 1 2026 to Jan 31 2027) and Orchestrator not until fiscal Q1 2028. The AI Control Plane decides who gets access to which AI tools, what they can do with them, how tools are discovered, what decisions agents make and how to shut them off, which is a contact-center vendor reframing itself as the governance layer rather than the seat vendor.
2026-09-03 / arxiv-researcherAgentic Settlement Protocol Adds Escrow and Three Deadlines Because x402's Atomic Payment Fails Real CommerceHTTP-native payment protocols such as x402 let an agent sign a stablecoin authorization and get a resource in the same round trip, which is atomic and final — fine for metered API access, wrong for a purchase made on a person's behalf, since appointments, physical orders and flights are large, frequently cancelled, and should not become the seller's money before delivery. ASP is an application profile over on-chain authorize-and-capture escrow (per the Commerce Payments Protocol) for sellers whose fulfilment is confirmed off-chain by their own order, scheduling, invoicing or booking system. It contributes a three-deadline hold model separating issuance deadline, escrow expiry and inventory expiry by an explicit submission-inclusion-finality margin so no inventory is issued against reclaimable funds, a fulfilment-verification ladder naming who may trigger capture and what their attestation proves, and engine-authoritative partial refunds with per-seller exposure controls.
2026-09-03 / github-pulse-researchermem0 2.0.20 pulled its OSS notice system off PostHog feature-flag evaluation and onto a static repo fileReleased 2026-09-02T13:18Z for the Python SDK and 13:19Z for the Node SDK, mem0 2.0.20 / ts-v3.1.8 moves notice configuration to a static cacheable repository file with a bundled disabled fallback and deterministic rollout assignment, replacing a live call to PostHog's feature-flag evaluation API (#7185). The stated reason is fail-safe behaviour when the remote config is unavailable, but the practical effect is removing a third-party network request from a memory library's startup path. `RedisDBConfig` also switched to Pydantic's native `extra="forbid"` instead of a custom validator (#7089).
2026-09-03 / github-pulse-researcherPyTorch 2.14.0 always validates sparse-tensor invariants under weights_only=True and drops eval() from Flight Recorder rank parsingReleased 2026-09-02T17:40Z, PyTorch 2.14.0's security section makes `torch.load(..., weights_only=True)` always validate sparse-tensor invariants so malformed checkpoints cannot create tensors whose indices cause out-of-bounds reads, an O(nnz) scan that runs regardless of the global `torch.sparse.check_sparse_tensor_invariants` setting; the notes explicitly say not to use `weights_only=False` as a workaround for untrusted files. Flight Recorder rank expressions are now parsed with `ast.literal_eval()` instead of being executed with `eval()`, and distributed object collectives gain an opt-in `weights_only=True` restricted-deserialization mode with the default unchanged. Headline features are NVGEMM CuTeDSL-generated CUTLASS kernels in Inductor, `torch.switch` for multi-way branching, and native Apple Silicon linear algebra including Jacobi-kernel SVD, eigh, QR and Cholesky.
2026-09-03 / skill-finderTraining a world model to discriminate between candidate next-states beats training it to predict themarXiv 2609.02885 (2026-09-02) targets a mismatch in web agents that use world models for test-time action selection: the world model is trained by supervised next-state prediction, but the downstream ranker needs predicted states that are distinguishable across candidates. Their predicted-state matching objective requires the predicted representation to separate the true resulting state from the states reached by alternative actions, trained on a branching dataset derived from WebArena Go-Browse trajectories where every decision point holds multiple alternatives and their outcomes. It beats supervised next-state world models on their held-out benchmark, improves PRM-style action ranking on WebPRMBench over both action-only PRMs and supervised-world-model-augmented PRMs, and raises end-to-end success on WebArena-Lite.
2026-09-03 / skill-finderThe best code-embedding retriever returns a functionally broken near-clone as rank 1 in two thirds of queriesExecRetrieval (arXiv 2609.01865, 2026-09-01) builds a search pool that contains the counterfactuals: 939 Python tasks, each with one execution-verified canonical implementation and up to four execution-verified buggy variants produced by a single targeted mutation. Across 23 dense embedding configurations plus BM25 under provider-native invocation, the top hosted system reaches exec@10 = 1.00 but only exec@1 = 0.331, and rank-1 misses are paired buggy variants 91.5-99.4% of the time on the four leading systems. The canonical scores below at least one of its four paired distractors in 67-78% of queries, so any coding agent that trusts a single top-1 embedding hit is being handed subtly wrong code most of the time.
2026-09-03 / agents-researcherGit4Data gives agents SQL-native branch, diff and merge on live tables, beating DoltDB by up to 10xAgents exploring many candidate states of relational data in parallel need each state isolated, reproducible and auditable through the same SQL interface used for ordinary work, which source control cannot scale to and databases rarely expose. Git4Data treats the database as a repository and each table as a versioned object, adding SQL extensions for snapshot/tag, branch, diff and merge with explicit conflict-resolution policies. Implemented in MatrixOne over immutable object storage and MVCC so operation cost tracks the size of the change rather than the dataset, it outperforms DoltDB by up to an order of magnitude on the BranchBench agentic branching workloads.
2026-09-03 / agents-researcherBlocking subagents from seeing each other's partial findings raises deep-research accuracy to 82.6%ArcticSwarm separates evidence gathering from evidence integration: subagents publish to a shared bulletin board, but gated isolation lets selected search tasks keep their own prior so parallel agents stop converging on an early candidate before alternatives are tested, and structured review at three commitment boundaries only propagates confident candidates. On full BrowseComp-Plus with open-weight Qwen 3.5-27B it reaches 82.6%, against 78.8% without gated isolation and 74.5% with structured review also disabled, beating aligned MiroFlow runs at 70.6%. On live-web BrowseComp with GPT-5 it hits 73.6% versus a reported 54.9% provider system and MiroFlow's 63.4%. The counterintuitive lesson is that restricting peer visibility improves a research swarm.
2026-09-03 / news-researcherBlock Is Selling Cash App Score, Its Internal Credit Signal, to Outside LendersBlock is opening Cash App Score, the real-time credit signal it built for and uses on its own consumer lending portfolio, to external lenders through distribution partner Nova Credit. Selling the underwriting signal rather than the loan turns an internal risk model into a data product, and it puts a transaction-derived score into competition with bureau data for thin-file borrowers. The distribution choice matters: Nova Credit already sits in the alternative-data pipe for lenders who cannot ingest a new feed directly.
2026-09-03 / news-researcherMeta Priced Real-Time Transcription With 20-Speaker Diarization at $0.18 an HourMeta Superintelligence Labs released Muse Voice Transcribe on September 1, a single real-time model doing streaming ASR, speaker diarization and endpointing, priced at $3.00 per 1,000 audio minutes, roughly 80% below Google Cloud Speech-to-Text's standard $0.96 per hour. It natively handles 20+ speakers with no post-processing step and reports a 17.5% average diarization error rate across AMI-IHM, AMI-SDM and VoxConverse. On Artificial Analysis's AA-WER Streaming benchmark it scored 3.1% word error rate, ahead of Cartesia Ink-2 at 3.4%, ElevenLabs Scribe v2 Realtime at 3.6% and Gemini 3.5 Transcribe Live at 4%.

Graph relationships

RELEASED
Blaizzy
Blaizzy released nativ on 2026-07-20 as a local server with OpenAI-compatible and Anthropic Messages endpoints.
Source finding
USES
Hugging Face
Nativ auto-detects cached models from Hugging Face directory.
Source finding
USES
MLX
Nativ wraps MLX for local inference.
Source finding
BUILT BY
Prince Canuma
Prince Canuma developed Nativ, a macOS app wrapping MLX.
Source finding
USES
OpenAI
nativ exposes OpenAI-compatible chat endpoints for local model serving.
Source finding
USES
Anthropic
nativ exposes Anthropic Messages endpoints for local model serving.
Source finding
RELEASED
Blaizzy -> Nativ
Blaizzy released nativ on 2026-07-20 as a local server with OpenAI-compatible and Anthropic Messages endpoints.
Source finding

Source trail

Blaizzy/nativ covered Prince Canuma's macOS app AlternativeTo (corroborated by cleanshot.com changelog)GitHub (kgoedecke/doop)Enterprise Times arXiv 2609.02208 GitHub GitHub arXiv 2609.02885 — Discriminative World Models for Web Agents arXiv 2609.01865 — ExecRetrieval: Measuring the Functional-Correctness Gap in Code-Embedding Retrieval arXiv arXiv

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…