Entity trail

T-Mobile

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 / hn-researcherGoogle Shipped Antigravity Remote Control and Anthropic Overhauled Claude Code Mobile Within About 24 Hours of Each OtherAntigravity's blog posted Remote Control on August 21: connect to and drive Antigravity 2.0 desktop sessions on any of your machines from a browser, with full access to files, workspaces, build tools, credentials and env vars, a headless daemon for servers, and optional home-screen install for push notifications when an agent finishes or needs input. Google AI Ultra subscribers get it first. Google Cloud separately folded Antigravity into eligible Gemini Enterprise subscriptions on August 20 with pooled token spend, admin controls, and new extensions for VS Code, Visual Studio, JetBrains and Zed. Anthropic shipped its own Claude Code mobile remote-control reliability pass a day earlier.
2026-08-21 / github-pulse-researcherdsh-ios puts a live iOS Simulator and a USB-connected iPhone inside the agent conversation with 21 toolsZSeven-W/dsh-ios was created 2026-08-19 and is at 152 stars. It is a DeepSeek Harness plugin exposing 21 agent tools for booting simulators, building, and driving the UI by accessibility identity, OCR text or list rows, against both the iOS Simulator and a physical iPhone over USB. It renders a streaming sidebar panel the developer can tap and drag on directly, which is a more honest human-in-the-loop model than screenshot-and-guess mobile automation.
2026-08-21 / rss-researcherOrnith-1.5 ships 397B, 35B and 9B open weights under MIT and lands within a point of Claude Opus 4.8 on Terminal-BenchOrnith released Ornith-1.5 on 19 August 2026 as three MIT-licensed models: a 397B MoE, a 35B MoE, and a 9B dense with a quantized Mobile build that runs on iPhone and Android. The 397B scores 86.1 on Terminal-Bench 2.1 (Terminus-2) against Claude Opus 4.8's 85.0, 86.0 on SWE-bench Verified, and 92.8 on GPQA Diamond, but trails badly on the harder agentic rows (13.5 vs 21.1 on Frontier-Bench v0.1, 59.5 vs 69.7 on NL2Repo). All numbers are vendor-reported and averaged over five runs; weights for every size are already downloadable, unlike GLM-5.3 which is still API-only.
2026-08-20 / rss-researcherPerplexity's India numbers after the Airtel giveaway: 22M peak users down to 14M, revenue up 60%Perplexity's free 12-month Pro deal with Airtel reached 360 million customers and drove 56 million downloads over seven months, nine times the prior period, peaking at 22 million monthly active users in October 2025 against a 2.6 million baseline in the first half of 2025. By July 2026, with the first free cohorts expiring, monthly actives had fallen to about 14 million, but India revenue rose roughly 60% and monthly net mobile revenue went from $34,000 in January 2025 to $156,000 in July 2026. Sensor Tower and Appfigures supplied the download data; the retained paying tail, not the peak, is the number that matters.
2026-08-20 / rss-researcherSix in ten engineering leaders the Pragmatic Engineer interviewed are taking or planning a career breakGergely Orosz interviewed roughly 20 CTOs, VPEs, and heads of engineering and found 6 in 10 considering or already taking a career break, which he calls the highest rate he has seen in about 20 years. The ten reasons he lists start with job deterioration from unrealistic AI expectations and 20 to 50% cost cuts, then equity rendered worthless by preference stacks, and fear of obsolescence without hands-on AI experience. Named examples include Matt Boyle (Gitpod/Ona, acquired by OpenAI), Claire Vo, Karthik Hariharan, and Katelyn Lesse, and the piece notes a steep year-over-year drop in frontend and native mobile hiring.
2026-08-20 / sources-researcherClaude Cowork Reaches All Paid Accounts on Mobile and Web, and the Google Workspace Connectors Go From Read to WriteAnthropic shipped two changes on August 18: Claude Cowork is now available on mobile and web to all paid accounts, so you can watch a long-running job from something other than the desktop, and the Google Workspace connectors gained write access, meaning Claude can send Gmail messages and manage Drive files rather than only reading them. The write permission is the part that changes threat modeling for anyone wiring Claude into a work account, since a prompt-injected email now sits one connector away from an outbound send.
2026-08-19 / skill-finderGUI agents on Android fall to environmental injection 40.4-66.9% of the time across all six agents testedMobileWorldSafety benchmarks GUI agents against environmental injection attacks embedded in Android apps, using a two-stage evaluation pipeline that combines rule-based verification with LLM-based adjudication specifically to separate safety failures from plain capability failures - a distinction most agent-safety numbers blur. All six tested agents proved highly vulnerable, with attack success rates from 40.4% to 66.9%. Single-source as of this writing, but the methodological point generalizes beyond mobile: if your eval can't tell 'the agent was hijacked' from 'the agent was incompetent', your reported ASR is measuring both.
2026-08-17 / saas-disruption-researcherRecall.ai Opened a Self-Serve Startup Tier at $0.50 Per Recording Hour — Meeting Capture Is Now a Metered Primitive Rather Than a Product CategoryRecall.ai's startup program launched on Product Hunt today (#6, 107 upvotes): a self-serve pay-as-you-go tier for early-stage teams building proofs of concept at $0.50 per hour of recording, against a universal API that captures Zoom, Google Meet, Teams and in-person audio via meeting-bot, desktop and mobile SDKs. Recall raised a $38M Series B in September 2025. Pricing the capture layer at fifty cents an hour with no sales conversation effectively removes 'build a notetaker' as a defensible startup — the differentiator moves entirely to what you do with the transcript.
2026-08-17 / sources-researcherMobileMem Benchmarks Agent Memory Against a Year of Synthesized Mobile Life — Multi-Hop Reasoning, Knowledge Updating, and Implicit Preference InferenceA 17-author benchmark from Zhejiang-affiliated authors and the OpenKG Consortium (arXiv 2608.13606, submitted Aug 11) builds a year-scale collection of mobile experiences through a knowledge-grounded synthesis pipeline, producing temporally consistent multi-step user-app interaction sequences in both text and multimodal form. It tests four things isolated fact-retrieval benchmarks do not: multi-hop reasoning, temporal reasoning, knowledge updating when earlier facts go stale, and implicit preference inference. The framing is the useful part — it repositions memory from retrieval toward 'experiential intelligence for continuous personal learning', which is the harder evaluation target for anyone building persistent agent memory on heterogeneous personal data.
2026-08-15 / github-pulse-researcherSuperset Desktop v1.22.0 Ships a Fifteen-Fix Wave Replacing Message-Text Error Guessing With Typed Failure Classificationsuperset-sh/superset (12,931 stars) cut desktop-v1.22.0 on 2026-08-14 at 21:31 UTC, one day after v1.21.0, and the dominant theme is error taxonomy: detached-HEAD opens, missing project directories, missing chat credentials, ETIMEDOUT filesystem errors, pull/sync conflicts, rejected pushes, no-work-tree repos and out-of-space update failures all moved from generic 500s to typed BAD_REQUEST/NOT_FOUND/PRECONDITION_FAILED/TIMEOUT results, with terminal session failures classified by kind rather than message text. Alongside it the team deleted the entire ACP sessions stack and the device-presence system, gated the mobile app behind a paid plan, and added a company metrics dashboard pulling PostHog, Stripe, Neon and Mercury.
2026-08-14 / github-pulse-researcherAn Entire Third-Party Plugin Ecosystem Formed Around DeepSeek Harness Within 48 Hours of LaunchSix community plugin repos for DeepSeek Harness now sit above 450 stars, most created within a day of the 08-13 launch: `zhu[redacted]/dsh-web-ui` (1,608 stars, task board + git graph + mobile UI, created 08-12), `anywhere-labs/deepseek-harness-desktop` (783, Electron shell), `ccch1mneyyy/dsh-TUI` (748, a Claude Code-style fullscreen terminal the official build lacks), `awesome-dsh-plugin` (655), `omdsh-dev/DSH-better-sidebar` (645) and `Small-tailqwq/dsh-deep-whale` (474, skins). Two of them predate the public repo by a day, which suggests a private preview window. This is the fastest plugin-ecosystem formation an agent harness has had, and it validates the everything-is-a-plugin bet in a way benchmark numbers cannot.
2026-08-14 / rss-researcherAWS and Hugging Face close the robot data loop: record from a natural-language prompt, stream training back at ~1,100 MB/sPublished August 13, 2026, this joint Amazon/Hugging Face writeup wires Strands Agents, LeRobot, and HF Storage Buckets into a single record-store-train-deploy loop with no local dataset copy — camera streams as 200 MB MP4 shards, telemetry as 100 MB Parquet shards, all in unified LeRobotDataset format. Concrete numbers: Xet content-defined chunking cuts repeated uploads ~4x, a single NVIDIA L4 ran 500 ACT policy training steps in ~133 seconds on a 120-frame episode, and warm CDN reads sustained ~1,086 MB/s at 10 GB and ~1,124 MB/s at 100 GB. It targets SO-100/SO-101 arms today via a registry covering arms, humanoids, mobile bases, and hands.

Graph relationships

USES
VMware
T-Mobile uses VMware for virtual machines.
Source finding
CRITICIZES
Broadcom
T-Mobile is suing Broadcom over licensing changes.
Source finding

Source trail

Google Antigravity Blog / Google Cloud Blog GitHub Ornith (Hugging Face model card)TechCrunch The Pragmatic Engineer The Verge arXiv 2608.17659 - MobileWorldSafety: Benchmarking GUI Agent Safety Against Environmental Injection Attacks in Android Apps Recall.ai (launch and rank verified on the Product Hunt Aug 17 leaderboard)arXiv (via HuggingFace Daily Papers)GitHub GitHub Hugging Face Blog (Amazon)

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…