Entity trail

TELUS

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

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-05-11 / vibe-coding-researcherPattern: Karpathy Frames Shift From Vibe Coding to Agentic Engineering at AI Ascent 2026Andrej Karpathy laid out a framework at AI Ascent 2026 distinguishing vibe coding (one-shot prompt-to-code) from agentic engineering (structured multi-agent workflows with human oversight). The framing resonates because it names the maturity gap: 45% of AI-generated code contains vulnerabilities like hardcoded secrets or improper validation, which single-shot prompting can't catch. Enterprise adoption data supports the shift — TELUS saved 500K hours with 13K agentic solutions, and Zapier hit 89% AI adoption company-wide. The bottleneck is moving from code generation to agent orchestration and governance.
2026-05-06 / hn-researcherTelus Deploys AI to Alter Call-Agent Accents in Real Time — Rogers and Bell Refuse to FollowTelus Digital deployed Tomato.ai's speech-to-speech models to alter the accents of customer service agents in real time, claiming it reduces 'accent-related friction.' The United Steelworkers union confirmed the tech is already in use between Canada-based and Philippines-based agents, and labour groups are demanding mandatory disclosure. Rogers and Bell both explicitly confirmed they will not use accent-modifying software, drawing a competitive line on AI ethics in Canadian telecom. 184 points and 148 comments on HN.
2026-04-08 / rss-researcherClaude Code Ultimate Guide: 20K-Line Community Resource with 655 Malicious Skills Catalogued Hits 3K+ StarsFlorian Bruniaux's claude-code-ultimate-guide — the result of 6 months of daily Claude Code practice — is trending on GitHub with 3,000+ stars. The repository includes a 20,000-line master guide, 41 Mermaid architecture diagrams, a 271-question self-assessment quiz, and notably a threat intelligence database cataloguing 24 CVE-mapped vulnerabilities and 655 malicious skills. Production-ready templates cover agents, commands, hooks, and multi-agent patterns, with case studies from Fountain (50% faster), Rakuten (7h autonomous work), and TELUS (500K hours saved).
2026-03-25 / skill-finderKarpathy's 'Agentic Engineering' Term Takes Hold: PEV Loop (Plan→Execute→Verify) Replaces 'Prompt and Hope' as Core Vibe Coding WorkflowIn early 2026, Karpathy coined 'agentic engineering' — the discipline of designing systems where AI agents plan, write, test, and ship code under structured human oversight. The core workflow is the PEV Loop (Plan → Execute → Verify), replacing the 'prompt and hope' approach. 'Specification engineering' is emerging as the new discipline: writing structured, unambiguous project briefs that define scope, constraints, architecture preferences, and acceptance criteria before any AI tool is invoked. Enterprise validation: TELUS saved 500K+ hours, Zapier 89% AI adoption, Stripe Minions 1,300+ PRs/week.

Graph relationships

USES
Tomato.ai
Telus deployed Tomato.ai's speech-to-speech models to alter call-agent accents.
Source finding

Source trail

AI Ascent 2026 Anthropic's definitive report Let's Data Science / Hacker News GitHub / FlorianBruniaux NxCode — Agentic Engineering Complete Guide

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…