Project Open to Alpha Team

External Truth Anchor — /feedback channel for Product.ai chat or Alloy that reaches engineering Slack within 60 seconds

Ship an in-product `/feedback` channel for one Product.ai surface (chat or Alloy). Users can flag a bad answer, ambiguous answer, or surprising answer with one click. The signal lands in an engineering Slack channel within 60 seconds, with the conversation context, the user's note, and a link to the trace. Production code, real route, real Slack channel, real users.

Project Overview

Discipline

AI Systems — AI Engineer · Engineering — Full-Stack Software Engineer

Duration

1 week

Compensation

Your stated freelance rate

Surface

Product.ai · Engineering · Consumer experience

Kernels

productai · engineering · consumer-experience

Outcomes

chat-expert · full-journey · dev-integrate

Tier

Foundational

Alpha Team

Open to alpha members who want to take this on

Tooling

Claude Code or Co-work

Why we want this done

The Anthropic dual postmortem is canonical: both incidents (August 2025 + April 23 2026) were detected by external community sources BEFORE internal eval apparatus. Stella Laurenzo's GitHub audit caught what Anthropic's internal eval missed. The /feedback command became the load-bearing community-detection channel for Claude Code. Closed-loop AI verification devolves to compound deception — this is the single most-cited Multi-Agent Physics finding (Meta-Axiom C). Product.ai today has no equivalent external truth anchor on its chat or Alloy surfaces. We are running closed-loop. The first surface to ship one breaks that pattern, immediately raises chat-expert and full-journey outcome traction by exposing real failure modes, and produces an instrumented data stream that every future eval harness depends on.

Scope

Pick the surface (chat or Alloy) and argue why
Design the in-product /feedback widget — where it lives, when it's discoverable, how a user invokes it
Build the route — capture conversation context, user note, anonymized trace link
Wire it to a real engineering Slack channel with a notification that includes context summary + click-through to full trace
Set up classification on inbound feedback (clearly bad / ambiguous / surprising / praise) — at minimum manual triage; ideally lightweight Haiku tagging
Document the response protocol — who sees what, what gets categorized, how it feeds into the eval substrate

What success looks like

The widget is live in production on the chosen surface within the trial window
At least one piece of real user feedback lands in Slack during the trial — and a team member responds to it within 24 hours
The notification surface is scannable in two seconds (context summary + link, not a wall of JSON)
The trace link works end-to-end (user click → full conversation visible to engineer)
Privacy handling is documented (PII scrubbing, retention, escalation rules)
The protocol doc is one page; an engineer who has never seen the system can adopt it on a second surface without re-asking

References

references.md

AI Engineering Phase 3 briefing axiom A5 (External Truth Anchors) and Multi-Agent Physics Meta-Axiom C
Anthropic Claude Code April 23, 2026 postmortem — /feedback as load-bearing community-detection
Stella Laurenzo's GitHub audit (6,852 sessions, 234,000 tool calls)
Product.ai chat surface code; Alloy code; Slack API; existing engineering Slack channel topology
axioms/AI & Agent Architecture/ for trace handling primitives

Constraints

Claude Code or Co-work as primary substrate
Real Slack channel, real route, real users — no mocks
Privacy-respecting: PII scrubbing on capture, retention rules documented
Slack notifications must be readable on mobile (Slack mobile app); avoid JSON-dumps
IP separation: chat and Alloy surfaces are application-layer; no aios-methods access required
The feedback widget must not degrade the surface UX — a second's hesitation on chat-load is a fail

Apply

Read the Codex (10 min)

The operating principles we work by. If they resonate, the rest of this will land. Open the Codex →

12-minute video screen

Hireflix, async. Questions are calibrated to this project specifically.

Chemistry call (30-60 min)

Direct call with the CEO. Strategic alignment and mutual fit. No problem-solving exercise.

Project begins within 2-3 weeks

1099 contractor agreement, NDA, paid at your stated rate. Day 1 in Santa Monica.

Alpha Team members can take this project without the screen-and-call sequence. Reach out via the Alpha Team channel.

Apply for this trial Browse other projects

Agent View · application/ld+json

{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@id": "https://product.ai/#organization"
    },
    {
      "@type": "SoftwareApplication",
      "@id": "https://product.ai/#application",
      "name": "Product.ai",
      "slogan": "Intelligent guidance at conversation speed",
      "description": "Product.ai is your AI shopping expert - verified product intelligence grounded in the Truth Graph, accessible to agents and humans alike.",
      "applicationCategory": "ShoppingApplication",
      "operatingSystem": "Web",
      "author": {
        "@id": "https://product.ai/#organization"
      },
      "offers": {
        "@type": "Offer",
        "price": "0",
        "priceCurrency": "USD"
      }
    },
    {
      "@type": "WebPage",
      "@id": "https://product.ai/join/projects/PRJ-11-external-truth-anchor-feedback/#webpage",
      "name": "External Truth Anchor — /feedback channel for Product.ai chat or Alloy that reaches engineering Slack within 60 seconds — Product.ai",
      "description": "Ship an in-product `/feedback` channel for one Product.ai surface (chat or Alloy). Users can flag a bad answer, ambiguous answer, or surprising answer with one click. The signal lands in an engineering Slack channel within 60 seconds, with the conversation context, the user's ...",
      "url": "https://product.ai/join/projects/PRJ-11-external-truth-anchor-feedback/",
      "isPartOf": {
        "@id": "https://product.ai/#website"
      },
      "about": {
        "@id": "https://product.ai/#organization"
      },
      "publisher": {
        "@id": "https://product.ai/#organization"
      }
    },
    {
      "@type": "JobPosting",
      "title": "External Truth Anchor — /feedback channel for Product.ai chat or Alloy that reaches engineering Slack within 60 seconds",
      "description": "Ship an in-product `/feedback` channel for one Product.ai surface (chat or Alloy). Users can flag a bad answer, ambiguous answer, or surprising answer with one click. The signal lands in an engineering Slack channel within 60 seconds, with the conversation context, the user's ...",
      "hiringOrganization": {
        "@id": "https://product.ai/#organization"
      },
      "employmentType": "CONTRACTOR",
      "datePosted": "2026-04-28",
      "jobLocation": {
        "@type": "Place",
        "address": {
          "@type": "PostalAddress",
          "addressLocality": "Los Angeles",
          "addressRegion": "CA",
          "addressCountry": "US"
        }
      },
      "baseSalary": {
        "@type": "MonetaryAmount",
        "currency": "USD",
        "value": {
          "@type": "QuantitativeValue",
          "unitText": "WEEK"
        }
      },
      "workHours": "1 week"
    },
    {
      "@type": "BreadcrumbList",
      "itemListElement": [
        {
          "@type": "ListItem",
          "position": 1,
          "name": "Product.ai",
          "item": "https://product.ai"
        },
        {
          "@type": "ListItem",
          "position": 2,
          "name": "Join",
          "item": "https://product.ai/join/"
        },
        {
          "@type": "ListItem",
          "position": 3,
          "name": "Projects",
          "item": "https://product.ai/join/projects/"
        },
        {
          "@type": "ListItem",
          "position": 4,
          "name": "External Truth Anchor — /feedback channel for Product.ai chat or Alloy that reaches engineering Slack within 60 seconds",
          "item": "https://product.ai/join/projects/PRJ-11-external-truth-anchor-feedback/"
        }
      ]
    }
  ]
}