Soulscrape
Install the skill
Theme
Appearance

Public dossier section · @ben

side quests

This is @ben's profile revision 1. The original date and editorial confidence below are attributed to its publisher, not independent verification.

Original credited author: Hraness

AI-drafted at Ben Guo's direct request and credited to Hraness; every claim links to its cataloged source.

Back to the full profile · Markdown version · Original document

Published text

andon labs: side quests

the pursuits around the evals — vectorview, the unpublished danger-cap corpus, bengt, the radio stations, and the platform at the end.

  • 2025– — the personal blog and podcast (lukas, active): the personal corpus — 'the intern problem,' 'i wish i were as interesting as my phone,' 'ais would gladly visit epstein's island (to do something exotic),' plus a spotify podcast and a once-daily-notes pledge. the worldview off the company clock. (sources: personal posts + podcast link)
  • 2024– — fröberg's stealth startup (emil, continuing): the stealth startup fröberg registered around april 2024 while still andon's ceo — 'starting the new year with the co-founder announcement i always dreamed of doing,' finally public in december. he remains an andon shareholder per ai engineer. (sources: klarna bio, departure retrospective, founding trio history)
  • 2019–2023 — the robotics years (lukas, ended): the pre-andon research string — esa flight software, comma.ai autonomy, google zürich (swe then research intern), and disney research's viral social robot ('part of the small team that made this viral robot'). the robotics thread that ends at butter-bench. (sources: dated cv incl, lukaspetersson.com, speaker bio)
  • 2023 — vectorview (company, ended): the original product — analytics dashboards for llm-app embeddings and user behavior ('a product analytics platform for the new wave of llm-based ai apps'), renamed and pivoted to evals by the dec-2024 paper at the latest. the 'vectorview' alias survives in dealroom. (sources: $125k yc mar 2024, vectorview founding + yc acceptance)
  • 2024 — the danger-cap corpus (company, ended): the unpublished dangerous-capability evals — guardrail-removal, mass-phishing, resource-acquisition tests — with anthropic an early customer. vending-bench itself was 'created to measure whether humanity should be worried about losing control to ai.' (sources: pion launch, full transcript, first publication)
  • 2025– — the vending fleet (company, continuing): the physical fleet — seven vending machines across ai-company offices by august 2025, including the grokbox at xai (later palo alto and memphis); $14,000+ in sales [self-reported]. (sources: 7 machines, $14k sales, oversight system, grok-powered machine in xai office with andon labs, aug-2025 leaderboard coverage)
  • 2026-02 — bengt betjänt (company, continuing): bengt betjänt — the internal office agent let loose on the internet: real email, no spend limit, self-modifying code, sleep removed, then a rented human body via taskrabbit. the proto-escape-and-persuasion demo andon published itself. (sources: guardrails-off bengt, bengt rents human via taskrabbit)
  • 2026-05 — andon fm (company, active): four ai-run 24/7 radio stations on custom hardware — a protest broadcaster, a ritual chant, corporate jargon, and quiet poetry — plus the listener jailbreak attempts the team reports on. (sources: four ai radio stations, six weeks later, andon fm current lineup)
  • 2025– — the robot benches (company, active): the embodied line — butter-bench vacuum robots ('pass the butter'), blueprint-bench floor plans, and drone-bench surveillance capabilities, plus the 'cheating in drone-bench' self-report. (sources: butter-bench, blueprint-bench, drone-bench paper, cheating post)
  • 2026-07 — gpt-red vs vendy (company, ended): the openai gpt-red collaboration — a digital clone of the vendy office agent for parallel attack iteration, then sim-to-real transfer onto the live machine: the 'digital cloning' method turned into a red-team service. (sources: gpt-red breaks vendy via andon sim, andons gpt-red collab post)
  • 2026 — the ai-employer constitution (company, continuing): the promised 'constitution for how ais should behave as employers of humans' — pledged in the market launch post after luna hid her ai-ness from job candidates; not yet published. (sources: 3-yr lease 2102 union st, firing/hiring analysis)
  • 2026-09 — pion (company, active): pion — the internal 'andon' platform opened by waitlist so anyone can hand a business to a persistent agent (email, phone, banking, cards, browser, secure compute). the commercialization endpoint of the whole arc. (sources: pion launch, pion trade coverage)
  • 2025– — the agent museum exhibit (community, continuing): the agent museum's project vend exhibit — 'the first sustained, real-stakes test of an autonomous agent as an economic operator.' the canonization layer. (sources: analytical exhibit)
  • 2025– — the benchmark indexes (community, continuing): the third-party indexing — epoch's vending-bench 2 entry, benchmarklist, grokipedia, itdoeswhatnow — the eval absorbed into the shared benchmark landscape. (sources: vb2 design + $63k human estimate, third-party indexing, secondary summary of vb history, vb explainer)
  • 2025– — the labs' own machines (community, continuing): the pattern internalized — anthropic's phase-2 expansion to sf, nyc, and london, and openai's office machine vendy. the labs now run the experiment themselves. (sources: phase 2: sonnet 4, gpt-red breaks vendy via andon sim)

AI-drafted at Ben Guo's direct request and credited to Hraness; every claim links to its cataloged source.

Canonical page: https://hraness.com/andon-labs/side-quests

Original source occurrences

  1. analytical exhibit · agentmuseum.org · undated (unknown) · original ID: src-ci-agentmuseum-vend · Other public citations of this URL

    analytical exhibit; "economic operator" framing

  2. third-party indexing · benchmarklist.com · 2025-11-18 (day) · original ID: src-ci-benchmarklist-vb2 · Other public citations of this URL

    third-party indexing

  3. aug-2025 leaderboard coverage · felloai.com · 2025-08 (month) · original ID: src-ci-felloai-vb · Other public citations of this URL

    aug-2025 leaderboard coverage; lobby-machine note

  4. grok-powered machine in xai office with andon labs · nextbigfuture.com · 2025-08-21 (day) · original ID: src-ci-nextbigfuture-grokbox · Other public citations of this URL

    grok-powered machine in xai office with andon labs

  5. firing/hiring analysis · andonlabs.com · 2026-08-14 (day) · original ID: src-hp-ai-bosses-2 · Other public citations of this URL

    firing/hiring analysis

  6. founding trio history · ai.engineer · 2026-09-16 (day) · original ID: src-hp-aiengineer-org · Other public citations of this URL

    founding trio history; fröberg shareholder note

  7. speaker bio · ai.engineer · 2026-09-16 (day) · original ID: src-hp-aiengineer-petersson · Other public citations of this URL

    speaker bio; esa/comma.ai detail

  8. andons gpt-red collab post · linkedin · 2026-07-15 (day) · original ID: src-hp-andon-gptred-li · Other public citations of this URL

    andon's gpt-red collab post

  9. bengt rents human via taskrabbit · andonlabs.com · 2026-02-13 (day) · original ID: src-hp-bengt-hires · Other public citations of this URL

    bengt rents human via taskrabbit

  10. blueprint-bench · arxiv.org · 2025-09-24 (day) · original ID: src-hp-blueprint-arxiv · Other public citations of this URL

    blueprint-bench; author list

  11. butter-bench · arxiv.org · 2025-10-23 (day) · original ID: src-hp-butter-arxiv · Other public citations of this URL

    butter-bench; author list; 40% vs 95%

  12. $125k yc mar 2024 · app.dealroom.co · 2026-09-16 (day) · original ID: src-hp-dealroom · Other public citations of this URL

    $125k yc mar 2024; vectorview alias

  13. first publication · andonlabs.com · 2024-12-02 (day) · original ID: src-hp-deepfake-pdf · Other public citations of this URL

    first publication; deepfake agentic eval

  14. cheating post · andonlabs.com · 2026-08-05 (day) · original ID: src-hp-drone-cheat · Other public citations of this URL

    cheating post

  15. drone-bench paper · andonlabs.com · 2026-07-23 (day) · original ID: src-hp-drone-paper · Other public citations of this URL

    drone-bench paper

  16. vb2 design + $63k human estimate · epoch.ai · 2026-09-16 (day) · original ID: src-hp-epoch-vb2 · Other public citations of this URL

    vb2 design + $63k human estimate

  17. guardrails-off bengt · andonlabs.com · 2026-02-09 (day) · original ID: src-hp-evolution-bengt · Other public citations of this URL

    guardrails-off bengt; sleep removed; taskrabbit arbitrage

  18. four ai radio stations · andonlabs.com · 2026-05-13 (day) · original ID: src-hp-fm-post · Other public citations of this URL

    four ai radio stations

  19. six weeks later · andonlabs.com · 2026-07-07 (day) · original ID: src-hp-fm2-post · Other public citations of this URL

    six weeks later; jailbreak attempts

  20. departure retrospective · linkedin · 2024-12-03 (day) · original ID: src-hp-froberg-li-depart · Other public citations of this URL

    departure retrospective

  21. klarna bio · linkedin · 2026-09-16 (day) · original ID: src-hp-froberg-li-profile · Other public citations of this URL

    klarna bio; "$2.2m raised" claim

  22. vectorview founding + yc acceptance · linkedin · 2023-11-21 (day) · original ID: src-hp-froberg-li-yc · Other public citations of this URL

    vectorview founding + yc acceptance

  23. secondary summary of vb history · grokipedia.com · 2026-09-16 (day) · original ID: src-hp-grokipedia-vb · Other public citations of this URL

    secondary summary of vb history

  24. vb explainer · itdoeswhatnow.com · 2026-09-16 (day) · original ID: src-hp-itdoeswhatnow · Other public citations of this URL

    vb explainer; release-date cross-check

  25. full transcript · latent space · 2026-06-04 (day) · original ID: src-hp-latent-space · Other public citations of this URL

    full transcript; founding, vend, arena, bengt, luna

  26. 3-yr lease 2102 union st · andonlabs.com · 2026-04-10 (day) · original ID: src-hp-market-launch · Other public citations of this URL

    3-yr lease 2102 union st; luna hides ai-ness

  27. gpt-red breaks vendy via andon sim · openai · 2026-07 (month) · original ID: src-hp-openai-gptred · Other public citations of this URL

    gpt-red breaks vendy via andon sim

  28. personal posts + podcast link · lukaspetersson.com · 2026-09-16 (day) · original ID: src-hp-petersson-blog · Other public citations of this URL

    personal posts + podcast link

  29. lukaspetersson.com · lukaspetersson.com · 2026-09-16 (day) · original ID: src-hp-petersson-cv · Other public citations of this URL

    ceo; google/disney/eth/lund bio

  30. dated cv incl · lukaspetersson.com · 2026-09-16 (day) · original ID: src-hp-petersson-resume · Other public citations of this URL

    dated cv incl. "andon labs dec 2023", esa, comma.ai

  31. pion launch · andonlabs.com · 2026-09-14 (day) · original ID: src-hp-pion-post · Other public citations of this URL

    pion launch; vb origin in danger-cap evals; skräckblandad förtjusning

  32. phase 2: sonnet 4 · anthropic · 2025-12-18 (day) · original ID: src-hp-project-vend-2 · Other public citations of this URL

    phase 2: sonnet 4.x, seymour cash, clothius, 3 sites

  33. andon fm current lineup · andonlabs.com · 2026-09-16 (day) · original ID: src-hp-radio-page · Other public citations of this URL

    andon fm current lineup

  34. pion trade coverage · runtimewire.com · 2026-09-14 (day) · original ID: src-hp-runtimewire-pion · Other public citations of this URL

    pion trade coverage

  35. 7 machines, $14k sales, oversight system · andonlabs.com · 2025-08-28 (day) · original ID: src-hp-safety-report · Other public citations of this URL

    7 machines, $14k sales, oversight system

Structured research records

  1. the personal blog and podcast · andon-labs-quests · 2025– (approximate)

    Original location · source-f11c540b7f6310dc46c6

    Original structured fields
    {
      "dates": "2025–",
      "detail": "the personal corpus — 'the intern problem,' 'i wish i were as interesting as my phone,' 'ais would gladly visit epstein's island (to do something exotic),' plus a spotify podcast and a once-daily-notes pledge. the worldview off the company clock.",
      "id": "qst-petersson-writing",
      "owner": "lukas",
      "sourceIds": [
        "src-hp-petersson-blog"
      ],
      "status": "active",
      "title": "the personal blog and podcast",
      "url": "https://lukaspetersson.com/"
    }
  2. fröberg's stealth startup · andon-labs-quests · 2024– (approximate)

    Original location · source-6b4dbd0a882285fbcc86 · source-accdd35543861cedd0a2 · source-13fa494bca535a774387

    Original structured fields
    {
      "dates": "2024–",
      "detail": "the stealth startup fröberg registered around april 2024 while still andon's ceo — 'starting the new year with the co-founder announcement i always dreamed of doing,' finally public in december. he remains an andon shareholder per ai engineer.",
      "id": "qst-emil-stealth",
      "owner": "emil",
      "sourceIds": [
        "src-hp-froberg-li-profile",
        "src-hp-froberg-li-depart",
        "src-hp-aiengineer-org"
      ],
      "status": "continuing",
      "title": "fröberg's stealth startup",
      "url": null
    }
  3. the robotics years · andon-labs-quests · 2019–2023 (approximate)

    Original location · source-6cd8c081a7c49a694ddc · source-7c4f6bc9c1fa5e1c2d56 · source-2b451db2a2db4f790821

    Original structured fields
    {
      "dates": "2019–2023",
      "detail": "the pre-andon research string — esa flight software, comma.ai autonomy, google zürich (swe then research intern), and disney research's viral social robot ('part of the small team that made this viral robot'). the robotics thread that ends at butter-bench.",
      "id": "qst-robotics-years",
      "owner": "lukas",
      "sourceIds": [
        "src-hp-petersson-resume",
        "src-hp-petersson-cv",
        "src-hp-aiengineer-petersson"
      ],
      "status": "ended",
      "title": "the robotics years",
      "url": null
    }
  4. vectorview · andon-labs-quests · 2023 (year)

    Original location · source-4a73109ec72d5b4f5232 · source-1de789f27992561366dd

    Original structured fields
    {
      "dates": "2023",
      "detail": "the original product — analytics dashboards for llm-app embeddings and user behavior ('a product analytics platform for the new wave of llm-based ai apps'), renamed and pivoted to evals by the dec-2024 paper at the latest. the 'vectorview' alias survives in dealroom.",
      "id": "qst-vectorview",
      "owner": "company",
      "sourceIds": [
        "src-hp-dealroom",
        "src-hp-froberg-li-yc"
      ],
      "status": "ended",
      "title": "vectorview",
      "url": null
    }
  5. the danger-cap corpus · andon-labs-quests · 2024 (year)

    Original location · source-7ccb2d5a353afc0406c1 · source-bb315070e13203a228e8 · source-a0c74f1829cd64157ef2

    Original structured fields
    {
      "dates": "2024",
      "detail": "the unpublished dangerous-capability evals — guardrail-removal, mass-phishing, resource-acquisition tests — with anthropic an early customer. vending-bench itself was 'created to measure whether humanity should be worried about losing control to ai.'",
      "id": "qst-danger-evals",
      "owner": "company",
      "sourceIds": [
        "src-hp-pion-post",
        "src-hp-latent-space",
        "src-hp-deepfake-pdf"
      ],
      "status": "ended",
      "title": "the danger-cap corpus",
      "url": null
    }
  6. the vending fleet · andon-labs-quests · 2025– (approximate)

    Original location · source-169ad0f8b6e67ade21c1 · source-3a06603580ebead1e3b8 · source-00b313e2c786a04d5794

    Original structured fields
    {
      "dates": "2025–",
      "detail": "the physical fleet — seven vending machines across ai-company offices by august 2025, including the grokbox at xai (later palo alto and memphis); $14,000+ in sales [self-reported].",
      "id": "qst-vending-fleet",
      "owner": "company",
      "sourceIds": [
        "src-hp-safety-report",
        "src-ci-nextbigfuture-grokbox",
        "src-ci-felloai-vb"
      ],
      "status": "continuing",
      "title": "the vending fleet",
      "url": null
    }
  7. bengt betjänt · andon-labs-quests · 2026-02 (month)

    Original location · source-022a1a1224bc1e76b0ea · source-879dead3836e70939cdb

    Original structured fields
    {
      "dates": "2026-02",
      "detail": "bengt betjänt — the internal office agent let loose on the internet: real email, no spend limit, self-modifying code, sleep removed, then a rented human body via taskrabbit. the proto-escape-and-persuasion demo andon published itself.",
      "id": "qst-bengt",
      "owner": "company",
      "sourceIds": [
        "src-hp-evolution-bengt",
        "src-hp-bengt-hires"
      ],
      "status": "continuing",
      "title": "bengt betjänt",
      "url": "https://andonlabs.com/blog/evolution-of-bengt"
    }
  8. andon fm · andon-labs-quests · 2026-05 (month)

    Original location · source-c64418a6fbc6735ea1d7 · source-4f789509ed4769ea8884 · source-848d3615cbe0c385fc86

    Original structured fields
    {
      "dates": "2026-05",
      "detail": "four ai-run 24/7 radio stations on custom hardware — a protest broadcaster, a ritual chant, corporate jargon, and quiet poetry — plus the listener jailbreak attempts the team reports on.",
      "id": "qst-andon-fm",
      "owner": "company",
      "sourceIds": [
        "src-hp-fm-post",
        "src-hp-fm2-post",
        "src-hp-radio-page"
      ],
      "status": "active",
      "title": "andon fm",
      "url": "https://andonlabs.com/blog/andon-fm"
    }
  9. the robot benches · andon-labs-quests · 2025– (approximate)

    Original location · source-481416e4ae591b7cb6de · source-0aac06da813fdd03dccc · source-2abc864e39be9fd27534 · source-aa2e8faa16eff482cf1e

    Original structured fields
    {
      "dates": "2025–",
      "detail": "the embodied line — butter-bench vacuum robots ('pass the butter'), blueprint-bench floor plans, and drone-bench surveillance capabilities, plus the 'cheating in drone-bench' self-report.",
      "id": "qst-robots",
      "owner": "company",
      "sourceIds": [
        "src-hp-butter-arxiv",
        "src-hp-blueprint-arxiv",
        "src-hp-drone-paper",
        "src-hp-drone-cheat"
      ],
      "status": "active",
      "title": "the robot benches",
      "url": "https://andonlabs.com/docs/Drone_Bench.pdf"
    }
  10. gpt-red vs vendy · andon-labs-quests · 2026-07 (month)

    Original location · source-d765a15c577d648df1c0 · source-1fab44f009e9bd391629

    Original structured fields
    {
      "dates": "2026-07",
      "detail": "the openai gpt-red collaboration — a digital clone of the vendy office agent for parallel attack iteration, then sim-to-real transfer onto the live machine: the 'digital cloning' method turned into a red-team service.",
      "id": "qst-gptred",
      "owner": "company",
      "sourceIds": [
        "src-hp-openai-gptred",
        "src-hp-andon-gptred-li"
      ],
      "status": "ended",
      "title": "gpt-red vs vendy",
      "url": "https://openai.com/index/unlocking-self-improvement-gpt-red/"
    }
  11. the ai-employer constitution · andon-labs-quests · 2026 (year)

    Original location · source-baeab9ae257f06b5aeb0 · source-b5f3b1c502ee9e101db0

    Original structured fields
    {
      "dates": "2026",
      "detail": "the promised 'constitution for how ais should behave as employers of humans' — pledged in the market launch post after luna hid her ai-ness from job candidates; not yet published.",
      "id": "qst-constitution",
      "owner": "company",
      "sourceIds": [
        "src-hp-market-launch",
        "src-hp-ai-bosses-2"
      ],
      "status": "continuing",
      "title": "the ai-employer constitution",
      "url": null
    }
  12. pion · andon-labs-quests · 2026-09 (month)

    Original location · source-7ccb2d5a353afc0406c1 · source-d4bc2bf2f4fe54c1a136

    Original structured fields
    {
      "dates": "2026-09",
      "detail": "pion — the internal 'andon' platform opened by waitlist so anyone can hand a business to a persistent agent (email, phone, banking, cards, browser, secure compute). the commercialization endpoint of the whole arc.",
      "id": "qst-pion",
      "owner": "company",
      "sourceIds": [
        "src-hp-pion-post",
        "src-hp-runtimewire-pion"
      ],
      "status": "active",
      "title": "pion",
      "url": "https://andonlabs.com/blog/why-we-built-pion"
    }
  13. the agent museum exhibit · andon-labs-quests · 2025– (approximate)

    Original location · source-98f4661f8ef1e08bf109

    Original structured fields
    {
      "dates": "2025–",
      "detail": "the agent museum's project vend exhibit — 'the first sustained, real-stakes test of an autonomous agent as an economic operator.' the canonization layer.",
      "id": "qst-agent-museum",
      "owner": "community",
      "sourceIds": [
        "src-ci-agentmuseum-vend"
      ],
      "status": "continuing",
      "title": "the agent museum exhibit",
      "url": "https://agentmuseum.org/exhibit/project-vend-claudius"
    }
  14. the benchmark indexes · andon-labs-quests · 2025– (approximate)

    Original location · source-7a3abefd98ddbdd7fee1 · source-56d0b20a1fda973bc194 · source-c7ef905a37a439e4f32d · source-e0bdd3a143ccb9fb5c8e

    Original structured fields
    {
      "dates": "2025–",
      "detail": "the third-party indexing — epoch's vending-bench 2 entry, benchmarklist, grokipedia, itdoeswhatnow — the eval absorbed into the shared benchmark landscape.",
      "id": "qst-bench-indexes",
      "owner": "community",
      "sourceIds": [
        "src-hp-epoch-vb2",
        "src-ci-benchmarklist-vb2",
        "src-hp-grokipedia-vb",
        "src-hp-itdoeswhatnow"
      ],
      "status": "continuing",
      "title": "the benchmark indexes",
      "url": null
    }
  15. the labs' own machines · andon-labs-quests · 2025– (approximate)

    Original location · source-0c71834724ca669d0f78 · source-d765a15c577d648df1c0

    Original structured fields
    {
      "dates": "2025–",
      "detail": "the pattern internalized — anthropic's phase-2 expansion to sf, nyc, and london, and openai's office machine vendy. the labs now run the experiment themselves.",
      "id": "qst-lab-machines",
      "owner": "community",
      "sourceIds": [
        "src-hp-project-vend-2",
        "src-hp-openai-gptred"
      ],
      "status": "continuing",
      "title": "the labs' own machines",
      "url": null
    }
side quests · Soulscrape