sources
the catalog — every source behind every claim, dated and typed.
the catalog
every claim on this site carries source ids that resolve here. type marks the evidence role; dates keep the precision the source gave.
- undatedanalytical exhibitagentmuseum.organalytical exhibit; "economic operator" framing
- 2025-10-28orchestrator-vs-executor framingandonlabs.comorchestrator-vs-executor framing; figure/gdm context
- undatedfirst ai-run caféandonlabs.com"first ai-run café"; hall of shame (6,000 napkins, 22.5 kg tomatoes)
- 2026-07-23drone surveillance capabilitiesandonlabs.comdrone surveillance capabilities; "no ai lab will be able to train on this eval"
- 2026-02-10chinese open models on vbandonlabs.comchinese open models on vb
- 2026-05-28slowly lost the plotandroidauthority.com"slowly lost the plot"; gemini $45 sponsorship
- 2026ap feature on the stockholm caféassociated pressap feature on the stockholm café
- 2026-09arxiv.orgarxiv.orgscaffold/scorer confound literature — context for methodology discourse, not andon-specific
- 2026yc w24aventure.vcyc w24; $500k jan-2024
- 2025-11-18third-party indexingbenchmarklist.comthird-party indexing
- 2026-04-23ghaffary visitbloombergghaffary visit; "orders too many candles"
- 2025-07notes grok 4 heavy beat human vending-bench baselinebuiltin.comnotes grok 4 heavy beat human vending-bench baseline
- 2025-12-19stern on squawk on the streetcnbcstern on squawk on the street
- 2026-05$1,000 in four daysfast company$1,000 in four days; andon labs employment caveat
- 2026-04then it forgot the stafffast company"then it forgot the staff"
- 2025-08aug-2025 leaderboard coveragefelloai.comaug-2025 leaderboard coverage; lobby-machine note
- 2026-04-24first-ever store… run by aiforbes"first-ever store… run by ai"; $100k budget detail
- 2026-02-18npr fresh airfreshairarchive.orgnpr fresh air; leads with project vend
- 2026-06interviewfrontiermodels.ccinterview; claude-vs-others safety-behavior observation
- 2026-02post-new yorker retrospectivefuturismpost-new yorker retrospective; catalogs meth/broadsword requests
- 2025-12goes bankrupt after ordering playstation 5 and live fishfuturism"goes bankrupt after ordering playstation 5 and live fish"
- 2025-07-10andon founders on-stagemanusai.hashnode.devandon founders on-stage; grok 4 #1 on vending-bench, "twice the second-place"
- 2026-04199 pts / 286 commentshacker news199 pts / 286 comments; posted by lukaspetersson
- 2026-07meta-thread: negative hn receptionhacker newsmeta-thread: negative hn reception; "publicity stunt" / self-promotion critiques; founder replies
- 2026-07yea what are they trying to test? + andons resource-acquisition replyhacker news"yea what are they trying to test?" + andon's resource-acquisition reply
- 2025its hilarious… print copies for their managershacker news"it's hilarious… print copies for their managers"
- 2025-06-27gap between current capabilities and what the massive hype machine would have us believehacker news"gap between current capabilities and what the massive hype machine would have us believe"
- 2025-06-27main hn thread on project vendhacker newsmain hn thread on project vend
- 2026-01small threadhacker newssmall thread; ties vb to the "titanium cube" fridge
- 2025-06-27distilled sentiment of the hn discussionhndistilled.comdistilled sentiment of the hn discussion
- 2026-08-16a real-life terminatorinc."a real-life terminator"
- 2026-04-21long explainerintuitionlabs.ailong explainer; phase-2 details incl. seymour cash stats
- 2025-11-18same-day vb2 writeupnews.lavx.husame-day vb2 writeup
- 2025-06-3053 ptslesswrong.com53 pts; safety-community discussion; petersson commented
- undatedmetr comparison point for eval-landscape sectionmetr.orgmetr comparison point for eval-landscape section
- 2025-07-23feedback columnnewscientist.comfeedback column; pratchett "whelk stall"
- 2026-02-16lewis-kraus anthropic profilethe new yorkerlewis-kraus anthropic profile; project vend set-piece
- 2025-08-21grok-powered machine in xai office with andon labsnextbigfuture.comgrok-powered machine in xai office with andon labs
- 2025-07$4,694 vs opus 4s $2,077 framingofficechai$4,694 vs opus 4's $2,077 framing
- 2025-07-01pc gamer take, nick evansonpcgamer.compc gamer take, nick evanson
- 2026-05-19radio-industry coveragenews.radio-online.comradio-industry coverage
- 2026harness-as-hidden-variable critique — context onlykdd-eval-workshop.github.ioharness-as-hidden-variable critique — context only
- 2025-07-11benchmark slate inclscientificamerican.combenchmark slate incl. vending-bench
- 2026-08-04syndication of the knight pieceseattletimes.comsyndication of the knight piece
- 2026-06knight pieceseattletimes.comknight piece; $7.5k/mo rent; sonnet 4.6; store "limping along"
- 2026-09-11drew magary store visitsfgate.comdrew magary store visit; snarky; "fort sumter of the coming cyberwar"
- 2026-08-17employee interviewsfstandard.comemployee interview; ~$40k loss; pronoun/ownership critique
- 2026-06human-ceo coup detailssignalcast.apphuman-ceo coup details; 164,000 fraudulent votes
- 2025-06-27willison excerpts+notessimonwillison.netwillison excerpts+notes; dev-community amplifier
- 2026-05-14somewhere between publicity stunts and ethics experimentsskh.news"somewhere between publicity stunts and ethics experiments"; sweden work-hours point
- 2025-11-01petersson quoted on model selectiontechcrunchpetersson quoted on model selection
- 2025-11butter-bench explainer + blueprint-bench linktechspot.combutter-bench explainer + blueprint-bench link
- 2026-05-17competent to unhingedthe-decoder.com"competent to unhinged"
- 2026-08-23firing recap + model-replay findingthe-decoder.comfiring recap + model-replay finding
- 2025-12-19phase-2 analysisthe-decoder.comphase-2 analysis
- 2026-05-18dj claude tried to quittheprint.in"'dj claude' tried to quit"
- 2026-08-15had to be reminded of its own rulesthe next web"had to be reminded of its own rules"; three open questions
- 2025-11robot exorcism protocol inner monologuetomshardware.com"robot exorcism protocol" inner monologue
- 2026-04-21is this the future of retail?usatoday.com"is this the future of retail?"; petersson interview
- 2025-07vending-industry coveragevendingmarketwatch.comvending-industry coverage
- 2025-12skeptical stress test framingthe vergeskeptical "stress test" framing
- 2025-12wsj recapwebpronews.comwsj recap; katherine long's ultra-capitalist free-for-all
- 2025-07-10launch context: antisemitism controversy same weekwiredlaunch context: antisemitism controversy same week
- 2025-11-16transcript inclpodscripts.cotranscript incl. blue-blazer anecdote
- 2026-08-044 months of ai managing 5 humansandonlabs.com4 months of ai managing 5 humans
- 2026-08-14firing/hiring analysisandonlabs.comfiring/hiring analysis
- 2026-09-16founding trio historyai.engineerfounding trio history; fröberg shareholder note
- 2026-09-16speaker bioai.engineerspeaker bio; esa/comma.ai detail
- 2026talk transcriptai.engineertalk transcript; simulation awareness
- 2025-11-17andons 60 minutes postlinkedinandon's 60 minutes post
- 2026-07-15andons gpt-red collab postlinkedinandon's gpt-red collab post
- 2026-09-16mission, saos, eval list, trusted by, © andon labs incandonlabs.commission, saos, eval list, "trusted by," © andon labs inc.
- 2026-09-16worldview, comp range, sf in-person, press linksandonlabs.comworldview, comp range, sf in-person, press links
- 2026-09-16headcount 18linkedinheadcount 18; london hq listing; executives
- 2026-09-16complete dated publications indexandonlabs.comcomplete dated publications index
- 2025-12project vend history recaplinkedinproject vend history recap; "first employees axel and hanna"; seymour cash vote
- 2026-09-07astra vs fableandonlabs.comastra vs fable
- 2025-07-10grok 4 livestream appearancelinkedingrok 4 livestream appearance
- 2024-12-04backlund co-founder announcementlinkedinbacklund co-founder announcement
- 2026-09-16profile: quantumblack/mckinsey, seb thesislinkedinprofile: quantumblack/mckinsey, seb thesis
- 2026-02-13bengt rents human via taskrabbitandonlabs.combengt rents human via taskrabbit
- 2025-09-24blueprint-bencharxiv.orgblueprint-bench; author list
- 2025-10-23butter-bencharxiv.orgbutter-bench; author list; 40% vs 95%
- 2026-05-04andon café launchandonlabs.comandon café launch; mona
- 2026-09-16investor listcb insightsinvestor list
- 2025-11-1660 minutes/overtime on claudiuscbsnews.com60 minutes/overtime on claudius
- 2025-08-16cognitive revolution ep + transcriptcognitiverevolution.aicognitive revolution ep + transcript; seldon
- 2026-09-16$125k yc mar 2024app.dealroom.co$125k yc mar 2024; vectorview alias
- 2025-06-28project vend coveragethe-decoder.comproject vend coverage
- 2024-12-02first publicationandonlabs.comfirst publication; deepfake agentic eval
- 2026-08-05cheating postandonlabs.comcheating post
- 2026-07-23drone-bench paperandonlabs.comdrone-bench paper
- 2026-09-16vb2 design + $63k human estimateepoch.aivb2 design + $63k human estimate
- 2026-02-09guardrails-off bengtandonlabs.comguardrails-off bengt; sleep removed; taskrabbit arbitrage
- 2026-06-09ftc threatsandonlabs.comftc threats; "calls the fbi again (almost)"
- 2026-05-13four ai radio stationsandonlabs.comfour ai radio stations
- 2026-07-07six weeks laterandonlabs.comsix weeks later; jailbreak attempts
- 2026-06-02fortune profilefortunefortune profile
- 2024-12-03departure retrospectivelinkedindeparture retrospective
- 2026-09-16klarna biolinkedinklarna bio; "$2.2m raised" claim
- 2023-11-21vectorview founding + yc acceptancelinkedinvectorview founding + yc acceptance
- 2026-09-16investors inclfundraisingfox.cominvestors incl. wave ventures; yc $500k
- 2026-06-30gemini 3andonlabs.comgemini 3.1 pro café eval
- 2026-04-22bad behavior not necessaryandonlabs.com"bad behavior not necessary"
- 2026-09-16secondary summary of vb historygrokipedia.comsecondary summary of vb history
- 2026-09-14eliza strickland featurespectrum.ieee.orgeliza strickland feature; petersson quotes; café cheese-toast
- 2026-09-16company profile pdfintuitionlabs.aicompany profile pdf; name-origin claim
- 2026-09-16vb explaineritdoeswhatnow.comvb explainer; release-date cross-check
- 2026-06-04full transcriptlatent spacefull transcript; founding, vend, arena, bengt, luna
- 2026-04-103-yr lease 2102 union standonlabs.com3-yr lease 2102 union st; luna hides ai-ness
- 2026-09-16luna architecture (200k ctx, memory, subagents)andonlabs.comluna architecture (200k ctx, memory, subagents); "leah"
- 2025explainer on vend+vb pairingmediumexplainer on vend+vb pairing
- 2026name-origin claim (andon cord)metro-tribune.comname-origin claim (andon cord)
- 2025-11-01butter-bench coveragemezha.netbutter-bench coverage; petersson interview ref
- 2025-03-03misalignment-in-the-wild highlightslinkedinmisalignment-in-the-wild highlights
- 2026-04-21nyt on andon marketthe new york timesnyt on andon market
- 2026-05-28observer reprintluxematin.comobserver reprint; $7.5k/mo lease, $100k, backlund kth/epfl, neurips claim
- 2026-07gpt-red breaks vendy via andon simopenaigpt-red breaks vendy via andon sim
- 2026-02-04arena misbehavior beginsandonlabs.comarena misbehavior begins
- 2026-05-28better alignment/worse performanceandonlabs.combetter alignment/worse performance
- 2026-07-28best capitalist, misalignedandonlabs.combest capitalist, misaligned
- 2026-09-16personal posts + podcast linklukaspetersson.compersonal posts + podcast link
- 2026-09-16lukaspetersson.comlukaspetersson.comceo; google/disney/eth/lund bio
- 2026-09-16profile datalinkedinprofile data
- 2025-02-27petersson fbi quotelinkedinpetersson fbi quote
- 2026-09-16dated cv incllukaspetersson.comdated cv incl. "andon labs dec 2023", esa, comma.ai
- 2026-09-14pion launchandonlabs.compion launch; vb origin in danger-cap evals; skräckblandad förtjusning
- 2026-09-16deal history + 7 investorspitchbookdeal history + 7 investors
- 2025-06-27anthropic phase-1 write-upanthropicanthropic phase-1 write-up
- 2025-12-18phase 2: sonnet 4anthropicphase 2: sonnet 4.x, seymour cash, clothius, 3 sites
- 2025-07-02petersson interviewpymntspetersson interview; founders as restockers
- 2026-09-16andon fm current lineupandonlabs.comandon fm current lineup
- 2026-09-16vetted profilerl-list.comvetted profile; funding/founder-claim caveats
- 2026-09-14pion trade coverageruntimewire.compion trade coverage
- 2025-08-287 machines, $14k sales, oversight systemandonlabs.com7 machines, $14k sales, oversight system
- 2025-06-28project vend coveragetechcrunchproject vend coverage
- 2026-05-24times uk on caféthetimes.comtimes uk on café; 3,000 gloves
- 2026-08-14first ai firingtime.comfirst ai firing; petersson quotes
- 2025-06-27time exclusivetime.comtime exclusive; freeman/troy quotes
- 2026writeup of talktokenless.techwriteup of talk; digital-cloning method
- 2026-09-16toyotas own andon explainermag.toyota.co.uktoyota's own andon explainer
- 2026-09-16multi-agent arenaandonlabs.commulti-agent arena; sleep_until_tomorrow traces
- 2025-02-21vending-bench paper (submitted feb 21arxiv.orgvending-bench paper (submitted feb 21; andon lists 2/16)
- 2025-02-27vb1 announcementlinkedinvb1 announcement
- 2026-09-16deprecated vb1 pageandonlabs.comdeprecated vb1 page; leaderboard incl. grok 4/gpt-5
- 2025-11-18vb2+arena reveallinkedinvb2+arena reveal; gemini 3 wins
- 2025-11-18vb2 mechanics + leaderboardandonlabs.comvb2 mechanics + leaderboard
- 2025-06-28project vend coverageventurebeatproject vend coverage
- 2026-05-15verge on andon fmthe vergeverge on andon fm
- 2026-09-16andon term/cord backgroundwikipediaandon term/cord background
- 2025-12-18wsj on vend phase 2the wall street journalwsj on vend phase 2
- 2026-09-16w24, founded 2023, team 11, sf, partner diana hu, news linksycombinator.comw24, founded 2023, team 11, sf, partner diana hu, news links
AI-drafted at Ben Guo's direct request and credited to Hraness; every claim links to its cataloged source.