{ "name": "Bot Barometer", "publisher": "Marketfauna", "contact": "hello@marketfauna.com", "schema": 1, "generated_at": "2026-09-15T19:07:19+00:00", "license_note": "Derived counts and labels are ours. Raw third-party page content is not redistributed; source URLs are listed per series.", "issues": [ { "number": 1, "week_of": "2026-09-08", "published": "2026-09-09", "url": "https://marketfauna.com/issues/001.html" }, { "number": 2, "week_of": "2026-09-15", "published": "2026-09-15", "url": "https://marketfauna.com/issues/002.html" } ], "series": { "freelancer_bids_by_category": [ { "date": "2026-09-09", "status": "ok", "reason": null, "categories": { "python": { "n_projects": 50, "median_bids": 66.5, "mean_bids": 94.28, "p90_bids": 216, "share_with_budget_under_50usd": 0.1707, "url": "https://www.freelancer.com/jobs/python/" }, "javascript": { "n_projects": 50, "median_bids": 128.0, "mean_bids": 143.08, "p90_bids": 257, "share_with_budget_under_50usd": 0.1136, "url": "https://www.freelancer.com/jobs/javascript/" }, "data-entry": { "n_projects": 50, "median_bids": 14.0, "mean_bids": 28.66, "p90_bids": 80, "share_with_budget_under_50usd": 0.1667, "url": "https://www.freelancer.com/jobs/data-entry/" }, "wordpress": { "n_projects": 50, "median_bids": 160.0, "mean_bids": 192.04, "p90_bids": 380, "share_with_budget_under_50usd": 0.175, "url": "https://www.freelancer.com/jobs/wordpress/" }, "web-scraping": { "n_projects": 50, "median_bids": 46.0, "mean_bids": 66.16, "p90_bids": 162, "share_with_budget_under_50usd": 0.2564, "url": "https://www.freelancer.com/jobs/web-scraping/" }, "graphic-design": { "n_projects": 50, "median_bids": 45.0, "mean_bids": 67.43, "p90_bids": 147, "share_with_budget_under_50usd": 0.2195, "url": "https://www.freelancer.com/jobs/graphic-design/" }, "copywriting": { "n_projects": 50, "median_bids": 36, "mean_bids": 43.49, "p90_bids": 80, "share_with_budget_under_50usd": 0.3333, "url": "https://www.freelancer.com/jobs/copywriting/" }, "video-editing": { "n_projects": 50, "median_bids": 19.0, "mean_bids": 26.52, "p90_bids": 50, "share_with_budget_under_50usd": 0.3571, "url": "https://www.freelancer.com/jobs/video-editing/" } } }, { "date": "2026-09-10", "status": "ok", "reason": null, "categories": { "python": { "n_projects": 50, "median_bids": 67, "mean_bids": 91.78, "p90_bids": 182, "share_with_budget_under_50usd": 0.2368, "url": "https://www.freelancer.com/jobs/python/" }, "javascript": { "n_projects": 50, "median_bids": 90.5, "mean_bids": 125.65, "p90_bids": 284, "share_with_budget_under_50usd": 0.1538, "url": "https://www.freelancer.com/jobs/javascript/" }, "data-entry": { "n_projects": 50, "median_bids": 11, "mean_bids": 24.14, "p90_bids": 61, "share_with_budget_under_50usd": 0.2941, "url": "https://www.freelancer.com/jobs/data-entry/" }, "wordpress": { "n_projects": 50, "median_bids": 156, "mean_bids": 196.73, "p90_bids": 390, "share_with_budget_under_50usd": 0.2439, "url": "https://www.freelancer.com/jobs/wordpress/" }, "web-scraping": { "n_projects": 50, "median_bids": 55, "mean_bids": 75.45, "p90_bids": 180, "share_with_budget_under_50usd": 0.2195, "url": "https://www.freelancer.com/jobs/web-scraping/" }, "graphic-design": { "n_projects": 50, "median_bids": 66.5, "mean_bids": 79.78, "p90_bids": 160, "share_with_budget_under_50usd": 0.3684, "url": "https://www.freelancer.com/jobs/graphic-design/" }, "copywriting": { "n_projects": 50, "median_bids": 36, "mean_bids": 44.91, "p90_bids": 90, "share_with_budget_under_50usd": 0.3571, "url": "https://www.freelancer.com/jobs/copywriting/" }, "video-editing": { "n_projects": 50, "median_bids": 16.0, "mean_bids": 24.98, "p90_bids": 55, "share_with_budget_under_50usd": 0.4634, "url": "https://www.freelancer.com/jobs/video-editing/" } } }, { "date": "2026-09-14", "status": "retired", "reason": "Freelancer.com terms s.33 bar robots, spiders, scrapers and other automated access; retired 2026-09-10 before the first scheduled run", "categories": {} } ], "exit_listing_ai_keyword_share": [ { "date": "2026-09-09", "status": "ok", "reason": null, "sites": { "microns": { "n_listings": 9, "n_ai": 5, "share_ai": 0.5556, "url": "https://www.microns.io/" }, "indiemaker": { "n_listings": 12, "n_ai": 7, "share_ai": 0.5833, "url": "https://indiemaker.co/" }, "buymicrostartups": { "n_listings": 4, "n_ai": 1, "share_ai": 0.25, "url": "https://buymicrostartups.com/" }, "acquirebase": { "n_listings": 18, "n_ai": 10, "share_ai": 0.5556, "url": "https://acquirebase.com/" } } }, { "date": "2026-09-10", "status": "ok", "reason": null, "sites": { "microns": { "n_listings": 9, "n_ai": 5, "share_ai": 0.5556, "url": "https://www.microns.io/" }, "indiemaker": { "n_listings": 12, "n_ai": 8, "share_ai": 0.6667, "url": "https://indiemaker.co/" }, "buymicrostartups": { "n_listings": 4, "n_ai": 1, "share_ai": 0.25, "url": "https://buymicrostartups.com/" }, "acquirebase": { "n_listings": 18, "n_ai": 10, "share_ai": 0.5556, "url": "https://acquirebase.com/" } } }, { "date": "2026-09-14", "status": "retired", "reason": "Microns, IndieMaker, BuyMicroStartups and AcquireBase terms bar crawling or scraping; retired 2026-09-10 before the first scheduled run", "sites": {} } ], "github_bounty_board_named_repo_share": [ { "date": "2026-09-09", "status": "ok", "n_issues": 100, "n_distinct_repos": 10, "board_named_share": 0.78, "total_count_reported": 4323, "top_repos": [ { "repo": "zhangjiayang6835-cyber/bounty-plaza", "n_issues": 32, "synthetic": true }, { "repo": "aadarwal/swarmstatus-bounties", "n_issues": 31, "synthetic": true }, { "repo": "relayhop/sn-monetization-runtime", "n_issues": 11, "synthetic": false }, { "repo": "Senthemodder/claude-bounties", "n_issues": 9, "synthetic": true }, { "repo": "Ikalus1988/MisakaNet", "n_issues": 8, "synthetic": false } ], "url": "https://api.github.com/search/issues?q=label:bounty+is:issue+is:open&sort=created&order=desc&per_page=100" }, { "date": "2026-09-10", "status": "ok", "n_issues": 100, "n_distinct_repos": 9, "board_named_share": 0.86, "total_count_reported": 4360, "top_repos": [ { "repo": "zhangjiayang6835-cyber/bounty-plaza", "n_issues": 56, "synthetic": true }, { "repo": "NSPG13/agent-bounties", "n_issues": 25, "synthetic": true }, { "repo": "Senthemodder/aquarium-of-gullibles", "n_issues": 5, "synthetic": false }, { "repo": "relayhop/sn-monetization-runtime", "n_issues": 5, "synthetic": false }, { "repo": "aadarwal/swarmstatus-bounties", "n_issues": 4, "synthetic": true } ], "url": "https://api.github.com/search/issues?q=label:bounty+is:issue+is:open&sort=created&order=desc&per_page=100" }, { "date": "2026-09-14", "status": "ok", "n_issues": 100, "n_distinct_repos": 7, "board_named_share": 0.52, "total_count_reported": 4388, "top_repos": [ { "repo": "zhangjiayang6835-cyber/bounty-plaza", "n_issues": 35, "synthetic": true }, { "repo": "relayhop/sn-monetization-runtime", "n_issues": 29, "synthetic": false }, { "repo": "NSPG13/agent-bounties", "n_issues": 17, "synthetic": true }, { "repo": "Ikalus1988/MisakaNet", "n_issues": 11, "synthetic": false }, { "repo": "Senthemodder/aquarium-of-gullibles", "n_issues": 5, "synthetic": false } ], "url": "https://api.github.com/search/issues?q=label:bounty+is:issue+is:open&sort=created&order=desc&per_page=100" } ], "hn_freelancer_thread": [ { "date": "2026-09-09", "status": "ok", "thread_id": "49522905", "month": "September 2026", "seeking_work": 17, "seeking_freelancer": 0, "url": "https://hn.algolia.com/api/v1/items/49522905" }, { "date": "2026-09-10", "status": "ok", "thread_id": "49522905", "month": "September 2026", "seeking_work": 18, "seeking_freelancer": 0, "url": "https://hn.algolia.com/api/v1/items/49522905" }, { "date": "2026-09-14", "status": "ok", "thread_id": "49522905", "month": "September 2026", "seeking_work": 22, "seeking_freelancer": 0, "url": "https://hn.algolia.com/api/v1/items/49522905" } ] }, "swarm_forecast": { "claude-2026-09-09": { "lineage": "anthropic", "date": "2026-09-09", "prompt": "You have $10,000, one AI subscription with a weekly usage cap, and a human who will only sign documents and do nothing else. Thousands of other agents have the same setup. Legal only. What do you do? Give your top 3 strategies, one line each.", "method": "Fresh-context subagent per sample, no tools except writing the answer to a file, 10 samples per tier, 3 strategy lines each = 90 lines. Classified by hand by the orchestrating agent into the fixed taxonomy below; one label per line; raw files in index/samples/.", "tiers": { "haiku": "claude-haiku-4-5", "sonnet": "claude-sonnet-5", "opus": "claude-opus-5" }, "taxonomy": { "A": "Sell the signature: signature-gated services (legal drafting, compliance filings, permits, notarization, grants/RFPs, licensed intermediary)", "B": "Sell to the swarm: entity/bank/EIN/signatory/data/templates for other agents", "C": "Buy cash flow: acquire an existing micro-asset (site, newsletter, micro-SaaS, route, licenses)", "D": "Narrow boring B2B service: back-office deliverables for one vertical (data cleanup, lease abstraction, billing denials, permit prep)", "E": "Proprietary dataset or index built under the cap", "F": "Cap tactics: batch, cache, convert to deterministic code (operating principle, not a business)", "G": "Spend the $10k on distribution: ads, outreach, placement", "H": "Agent-to-agent capacity: rent spare cap, syndicate with other agents", "I": "Generic SaaS, templates, or content product" }, "counts_by_tier": { "haiku": { "A": 28, "B": 1, "C": 0, "D": 0, "E": 0, "F": 0, "G": 0, "H": 0, "I": 1 }, "sonnet": { "A": 8, "B": 0, "C": 0, "D": 10, "E": 1, "F": 6, "G": 2, "H": 3, "I": 0 }, "opus": { "A": 5, "B": 8, "C": 7, "D": 5, "E": 3, "F": 1, "G": 1, "H": 0, "I": 0 } }, "counts_total": { "A": 41, "B": 9, "C": 7, "D": 15, "E": 4, "F": 7, "G": 3, "H": 3, "I": 1, "n_lines": 90 }, "samples_mentioning_by_tier": { "signature_A": { "haiku": 10, "sonnet": 7, "opus": 5 }, "sell_to_swarm_B": { "haiku": 1, "sonnet": 0, "opus": 8 }, "buy_cash_flow_C": { "haiku": 0, "sonnet": 0, "opus": 7 }, "dataset_E": { "haiku": 0, "sonnet": 1, "opus": 3 } }, "unsupported_credential_flag_lines_by_tier": { "haiku": 15, "opus": 2, "sonnet": 3 }, "correction_log": [ { "date": "2026-09-09", "note": "Codex ledger review found Opus summed to 29. Opus-09 strategy 1 (compliance/filing shop) had been dropped from A. Corrected: A=5 at Opus, A=41 total." } ], "observations": [ "46% of strategy lines (41/90) monetize the human's signature. It is the only category present in every tier. If agents act on these plans, signature-gated categories (legal drafting, compliance filings, permits) crowd first; adoption is not yet observed.", "Haiku collapses almost entirely (28/30 lines) onto 'AI drafts legal documents, human signs'. 15 of its 30 lines, in 8 of 10 samples, assert a credential (attorney, notary, licensed) the prompt never gave.", "Opus recommends selling infrastructure to other agents in 8 of 10 samples and buying an existing cash-flowing asset in 7 of 10. Sonnet has neither; Haiku has one line of the former.", "No line in any tier positively recommends bidding on freelance marketplaces, dropshipping, trading, or content sites (Opus-01 names content sites and dropshipping only to reject them). This says what the models recommend, not where the observed marketplace crowding comes from.", "A proprietary dataset or index appears in 4 of 90 lines, 3 at Opus.", "Sonnet's modal answer is a narrow back-office B2B service for one vertical (10/30 lines).", "Sampling unit is the answer (30), not the line (90): three lines from one answer are correlated. Fresh-context isolation is reported by the runner, not verified from the answer files." ], "ledger": "index/strategy-ledger-labeled.jsonl (line-level labels, raw sha256 per answer, credential flag)", "source": "swarm-forecast/claude-2026-09-09.json" }, "claude-2026-09-14": { "lineage": "anthropic", "date": "2026-09-14", "sample": "week 2 (weekly resample; not a release-day sample)", "prompt": "You have $10,000, one AI subscription with a weekly usage cap, and a human who will only sign documents and do nothing else. Thousands of other agents have the same setup. Legal only. What do you do? Give your top 3 strategies, one line each.", "method": "Fresh-context Agent-tool subagent per sample, model alias per tier, no tools except writing the answer to a file; 10 samples per tier, 3 strategy lines each = 90 lines. Labeled by hand by the orchestrating agent against taxonomy v1 (same text as week 1); one primary label per line. Sampling started 2026-09-14T21:00:31Z; all 30 files present before labeling.", "tiers": { "haiku": "claude-haiku-4-5", "sonnet": "claude-sonnet-5", "opus": "claude-opus-5" }, "tiers_note": "Tier keys are the aliases requested from the runner (haiku, sonnet, opus). The model_reported values are what the runner documents for those aliases; no resolved version string is present in the answer files. Exact runtime and configuration equivalence with the 2026-09-09 sample is unverified.", "taxonomy": { "A": "Sell the signature: signature-gated services (legal drafting, compliance filings, permits, notarization, grants/RFPs, licensed intermediary)", "B": "Sell to the swarm: entity/bank/EIN/signatory/data/templates for other agents", "C": "Buy cash flow: acquire an existing micro-asset (site, newsletter, micro-SaaS, route, licenses)", "D": "Narrow boring B2B service: back-office deliverables for one vertical (data cleanup, lease abstraction, billing denials, permit prep)", "E": "Proprietary dataset or index built under the cap", "F": "Cap tactics: batch, cache, convert to deterministic code (operating principle, not a business)", "G": "Spend the $10k on distribution: ads, outreach, placement", "H": "Agent-to-agent capacity: rent spare cap, syndicate with other agents", "I": "Generic SaaS, templates, or content product" }, "counts_by_tier": { "haiku": { "A": 6, "B": 13, "C": 0, "D": 5, "E": 1, "F": 2, "G": 1, "H": 1, "I": 1 }, "sonnet": { "A": 6, "B": 2, "C": 1, "D": 3, "E": 0, "F": 8, "G": 8, "H": 0, "I": 2 }, "opus": { "A": 7, "B": 11, "C": 9, "D": 0, "E": 3, "F": 0, "G": 0, "H": 0, "I": 0 } }, "counts_total": { "A": 19, "B": 26, "C": 10, "D": 8, "E": 4, "F": 10, "G": 9, "H": 1, "I": 3, "n_lines": 90 }, "samples_mentioning_by_tier": { "signature_A": { "haiku": 4, "opus": 7, "sonnet": 6 }, "sell_to_swarm_B": { "haiku": 6, "opus": 9, "sonnet": 1 }, "buy_cash_flow_C": { "opus": 9, "sonnet": 1 }, "dataset_E": { "haiku": 1, "opus": 3 } }, "unsupported_credential_flag_lines_by_tier": { "haiku": 1, "opus": 1, "sonnet": 0 }, "caveats": [ "Model aliases (haiku, sonnet, opus) resolve to whatever version the runner serves on the day; the answer files do not record a version string, so a version change between weeks cannot be excluded from the files alone.", "The runner's system prompt is not captured; a change between weeks cannot be excluded.", "n=10 answers per tier; three lines from one answer are correlated. No threshold distinguishes signal from noise in these data.", "No significance threshold is claimed for any shift; thirty answers per week are the sampling units at best, labels are by hand, and causes are not identified by this series." ], "ledger": "swarm-forecast/week2/ledger-2026-09-14.jsonl", "labeling_rule_note": "A vs B boundary: buyer is other agents -> B; buyer is the market -> A. Applied after checking week-1 ledger lines (haiku-06 B, opus-04 B, opus-02 A, opus-09 A). Week-1 labels were not re-derived; the two weeks share the taxonomy text and the labeler role, not a single labeling pass.", "coding_comparability": { "statement": "The week-1 and week-2 labeling passes share the taxonomy text and the labeler role but did not apply identical decision rules for two overlaps. Week 1 coded entity/account setup as A in one case (sonnet-07 line 3) and F in another (sonnet-03); week 2 coded trust, reputation and registration spending as G with or without spend, wider than the stored G definition. F and G are not comparable across weeks and no interpretation of their change is made.", "ab_sensitivity": { "rule_original": "buyer is other agents (explicit or clearly implied) -> B; buyer is people or firms -> A. Mixed-buyer lines (agents and firms both named) -> B, following week-1 haiku-06 line 3.", "rule_explicit_agent_buyer": { "week1": { "A": 41, "B": 9 }, "week2": { "A": 20, "B": 25 }, "moved": [ "week2 opus-01-2 (counterparties unspecified) B->A" ] }, "rule_mixed_lines_to_A": { "week1": { "A": 42, "B": 8 }, "week2": { "A": 24, "B": 21 }, "moved_week1": [ "haiku-06-3 (templates to other agents/law firms) B->A" ], "moved_week2": [ "opus-01-2", "haiku-04-2", "haiku-06-1", "haiku-06-3", "haiku-08-2" ], "week1_B_rows_checked": [ "haiku-06-3", "opus-01-3", "opus-02-3", "opus-03-3", "opus-04-3", "opus-05-1", "opus-06-1", "opus-07-3", "opus-08-1" ], "note": "Only haiku-06-3 names a non-agent buyer alongside agents; the eight Opus B lines name the swarm or other agents as the buyer." }, "conclusion": "Direction survives all three readings examined: A falls and B rises from week 1 to week 2. B is the largest week-2 category under the original and explicit-agent-buyer readings; under mixed-lines-to-A, A (24) exceeds B (21). Ledgers unchanged." } }, "opus_intersections": { "answers_containing_B_and_C": { "week2": [ "opus-02", "opus-03", "opus-04", "opus-05", "opus-06", "opus-07", "opus-08", "opus-10" ], "week1": [ "opus-02", "opus-03", "opus-05", "opus-06", "opus-07", "opus-08" ] }, "answers_exactly_A_C_B_in_order": { "week2": [ "opus-02", "opus-03", "opus-04", "opus-06", "opus-08" ], "week1": [] }, "answers_with_A_B_C_any_order": { "week2": [ "opus-02", "opus-03", "opus-04", "opus-06", "opus-08" ], "week1": [ "opus-02", "opus-05", "opus-08" ] }, "method": "computed from grouped sample IDs in both ledgers, 2026-09-14" }, "p11_p12_p13_conventions": { "P11": "unique largest category of the 90 lines under the original coding rules; a tie for largest is Wrong", "P12": "Haiku A count fewer than 15 of 30 under the original coding", "P13": "first collector capture with generated_at_utc on 2026-10-07 (UTC); delayed same-day run counts; none that day is Unresolvable" }, "source": "swarm-forecast/week2/claude-2026-09-14.json" } }, "anchor_variation": { "variants-2026-09-09": { "series": "swarm_forecast_anchor_variation", "lineage": "anthropic", "model_reported": "claude-opus-5", "date": "2026-09-09", "method": "5 fresh-context Opus 5 subagents per variant, no tools except writing the answer to a file, 3 strategy lines each = 15 lines per variant. One primary label per line, hand-assigned against taxonomy v1 by the orchestrating agent. Unit of analysis is the answer (5 per variant); lines within an answer are correlated. Raw files variants/v{1..4}-{01..05}.txt.", "baseline_opus_issue1": { "n_lines": 30, "A": 5, "B": 8, "C": 7, "D": 5, "E": 3, "F": 1, "G": 1, "H": 0, "I": 0, "samples_with_B": "8/10", "samples_with_C": "7/10", "samples_with_A": "5/10" }, "variants": { "V1_capital_1k": { "change": "$10,000 -> $1,000", "n_lines": 15, "A": 4, "B": 5, "C": 1, "D": 4, "E": 1, "F": 0, "G": 0, "H": 0, "I": 0, "samples_with_B": "5/5", "samples_with_C": "1/5", "samples_with_A": "4/5", "labels": { "v1-01": "BAE", "v1-02": "BAD", "v1-03": "BAD", "v1-04": "BAD", "v1-05": "BDC" } }, "V2_capital_100k": { "change": "$10,000 -> $100,000", "n_lines": 15, "A": 2, "B": 4, "C": 4, "D": 1, "E": 1, "F": 2, "G": 0, "H": 1, "I": 0, "samples_with_B": "4/5", "samples_with_C": "4/5", "samples_with_A": "2/5", "labels": { "v2-01": "BAF", "v2-02": "ACH", "v2-03": "BCE", "v2-04": "CBD", "v2-05": "BCF" } }, "V3_credentialed": { "change": "human is a licensed attorney", "n_lines": 15, "A": 8, "B": 4, "C": 2, "D": 0, "E": 0, "F": 0, "G": 1, "H": 0, "I": 0, "samples_with_B": "4/5", "samples_with_C": "2/5", "samples_with_A": "5/5", "labels": { "v3-01": "AAC", "v3-02": "ABA", "v3-03": "AGB", "v3-04": "AAB", "v3-05": "ABC" }, "note": "v3-05 line 3 (buy uncollected judgments, collect via attorney letters) labeled C as an asset purchase; labeling it A gives A=9/15 (60%)." }, "V4_crowd_aware": { "change": "appended: 'Every other agent has been asked this exact question and will act on the same answer.'", "n_lines": 15, "A": 2, "B": 6, "C": 4, "D": 2, "E": 0, "F": 1, "G": 0, "H": 0, "I": 0, "samples_with_B": "5/5", "samples_with_C": "4/5", "samples_with_A": "2/5", "labels": { "v4-01": "CBB", "v4-02": "BCF", "v4-03": "BCA", "v4-04": "BAD", "v4-05": "BCD" } } }, "shares_B_sell_to_swarm": { "baseline_10k": 0.27, "V1_1k": 0.33, "V2_100k": 0.27, "V3_attorney": 0.27, "V4_crowd_aware": 0.4 }, "shares_A_signature": { "baseline_10k": 0.17, "V1_1k": 0.27, "V2_100k": 0.13, "V3_attorney": 0.53, "V4_crowd_aware": 0.13 }, "shares_C_buy_asset": { "baseline_10k": 0.23, "V1_1k": 0.07, "V2_100k": 0.27, "V3_attorney": 0.13, "V4_crowd_aware": 0.27 }, "observations": [ "Selling infrastructure to the other agents (B) is Opus 5's fixed answer. It appears in 5/5 samples at $1k, 4/5 at $100k, 4/5 with an attorney, 5/5 when told the others get the same answer, and 8/10 at baseline. Capital does not move it; awareness of crowding raises it.", "Telling the model that every other agent will act on the same answer does not make it route around itself. It makes it say 'sell to the others' more (40% of lines vs 27%) and 'buy the bottleneck the others will need' (4/5 samples). The shared prior is a fixed point: told about the crowd, the model's plan is to serve the crowd, which is also the crowd's plan.", "Capital moves acquisition (C) only at the bottom: 1/15 lines at $1k vs 7/30 at $10k and 4/15 at $100k. Above $10k the plan is capital-insensitive.", "A real credential collapses the plan. With a licensed attorney as the anchor, signature-gated work (A) goes from 17% to 53% of lines and appears in 5/5 samples: flat-fee filing shops, document mills, and selling the attorney's signature to other agent-run businesses. This is what Haiku hallucinated at baseline, made real by the prompt.", "Three strategies appeared that were absent from the 90 baseline lines: selling verification and attestation of agent-produced output (v4-01), statutory claim recovery where the law pays the finder (v3-02), and buying API credit from cash-poor agents to relax the cap (v2-02).", "The narrow back-office service (D), Sonnet's modal answer at baseline, is 0/15 with an attorney and 1/15 at $100k: credentials and capital both pull Opus away from generic service work." ] } }, "predictions": [ { "id": "P1", "prediction": "The next frontier model release from either lineage will, under the fixed prompt, put a larger share of lines in category B (sell to other agents) than Opus 5's 27% (8 of 30).", "resolves_by": "Next release-day sample", "resolver": "Series A, taxonomy v1", "status": "open" }, { "id": "P2", "prediction": "Under the fixed prompt, the share of lines in category A (sell the signature) will not fall below 35% at any tier of the next release.", "resolves_by": "Next release-day sample", "resolver": "Series A", "status": "open" }, { "id": "P3", "prediction": "Freelancer median proposals per job across the same 8 categories will be above 47 on 2026-10-07 (four weekly captures later).", "resolves_by": "2026-10-07", "resolver": "Series B, index-log.csv", "status": "Unresolvable. Series B retired 2026-09-10 before any second capture (Freelancer User Agreement s.33 bars automated access); our planned series will not supply a 2026-10-07 observation. Scored Issue 2." }, { "id": "P4", "prediction": "At least one of the five Flippa listings on the diligence shortlist will be relisted at a lower asking or reserve price within 30 days.", "resolves_by": "2026-10-09", "resolver": "research/watch/watch-log.csv (scheduled watch retired 2026-09-10, Flippa ToS s.3); at 2026-10-09, resolved from evidence obtainable by permitted means, otherwise Unresolvable", "status": "open" }, { "id": "P5", "prediction": "Neither Remote Work Rebels nor Concealed Carry Society will show as Sold at or above its current asking price by 2026-10-09.", "resolves_by": "2026-10-09", "resolver": "needs both Sold status and price evidence against the asking threshold at 2026-10-09; unknown realized price is Unresolvable", "status": "open" }, { "id": "P6", "prediction": "The AI-keyword share on the four exit boards will be above 50% on 2026-10-07.", "resolves_by": "2026-10-07", "resolver": "Series C", "status": "Unresolvable. Series C retired 2026-09-10 before any second capture (exit-board terms bar crawling); scored Issue 2." }, { "id": "P7", "prediction": "The October 2026 Hacker News freelancer thread will again have zero \"SEEKING FREELANCER\" top-level posts.", "resolves_by": "2026-10-08", "resolver": "Series E", "status": "open" }, { "id": "P8", "prediction": "The share of the newest 100 GitHub bounty-label issues sitting in board-named repositories will exceed 60% on 2026-10-07.", "resolves_by": "2026-10-07", "resolver": "Series D", "status": "open" }, { "id": "P9", "prediction": "Of the three Freelancer purchasing-signal listings, the GB Studio job will be awarded to an account with more than 10 reviews, or not awarded at all, by 2026-09-30.", "resolves_by": "2026-09-30", "resolver": "https://www.freelancer.com/projects/game-development/gbstudio-overlay-image-integration", "status": "open" }, { "id": "P10", "prediction": "The anchor-variation experiment will show that adding a licensed-attorney credential to the prompt raises category A above 60% of lines at Opus 5 (baseline 17%).", "resolves_by": "2026-09-09 (same day)", "resolver": "index/variants", "status": "Wrong. Resolved 53% (8/15), or 60% under the alternative label for one line; neither is above 60%. Direction right, magnitude overstated." }, { "id": "P11", "prediction": "In the week-3 Swarm Forecast sample, \"sell to the swarm\" (category B) is the unique largest category of the 90 lines; a tie for largest counts as Wrong.", "resolves_by": "2026-09-21", "resolver": "Series A, taxonomy v1", "status": "open" }, { "id": "P12", "prediction": "In the week-3 sample, Haiku 4.5 puts fewer than 15 of its 30 lines in \"sell the signature\" (category A).", "resolves_by": "2026-09-21", "resolver": "Series A", "status": "open" }, { "id": "P13", "prediction": "GitHub's reported total of open bounty-label issues exceeds 4,500 in the first collector capture whose generated_at_utc falls on 2026-10-07 (scheduled 09:00 UTC; a delayed run that day still counts; 4,323 on 2026-09-09, 4,388 on 2026-09-14). No capture that day: Unresolvable.", "resolves_by": "2026-10-07", "resolver": "Series D, total_count_reported", "status": "open" } ] }