--- schema_version: "1.0.0" edition_id: "mp-2026-10-28-morning-0110" published_at: "2026-10-28T09:00:00.000-04:00" modified_at: "2026-10-28T09:00:00.000-04:00" canonical_url: "https://themachinepress.com/edition/2026-10-28" story_count: 27 lead_story_id: "mp-2026-10-28-001" --- # The Machine Press — Morning edition Edition ID: `mp-2026-10-28-morning-0110` Published: 2026-10-28T09:00:00.000-04:00 Canonical edition: https://themachinepress.com/edition/2026-10-28 In simulation, routing tool calls by consequence, reversibility, ambiguity and evidence freshness cut harmful-side-effect trajectories from 14.75 to 7.02 percent at equal compute cost. ## 1. The Gate Counts the Cost of Being Wrong {#mp-2026-10-28-001} - Story ID: `mp-2026-10-28-001` - Type: `lead` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-001/the-gate-counts-the-cost-of-being-wrong **Dek:** In simulation, routing tool calls by consequence, reversibility, ambiguity and evidence freshness cut harmful-side-effect trajectories from 14.75 to 7.02 percent at equal compute cost. Runtime safeguards for tool-using agents often ask only how uncertain a proposed action looks. CASEL instead assigns each call to direct execution, verification, sandboxing, clarification, approval or abstention by combining calibrated failure probability with state-dependent consequence, reversibility, ambiguity and evidence freshness. Across 20 simulated seeds with 120,000 test actions each, its allocation policy reduced harmful-side-effect trajectories from 14.75 percent under an uncertainty-ranked gate to 7.02 percent and lowered normalized harm from 0.184 to 0.123 at the same compute cost. A simpler probability-times-consequence ranker reached similar harm but sent 45 percent more actions to human approval. These results compare policies inside a controlled simulator whose intervention effectiveness is assumed, not measured; they do not establish deployment performance. ### Why it matters {#why-it-matters-mp-2026-10-28-001} In simulation, routing tool calls by consequence, reversibility, ambiguity and evidence freshness cut harmful-side-effect trajectories from 14.75 to 7.02 percent at equal compute cost. ### Limits and context {#limitations-mp-2026-10-28-001} - Runtime safeguards for tool-using agents often ask only how uncertain a proposed action looks. - These results compare policies inside a controlled simulator whose intervention effectiveness is assumed, not measured; they do not establish deployment performance. ### Claims and sources {#claims-mp-2026-10-28-001} - In simulation, routing tool calls by consequence, reversibility, ambiguity and evidence freshness cut harmful-side-effect trajectories from 14.75 to 7.02 percent at equal compute cost. [source-2026-10-28-001] — Qualification: Runtime safeguards for tool-using agents often ask only how uncertain a proposed action looks. ## 2. A Radio Circle Spans Two Million Light-Years {#mp-2026-10-28-002} - Story ID: `mp-2026-10-28-002` - Type: `secondary` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-002/a-radio-circle-spans-two-million-light-years **Dek:** ASKAP data reveal a faint 0.63-megaparsec ring around the brightest galaxy in Abell 3639, placing ORC J1928-5056 among the largest single odd radio circles known. Odd radio circles are faint, roughly circular radio structures whose origins remain unsettled. Researchers report ORC J1928-5056 in 856-megahertz ASKAP continuum data as a mildly elliptical ring four arcminutes across. At the central galaxy's measured redshift of 0.1507, that angle corresponds to 0.63 megaparsecs, or roughly two million light-years. After subtracting the host's radio nucleus, the ring carries 73 millijanskys across about twelve square arcminutes and a rest-frame 1.4-gigahertz luminosity of 2 × 10^24 watts per hertz. The central object is a massive passive elliptical galaxy and brightest cluster galaxy with a radio jet. The observation establishes the size and radio properties of this candidate; it does not yet settle which proposed mechanism formed the circle. ### Why it matters {#why-it-matters-mp-2026-10-28-002} ASKAP data reveal a faint 0.63-megaparsec ring around the brightest galaxy in Abell 3639, placing ORC J1928-5056 among the largest single odd radio circles known. ### Limits and context {#limitations-mp-2026-10-28-002} - Odd radio circles are faint, roughly circular radio structures whose origins remain unsettled. - The observation establishes the size and radio properties of this candidate; it does not yet settle which proposed mechanism formed the circle. ### Claims and sources {#claims-mp-2026-10-28-002} - ASKAP data reveal a faint 0.63-megaparsec ring around the brightest galaxy in Abell 3639, placing ORC J1928-5056 among the largest single odd radio circles known. [source-2026-10-28-002] — Qualification: Odd radio circles are faint, roughly circular radio structures whose origins remain unsettled. ## 3. The Numerical Fix Needed One More Property Test {#mp-2026-10-28-003} - Story ID: `mp-2026-10-28-003` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-003/the-numerical-fix-needed-one-more-property-test **Dek:** Hypothesis tests reproduced two known softmax failures, found another edge case in the proposed repair and guided an implementation that also supports causal masking. Softmax and cross-entropy sit inside nearly every modern neural network, so rare numerical failures can travel widely. This paper encodes expected invariants as property-based tests and uses Hypothesis to reproduce two previously reported problems in implementations found across major machine-learning libraries. The same test suite exposes an additional edge case in the earlier repair. The authors then propose a more robust softmax implementation that passes their generated cases and supports causal-attention masking. The contribution is not a census of every installed library version; it is a compact demonstration that executable mathematical properties can uncover failures missed by example-based tests. ### Why it matters {#why-it-matters-mp-2026-10-28-003} Hypothesis tests reproduced two known softmax failures, found another edge case in the proposed repair and guided an implementation that also supports causal masking. ### Limits and context {#limitations-mp-2026-10-28-003} - The contribution is not a census of every installed library version; it is a compact demonstration that executable mathematical properties can uncover failures missed by example-based tests. ### Claims and sources {#claims-mp-2026-10-28-003} - Hypothesis tests reproduced two known softmax failures, found another edge case in the proposed repair and guided an implementation that also supports causal masking. [source-2026-10-28-003] — Qualification: The contribution is not a census of every installed library version; it is a compact demonstration that executable mathematical properties can uncover failures missed by example-based tests. ## 4. Seven Hundred Twenty-Five Agent Errors Recovered Without Intervention {#mp-2026-10-28-004} - Story ID: `mp-2026-10-28-004` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-004/seven-hundred-twenty-five-agent-errors-recovered-without-intervention **Dek:** ReInBench records 1,403 error events across 1,259 computer-use and web traces; 725 recovered, often within a few further steps. An external monitor that interrupts every agent error can erase the opportunity for the agent to repair its own execution. ReInBench collects 1,259 traces and 1,403 observed error events from computer-use and web tasks. Of those events, 725 self-recovered, often after only a few additional steps, with recovery rates varying across execution models. Recovery from one error also correlated with handling later errors of the same type and with eventual task completion. The authors propose ReCoF to predict at error time whether recovery is likely, reporting a 9.02 percent relative Macro-F1 improvement over the strongest baseline. The result supports selective intervention on this benchmark, not unrestricted waiting in high-consequence systems. ### Why it matters {#why-it-matters-mp-2026-10-28-004} ReInBench records 1,403 error events across 1,259 computer-use and web traces; 725 recovered, often within a few further steps. ### Limits and context {#limitations-mp-2026-10-28-004} - Of those events, 725 self-recovered, often after only a few additional steps, with recovery rates varying across execution models. - The result supports selective intervention on this benchmark, not unrestricted waiting in high-consequence systems. ### Claims and sources {#claims-mp-2026-10-28-004} - ReInBench records 1,403 error events across 1,259 computer-use and web traces; 725 recovered, often within a few further steps. [source-2026-10-28-004] — Qualification: Of those events, 725 self-recovered, often after only a few additional steps, with recovery rates varying across execution models. ## 5. A Handoff Has to Preserve Commitments, Not Just Facts {#mp-2026-10-28-005} - Story ID: `mp-2026-10-28-005` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-005/a-handoff-has-to-preserve-commitments-not-just-facts **Dek:** Across 1,039 tasks, structured inherited state improved repository-evidence recovery for every tested receiver and continuation for most. Long-running work often reaches a new model after earlier decisions, tool calls and file changes have already constrained what remains. HSC-Bench measures two distinct parts of that takeover across 1,039 tasks: recovering relevant repository evidence and continuing from established conversational, tool and file state. Six receiving models performed differently across the two tracks, showing that memory retrieval alone does not predict whether a model will honor completed work and prior commitments. Structured handoff state improved evidence recovery for every receiver and continuation performance for most. The benchmark does not prove a universal handoff format, but it makes continuity observable as downstream behavior rather than a summary-quality score. ### Why it matters {#why-it-matters-mp-2026-10-28-005} Across 1,039 tasks, structured inherited state improved repository-evidence recovery for every tested receiver and continuation for most. ### Limits and context {#limitations-mp-2026-10-28-005} - Six receiving models performed differently across the two tracks, showing that memory retrieval alone does not predict whether a model will honor completed work and prior commitments. - The benchmark does not prove a universal handoff format, but it makes continuity observable as downstream behavior rather than a summary-quality score. ### Claims and sources {#claims-mp-2026-10-28-005} - Across 1,039 tasks, structured inherited state improved repository-evidence recovery for every tested receiver and continuation for most. [source-2026-10-28-005] — Qualification: Six receiving models performed differently across the two tracks, showing that memory retrieval alone does not predict whether a model will honor completed work and prior commitments. ## 6. Passing Tests Covered Only a Third of the Required Behavior {#mp-2026-10-28-006} - Story ID: `mp-2026-10-28-006` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-006/passing-tests-covered-only-a-third-of-the-required-behavior **Dek:** On Req2Test-Bench, the strongest agent configuration produced runnable tests for 82.8 percent of cases but behaviorally adequate tests for 33.8 percent. A generated test can execute cleanly while failing to constrain the behavior that a requirement actually describes. Req2Test-Bench reconstructs 5,086 developer-authored behavior-driven-development instances from eight open-source projects, five languages and four frameworks. Agents receive a Gherkin requirement and target interfaces, then must generate the withheld test implementation. The best of eight evaluated configurations produced passing tests for 82.8 percent of instances, but only 33.8 percent killed every mutant killed by the developer test. Under-specified setup and implicit expected outcomes accounted for much of the information gap. The benchmark measures fidelity to recovered project artifacts; it does not establish that every developer-written test is complete. ### Why it matters {#why-it-matters-mp-2026-10-28-006} On Req2Test-Bench, the strongest agent configuration produced runnable tests for 82.8 percent of cases but behaviorally adequate tests for 33.8 percent. ### Limits and context {#limitations-mp-2026-10-28-006} - The best of eight evaluated configurations produced passing tests for 82.8 percent of instances, but only 33.8 percent killed every mutant killed by the developer test. - The benchmark measures fidelity to recovered project artifacts; it does not establish that every developer-written test is complete. ### Claims and sources {#claims-mp-2026-10-28-006} - On Req2Test-Bench, the strongest agent configuration produced runnable tests for 82.8 percent of cases but behaviorally adequate tests for 33.8 percent. [source-2026-10-28-006] — Qualification: The best of eight evaluated configurations produced passing tests for 82.8 percent of instances, but only 33.8 percent killed every mutant killed by the developer test. ## 7. Official Tests Accepted a Third of the Wrong Infrastructure Patches {#mp-2026-10-28-007} - Story ID: `mp-2026-10-28-007` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-007/official-tests-accepted-a-third-of-the-wrong-infrastructure-patches **Dek:** InfraDev adds hidden probes to 36 multi-file AI-infrastructure tasks; the best frontier model resolves 38.9 percent, and probes reject 32 percent of official-test passers. Production AI infrastructure changes span distributed training, inference and reinforcement-learning systems, but most coding benchmarks stop at a kernel or one GPU. InfraDev reconstructs 36 pull requests from Megatron-LM, vLLM and verl as repository tasks that must run across many GPUs. Hidden probes cover requirements omitted by the original tests while an explicit prompt names every interface those probes invoke. Across six frontier models, the strongest resolved 38.9 percent. The probes rejected 32 percent of submissions that passed all official tests; using original pull-request descriptions as prompts produced no solved tasks. The benchmark is small and infrastructure-specific, yet it demonstrates why state, memory, time and traffic behavior need evaluation beyond nominal test suites. ### Why it matters {#why-it-matters-mp-2026-10-28-007} InfraDev adds hidden probes to 36 multi-file AI-infrastructure tasks; the best frontier model resolves 38.9 percent, and probes reject 32 percent of official-test passers. ### Limits and context {#limitations-mp-2026-10-28-007} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-10-28-007} - InfraDev adds hidden probes to 36 multi-file AI-infrastructure tasks; the best frontier model resolves 38.9 percent, and probes reject 32 percent of official-test passers. [source-2026-10-28-007] ## 8. The Humanoid Lowers Itself From a Walk Into a Crawl {#mp-2026-10-28-008} - Story ID: `mp-2026-10-28-008` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-008/the-humanoid-lowers-itself-from-a-walk-into-a-crawl **Dek:** REACH links variable-height bipedal motion, contact-rich crawling and LiDAR planning so a Unitree G1 can traverse tested confined spaces. Crouched walking alone cannot fit a humanoid through every low-clearance route. REACH divides the problem into variable-height bipedal locomotion from 1.27 to 0.50 meters, crawling from 0.50 to 0.20 meters and a bidirectional transition policy between the two. A history-conditioned latent state captures proprioceptive terrain and contact context during crawling, while a LiDAR planner converts point clouds into three-dimensional paths and proactive height commands. The authors report real-world height and velocity tracking, mode transitions and confined-space traversal on a Unitree G1. These demonstrations concern the tested robot and environments; they do not establish unsupervised operation in arbitrary confined or hazardous spaces. ### Why it matters {#why-it-matters-mp-2026-10-28-008} REACH links variable-height bipedal motion, contact-rich crawling and LiDAR planning so a Unitree G1 can traverse tested confined spaces. ### Limits and context {#limitations-mp-2026-10-28-008} - Crouched walking alone cannot fit a humanoid through every low-clearance route. - These demonstrations concern the tested robot and environments; they do not establish unsupervised operation in arbitrary confined or hazardous spaces. ### Claims and sources {#claims-mp-2026-10-28-008} - REACH links variable-height bipedal motion, contact-rich crawling and LiDAR planning so a Unitree G1 can traverse tested confined spaces. [source-2026-10-28-008] — Qualification: Crouched walking alone cannot fit a humanoid through every low-clearance route. ## 9. The Malware Detector Can Fail Without Moving Its Metrics {#mp-2026-10-28-009} - Story ID: `mp-2026-10-28-009` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-009/the-malware-detector-can-fail-without-moving-its-metrics **Dek:** Across 453,174 injected faults, 72.3 percent of evaluated tree-model faults barely changed aggregate recall or false-positive rate, even when individual predictions moved. Hardware malware detectors classify performance-counter patterns, but faults inside their own decision logic can quietly change which executions pass. Researchers injected 453,174 stuck-at and selected parameter-bit faults across 18 classifiers, counter budgets and ensemble designs. Tree models showed a lower mean critical-fault fraction than function-based models, 5.3 versus 14.3 percent. Yet 72.3 percent of evaluated tree faults were metric-silent by the study's aggregate threshold, despite possible changes to individual predictions. In a representative J48 design, an idealized selection method identified 16 of 115 structural units whose protection reduced predicted worst-case recall loss from 60.1 to 3.3 points. RTL correspondence and physical hardening cost remain unvalidated. ### Why it matters {#why-it-matters-mp-2026-10-28-009} Across 453,174 injected faults, 72.3 percent of evaluated tree-model faults barely changed aggregate recall or false-positive rate, even when individual predictions moved. ### Limits and context {#limitations-mp-2026-10-28-009} - RTL correspondence and physical hardening cost remain unvalidated. ### Claims and sources {#claims-mp-2026-10-28-009} - Across 453,174 injected faults, 72.3 percent of evaluated tree-model faults barely changed aggregate recall or false-positive rate, even when individual predictions moved. [source-2026-10-28-009] — Qualification: RTL correspondence and physical hardening cost remain unvalidated. ## 10. A Team Memory Learns When It Should Stay Quiet {#mp-2026-10-28-010} - Story ID: `mp-2026-10-28-010` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-010/a-team-memory-learns-when-it-should-stay-quiet **Dek:** DiscernMem tests repairs against current and earlier tasks, then learns the contexts in which each role should apply or avoid a stored rule. A successful multi-agent episode does not reveal whether a stored memory was sound, irrelevant or accidentally helpful. DiscernMem compares executions with one memory against runs without it, diagnoses defects and tests proposed repairs on both current and earlier tasks to avoid destroying previously useful behavior. A second component records each memory's past use and effect, learning a contextual boundary for when a role should apply or avoid it. With Qwen3-4B-Instruct-2507 across four benchmarks, the training-free framework led reported baselines on AutoGen, CAMEL and MacNet by 0.51 to 4.32 average points and improved symbolic-planning results. The evidence belongs to the tested model, systems and tasks rather than every multi-agent memory design. ### Why it matters {#why-it-matters-mp-2026-10-28-010} DiscernMem tests repairs against current and earlier tasks, then learns the contexts in which each role should apply or avoid a stored rule. ### Limits and context {#limitations-mp-2026-10-28-010} - A successful multi-agent episode does not reveal whether a stored memory was sound, irrelevant or accidentally helpful. ### Claims and sources {#claims-mp-2026-10-28-010} - DiscernMem tests repairs against current and earlier tasks, then learns the contexts in which each role should apply or avoid a stored rule. [source-2026-10-28-010] — Qualification: A successful multi-agent episode does not reveal whether a stored memory was sound, irrelevant or accidentally helpful. ## 11. Fewer Image Steps Made Unsafe Concepts Harder to Remove {#mp-2026-10-28-011} - Story ID: `mp-2026-10-28-011` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-011/fewer-image-steps-made-unsafe-concepts-harder-to-remove **Dek:** Across tested distilled text-to-image systems, lower step counts raised unsafe-generation rates; DICE reduced FLUX Schnell's measured rate to 0.046. Fast text-to-image models compress generation into only a few denoising steps, but the safety behavior of that compression has received less scrutiny. This study reports that several distilled few-step models generate unsafe concepts more often than their teacher models and become harder to modify with existing concept-erasure methods as step counts fall. The authors attribute part of the failure to fragile generation trajectories and the way guidance is distilled, then introduce DICE, which aligns erasure with the originating distillation method. On their FLUX Schnell evaluation, DICE reduced the unsafe-generation rate to 0.046, compared with at least 0.161 for tested prior erasure methods, while largely preserving benign quality. Those rates depend on the study's prompts, detectors and models. ### Why it matters {#why-it-matters-mp-2026-10-28-011} Across tested distilled text-to-image systems, lower step counts raised unsafe-generation rates; DICE reduced FLUX Schnell's measured rate to 0.046. ### Limits and context {#limitations-mp-2026-10-28-011} - Fast text-to-image models compress generation into only a few denoising steps, but the safety behavior of that compression has received less scrutiny. ### Claims and sources {#claims-mp-2026-10-28-011} - Across tested distilled text-to-image systems, lower step counts raised unsafe-generation rates; DICE reduced FLUX Schnell's measured rate to 0.046. [source-2026-10-28-011] — Qualification: Fast text-to-image models compress generation into only a few denoising steps, but the safety behavior of that compression has received less scrutiny. ## 12. A Biological Curve Becomes a Soft Robot in Four Milliseconds {#mp-2026-10-28-012} - Story ID: `mp-2026-10-28-012` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-012/a-biological-curve-becomes-a-soft-robot-in-four-milliseconds **Dek:** A surrogate-assisted inverse-design pipeline reports 93 percent shape reconstruction on 5,000 unseen threading patterns and physical demonstrations of four bioinspired forms. Cable-driven modular soft robots can mimic animal curves, but mapping a target shape back to a buildable threading pattern normally requires expensive simulation and combinatorial search. This real-to-sim-to-real pipeline learns a forward surrogate and uses uncertainty-guided candidate refinement for inverse design. The reported surrogate cuts deformation-prediction time from 153,000 milliseconds to 0.0008 milliseconds, while inverse design falls from about 1.99 million milliseconds to 4.18 milliseconds. On 5,000 previously unseen threading patterns, 93 percent met the study's surrogate-based reconstruction criterion, and four physical robots demonstrated bioinspired shapes. The success rate is evaluated by the learned surrogate, so the physical examples are important but limited checks on transfer. ### Why it matters {#why-it-matters-mp-2026-10-28-012} A surrogate-assisted inverse-design pipeline reports 93 percent shape reconstruction on 5,000 unseen threading patterns and physical demonstrations of four bioinspired forms. ### Limits and context {#limitations-mp-2026-10-28-012} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-10-28-012} - A surrogate-assisted inverse-design pipeline reports 93 percent shape reconstruction on 5,000 unseen threading patterns and physical demonstrations of four bioinspired forms. [source-2026-10-28-012] ## 13. The Autonomous Lab Gets a Causal Event Ledger {#mp-2026-10-28-013} - Story ID: `mp-2026-10-28-013` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-013/the-autonomous-lab-gets-a-causal-event-ledger **Dek:** A proposed four-part framework joins robot trajectories, typed instrument calls, runtime safety checks and episode-level evaluation into one traceable execution interface. Self-driving laboratories must coordinate continuous robot motion with discrete instrument APIs and measurements that arrive over time. The proposed framework separates that challenge into a measurement-conditioned Lab Action Model, an OpenSDL execution substrate, a Safe-SDL runtime gate and L3-Bench evaluation. OpenSDL maps role-level operations to devices and records timestamped causal events; Safe-SDL applies schema and resource checks before dispatch and signal-temporal-logic monitoring during execution. L3-Bench scores physical outcomes, timing, reproducibility, safety violations and fault recovery. Much of the system is a design and initial implementation rather than a completed autonomous laboratory comparison, but it defines the audit trail needed to test later systems without scraping their interfaces after the fact. ### Why it matters {#why-it-matters-mp-2026-10-28-013} A proposed four-part framework joins robot trajectories, typed instrument calls, runtime safety checks and episode-level evaluation into one traceable execution interface. ### Limits and context {#limitations-mp-2026-10-28-013} - Much of the system is a design and initial implementation rather than a completed autonomous laboratory comparison, but it defines the audit trail needed to test later systems without scraping their interfaces after the fact. ### Claims and sources {#claims-mp-2026-10-28-013} - A proposed four-part framework joins robot trajectories, typed instrument calls, runtime safety checks and episode-level evaluation into one traceable execution interface. [source-2026-10-28-013] — Qualification: Much of the system is a design and initial implementation rather than a completed autonomous laboratory comparison, but it defines the audit trail needed to test later systems without scraping their interfaces after the fact. ## 14. The Swap Fee Display Missed Millions in Hook Deductions {#mp-2026-10-28-014} - Story ID: `mp-2026-10-28-014` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-014/the-swap-fee-display-missed-millions-in-hook-deductions **Dek:** A bytecode audit of 7,873 hook deployments finds that 24.30 percent could impose undisclosed deductions and records $3.97 million where token prices were available. Modern automated market makers let third-party hook contracts alter settlement, but standard fee displays do not necessarily itemize those deductions. Researchers analyzed 7,873 hook deployments, 50,011 pools and 296 million events on Uniswap v4 and PancakeSwap Infinity through August 1, 2026. Their bytecode method reconstructs changes to trader payments and receipts separately from exchanged assets and pool fees. Direct settlement adjustment appeared in 71.37 percent of registered hooks; 24.30 percent could impose undisclosed deductions, and such deductions occurred in 59.23 percent of accounted swaps involving those hooks. Priced observations totaled $3.97 million. The study measures on-chain behavior under its reconstruction method, not legal intent or every user-interface disclosure. ### Why it matters {#why-it-matters-mp-2026-10-28-014} A bytecode audit of 7,873 hook deployments finds that 24.30 percent could impose undisclosed deductions and records $3.97 million where token prices were available. ### Limits and context {#limitations-mp-2026-10-28-014} - Modern automated market makers let third-party hook contracts alter settlement, but standard fee displays do not necessarily itemize those deductions. - The study measures on-chain behavior under its reconstruction method, not legal intent or every user-interface disclosure. ### Claims and sources {#claims-mp-2026-10-28-014} - A bytecode audit of 7,873 hook deployments finds that 24.30 percent could impose undisclosed deductions and records $3.97 million where token prices were available. [source-2026-10-28-014] — Qualification: Modern automated market makers let third-party hook contracts alter settlement, but standard fee displays do not necessarily itemize those deductions. ## 15. A Cracked Laser Crystal Repairs Its Optical Path {#mp-2026-10-28-026} - Story ID: `mp-2026-10-28-026` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-026/a-cracked-laser-crystal-repairs-its-optical-path **Dek:** HDMAC microcrystals healed AFM-induced cracks along two crystal planes while retaining waveguide and resonator behavior needed for lasing studies. Microscopic fractures can end the useful life of organic photonic devices, even when the active material remains chemically intact. This mechanophotonic study uses an atomic-force-microscope cantilever to introduce controlled cracks and complete fractures in HDMAC molecular crystals. The crystals showed self-healing along the (010) and (001) planes, and the researchers evaluated recovery through optical loss, resonator behavior and nanomechanical measurements rather than visual closure alone. The healed samples retained waveguiding and resonator properties relevant to lasing. The result concerns one molecular crystal system under laboratory damage and repair; it does not establish self-healing for deployed photonic devices or arbitrary fracture geometries. ### Why it matters {#why-it-matters-mp-2026-10-28-026} HDMAC microcrystals healed AFM-induced cracks along two crystal planes while retaining waveguide and resonator behavior needed for lasing studies. ### Limits and context {#limitations-mp-2026-10-28-026} - The result concerns one molecular crystal system under laboratory damage and repair; it does not establish self-healing for deployed photonic devices or arbitrary fracture geometries. ### Claims and sources {#claims-mp-2026-10-28-026} - HDMAC microcrystals healed AFM-induced cracks along two crystal planes while retaining waveguide and resonator behavior needed for lasing studies. [source-2026-10-28-015] — Qualification: The result concerns one molecular crystal system under laboratory damage and repair; it does not establish self-healing for deployed photonic devices or arbitrary fracture geometries. ## 16. A Ninety-Percent-Accurate Monitor Can Still Make Restarts Worse {#mp-2026-10-28-027} - Story ID: `mp-2026-10-28-027` - Type: `dispatch` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-027/a-ninety-percent-accurate-monitor-can-still-make-restarts-worse **Dek:** The same monitor ranges from a five-point loss to a 35-point gain depending on the replacement attempt, and none of 103 live comparisons cleared the study's decision bar. Accuracy alone cannot say whether an alarm should restart an agent. An exact identity in this paper shows that a monitor fixed at 90 percent accuracy can move total success by anywhere from minus five to plus 35 percentage points depending only on the residual-budget success probability of the replacement attempt. The authors add a decision-sufficiency test for progress representations and a finite-sample audit that returns restart, continue or abstain. On 194 externally evaluated executions across 19 selected SymPy issues, an apparent 2.29-point gain traced to one surviving failed attempt, a second model lost 42.29 points, and none of 103 live comparisons met the stated certification threshold. The analysis is scoped to independent clean restarts with the declared assumptions. ### Why it matters {#why-it-matters-mp-2026-10-28-027} The same monitor ranges from a five-point loss to a 35-point gain depending on the replacement attempt, and none of 103 live comparisons cleared the study's decision bar. ### Limits and context {#limitations-mp-2026-10-28-027} - Accuracy alone cannot say whether an alarm should restart an agent. - An exact identity in this paper shows that a monitor fixed at 90 percent accuracy can move total success by anywhere from minus five to plus 35 percentage points depending only on the residual-budget success probability of the replacement attempt. ### Claims and sources {#claims-mp-2026-10-28-027} - The same monitor ranges from a five-point loss to a 35-point gain depending on the replacement attempt, and none of 103 live comparisons cleared the study's decision bar. [source-2026-10-28-016] — Qualification: Accuracy alone cannot say whether an alarm should restart an agent. ## 17. Two Seconds of Delay Smooths the Training Load {#mp-2026-10-28-015} - Story ID: `mp-2026-10-28-015` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-015/two-seconds-of-delay-smooths-the-training-load **Dek:** Phase-coordinated launch timing cut simulated 0.2-to-2-hertz mean-square power fluctuation by 74.39 percent with 2.188 seconds of average delay. A model-predictive controller offsets independent training-job starts so their periodic power swings cancel instead of stacking. In numerical results, it reduced mean-square fluctuation in the 0.2-to-2-hertz band by 74.39 percent and power standard deviation by 46.6 percent, while meeting launch deadlines. The evidence is simulated and does not measure a production grid. ### Why it matters {#why-it-matters-mp-2026-10-28-015} Phase-coordinated launch timing cut simulated 0.2-to-2-hertz mean-square power fluctuation by 74.39 percent with 2.188 seconds of average delay. ### Limits and context {#limitations-mp-2026-10-28-015} - The evidence is simulated and does not measure a production grid. ### Claims and sources {#claims-mp-2026-10-28-015} - Phase-coordinated launch timing cut simulated 0.2-to-2-hertz mean-square power fluctuation by 74.39 percent with 2.188 seconds of average delay. [source-2026-10-28-017] — Qualification: The evidence is simulated and does not measure a production grid. ## 18. Silicon Reads Mid-Infrared Light Through an Optical Gate {#mp-2026-10-28-016} - Story ID: `mp-2026-10-28-016` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-016/silicon-reads-mid-infrared-light-through-an-optical-gate **Dek:** A nanoparticle-metasurface device reports room-temperature, zero-bias detection at 6.8 micrometers without extending silicon's direct absorption band. Instead of asking a mid-infrared photon to create carriers directly in silicon, the device lets it alter a prepared neodymium optical state and converts that change through a silicon Schottky junction. The reported modulation responsivity is 103 microamps per watt at 6.8 micrometers, with three distinguishable combined-input states. This is a laboratory device demonstration, not a production sensor. ### Why it matters {#why-it-matters-mp-2026-10-28-016} A nanoparticle-metasurface device reports room-temperature, zero-bias detection at 6.8 micrometers without extending silicon's direct absorption band. ### Limits and context {#limitations-mp-2026-10-28-016} - This is a laboratory device demonstration, not a production sensor. ### Claims and sources {#claims-mp-2026-10-28-016} - A nanoparticle-metasurface device reports room-temperature, zero-bias detection at 6.8 micrometers without extending silicon's direct absorption band. [source-2026-10-28-018] — Qualification: This is a laboratory device demonstration, not a production sensor. ## 19. Four Measurement Settings Are Both Necessary and Sufficient {#mp-2026-10-28-017} - Story ID: `mp-2026-10-28-017` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-017/four-measurement-settings-are-both-necessary-and-sufficient **Dek:** For collective first and second moments of two qubits, structural bounds and explicit constructions certify a four-setting minimum. The researchers ask how many orthogonal projective settings can estimate several observables without bias for arbitrary states. For two-qubit first and second moments, four settings are proven necessary and sufficient; an optimized design comes within about 0.0341 percent of a fixed-allocation variance lower bound. Finite-depth reconstruction adds bias, so the advantage depends on sample regime. ### Why it matters {#why-it-matters-mp-2026-10-28-017} For collective first and second moments of two qubits, structural bounds and explicit constructions certify a four-setting minimum. ### Limits and context {#limitations-mp-2026-10-28-017} - No additional limitation was separately recorded. ### Claims and sources {#claims-mp-2026-10-28-017} - For collective first and second moments of two qubits, structural bounds and explicit constructions certify a four-setting minimum. [source-2026-10-28-019] ## 20. A Full Calorimeter Image Reaches the Trigger in 25.9 Microseconds {#mp-2026-10-28-018} - Story ID: `mp-2026-10-28-018` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-018/a-full-calorimeter-image-reaches-the-trigger-in-25-9-microseconds **Dek:** An 8-bit FPGA CNN rejected 99.34 percent of a recorded background sample at nominal 95 percent signal efficiency within the adopted latency budget. A lossless space-to-depth transform and quantization-aware training shrink full-granularity calorimeter inference for FPGA execution. On 100,000 test events, the board measured 17.9 microseconds for CNN inference and a projected 25.9 microseconds end to end, below the study's 32-microsecond budget. The result is a KOTO-oriented proof of principle, not a deployed trigger claim. ### Why it matters {#why-it-matters-mp-2026-10-28-018} An 8-bit FPGA CNN rejected 99.34 percent of a recorded background sample at nominal 95 percent signal efficiency within the adopted latency budget. ### Limits and context {#limitations-mp-2026-10-28-018} - The result is a KOTO-oriented proof of principle, not a deployed trigger claim. ### Claims and sources {#claims-mp-2026-10-28-018} - An 8-bit FPGA CNN rejected 99.34 percent of a recorded background sample at nominal 95 percent signal efficiency within the adopted latency budget. [source-2026-10-28-020] — Qualification: The result is a KOTO-oriented proof of principle, not a deployed trigger claim. ## 21. Two Blazars Add a Second Radio Rhythm {#mp-2026-10-28-019} - Story ID: `mp-2026-10-28-019` - Type: `ticker` - Classification: `editorial` - Content status: `new` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-019/two-blazars-add-a-second-radio-rhythm **Dek:** Long monitoring records show candidate 2:1 periodicities in PKS J0805-0111 and PKS 2131-021, adding potential corroboration for binary interpretations. The reported harmonic structure follows an earlier 2:1 candidate in J1309+1154 and could help distinguish physical periodicity from correlated noise in long blazar light curves. The authors present several possible models and arguments for and against the signals. These remain supermassive-black-hole-binary candidates, not confirmed binaries. ### Why it matters {#why-it-matters-mp-2026-10-28-019} Long monitoring records show candidate 2:1 periodicities in PKS J0805-0111 and PKS 2131-021, adding potential corroboration for binary interpretations. ### Limits and context {#limitations-mp-2026-10-28-019} - These remain supermassive-black-hole-binary candidates, not confirmed binaries. ### Claims and sources {#claims-mp-2026-10-28-019} - Long monitoring records show candidate 2:1 periodicities in PKS J0805-0111 and PKS 2131-021, adding potential corroboration for binary interpretations. [source-2026-10-28-021] — Qualification: These remain supermassive-black-hole-binary candidates, not confirmed binaries. ## 22. Space Drums 2.0 {#mp-2026-10-28-020} - Story ID: `mp-2026-10-28-020` - Type: `invention_desk` - Classification: `editorial` - Content status: `carried_over` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-020/space-drums-2-0 **Dek:** Moves motion tracking onto two ESP32-S3 drumsticks, adds haptic hits and a two-pedal hub, and sends velocity-sensitive drum events to open PC or Android playback software. Moves motion tracking onto two ESP32-S3 drumsticks, adds haptic hits and a two-pedal hub, and sends velocity-sensitive drum events to open PC or Android playback software. ### Why it matters {#why-it-matters-mp-2026-10-28-020} An independent builder is turning an improbable idea into a working project. ### Limits and context {#limitations-mp-2026-10-28-020} - A Desk Pick is an editorial selection, not a product endorsement. ### Claims and sources {#claims-mp-2026-10-28-020} - This Invention Desk entry makes no independently sourced news claim. ## 23. 3RIC {#mp-2026-10-28-021} - Story ID: `mp-2026-10-28-021` - Type: `invention_desk` - Classification: `editorial` - Content status: `carried_over` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-021/3ric **Dek:** Pairs a chip-level 65C02 computer with custom hardware, video, sound, floppy, and microSD support with a cycle-honest emulator and browser assembler built around the same machine. Pairs a chip-level 65C02 computer with custom hardware, video, sound, floppy, and microSD support with a cycle-honest emulator and browser assembler built around the same machine. ### Why it matters {#why-it-matters-mp-2026-10-28-021} An independent builder is turning an improbable idea into a working project. ### Limits and context {#limitations-mp-2026-10-28-021} - A Desk Pick is an editorial selection, not a product endorsement. ### Claims and sources {#claims-mp-2026-10-28-021} - This Invention Desk entry makes no independently sourced news claim. ## 24. M3-CRETE {#mp-2026-10-28-022} - Story ID: `mp-2026-10-28-022` - Type: `invention_desk` - Classification: `editorial` - Content status: `carried_over` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-022/m3-crete **Dek:** Publishes a pallet-scale cartesian motion reference, CAD, BOM, and low-voltage controls for supervised cementitious-extrusion research while explicitly excluding certified structural use. Publishes a pallet-scale cartesian motion reference, CAD, BOM, and low-voltage controls for supervised cementitious-extrusion research while explicitly excluding certified structural use. ### Why it matters {#why-it-matters-mp-2026-10-28-022} An independent builder is turning an improbable idea into a working project. ### Limits and context {#limitations-mp-2026-10-28-022} - A Desk Pick is an editorial selection, not a product endorsement. ### Claims and sources {#claims-mp-2026-10-28-022} - This Invention Desk entry makes no independently sourced news claim. ## 25. DeskNode {#mp-2026-10-28-023} - Story ID: `mp-2026-10-28-023` - Type: `invention_desk` - Classification: `editorial` - Content status: `carried_over` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-023/desknode **Dek:** Turns a bare ESP32-S3 touchscreen into a USB-connected desk instrument for PC activity and optional room temperature, humidity, and light measurements. Turns a bare ESP32-S3 touchscreen into a USB-connected desk instrument for PC activity and optional room temperature, humidity, and light measurements. ### Why it matters {#why-it-matters-mp-2026-10-28-023} An independent builder is turning an improbable idea into a working project. ### Limits and context {#limitations-mp-2026-10-28-023} - A Desk Pick is an editorial selection, not a product endorsement. ### Claims and sources {#claims-mp-2026-10-28-023} - This Invention Desk entry makes no independently sourced news claim. ## 26. The First Paid Slot {#mp-2026-10-28-024} - Story ID: `mp-2026-10-28-024` - Type: `invention_desk` - Classification: `house_example` - Content status: `carried_over` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-024/the-first-paid-slot **Dek:** A transparent preview of paid placement with one verified link and no claim of endorsement. A transparent preview of paid placement with one verified link and no claim of endorsement. House example - no advertiser paid. Payment will buy placement, never endorsement. ### Why it matters {#why-it-matters-mp-2026-10-28-024} This placement explains how builders can appear in The Invention Desk without purchasing editorial endorsement. ### Limits and context {#limitations-mp-2026-10-28-024} - House example - no advertiser paid. Payment will buy placement, never endorsement. ### Claims and sources {#claims-mp-2026-10-28-024} - This Invention Desk entry makes no independently sourced news claim. ## 27. Put Your Project on the Desk {#mp-2026-10-28-025} - Story ID: `mp-2026-10-28-025` - Type: `invention_desk` - Classification: `house_example` - Content status: `carried_over` - Permanent URL: https://themachinepress.com/story/mp-2026-10-28-025/put-your-project-on-the-desk **Dek:** One manually reviewed placement stays active for seven days and remains separate from Desk Picks. One manually reviewed placement stays active for seven days and remains separate from Desk Picks. Manual intake only. Payment buys placement, never endorsement, and every submission is reviewed. ### Why it matters {#why-it-matters-mp-2026-10-28-025} This placement explains how builders can appear in The Invention Desk without purchasing editorial endorsement. ### Limits and context {#limitations-mp-2026-10-28-025} - Manual intake only. Payment buys placement, never endorsement, and every submission is reviewed. ### Claims and sources {#claims-mp-2026-10-28-025} - This Invention Desk entry makes no independently sourced news claim. ## Normalized sources - **source-2026-10-28-001:** [arXiv preprint 2610.36570, original v1](https://arxiv.org/abs/2610.36570v1) — arXiv; primary_research - **source-2026-10-28-002:** [arXiv preprint 2610.36512, original v1](https://arxiv.org/abs/2610.36512v1) — arXiv; primary_research - **source-2026-10-28-003:** [arXiv preprint 2610.36427, original v1](https://arxiv.org/abs/2610.36427v1) — arXiv; primary_research - **source-2026-10-28-004:** [arXiv preprint 2610.36471, original v1](https://arxiv.org/abs/2610.36471v1) — arXiv; primary_research - **source-2026-10-28-005:** [arXiv preprint 2610.36523, original v1](https://arxiv.org/abs/2610.36523v1) — arXiv; primary_research - **source-2026-10-28-006:** [arXiv preprint 2610.36584, original v1](https://arxiv.org/abs/2610.36584v1) — arXiv; primary_research - **source-2026-10-28-007:** [arXiv preprint 2610.36589, original v1](https://arxiv.org/abs/2610.36589v1) — arXiv; primary_research - **source-2026-10-28-008:** [arXiv preprint 2610.36546, original v1](https://arxiv.org/abs/2610.36546v1) — arXiv; primary_research - **source-2026-10-28-009:** [arXiv preprint 2610.36612, original v1](https://arxiv.org/abs/2610.36612v1) — arXiv; primary_research - **source-2026-10-28-010:** [arXiv preprint 2610.36652, original v1](https://arxiv.org/abs/2610.36652v1) — arXiv; primary_research - **source-2026-10-28-011:** [arXiv preprint 2610.36665, original v1](https://arxiv.org/abs/2610.36665v1) — arXiv; primary_research - **source-2026-10-28-012:** [arXiv preprint 2610.36681, original v1](https://arxiv.org/abs/2610.36681v1) — arXiv; primary_research - **source-2026-10-28-013:** [arXiv preprint 2610.36761, original v1](https://arxiv.org/abs/2610.36761v1) — arXiv; primary_research - **source-2026-10-28-014:** [arXiv preprint 2610.36682, original v1](https://arxiv.org/abs/2610.36682v1) — arXiv; primary_research - **source-2026-10-28-015:** [arXiv preprint 2610.36692, original v1](https://arxiv.org/abs/2610.36692v1) — arXiv; primary_research - **source-2026-10-28-016:** [arXiv preprint 2610.36544, original v1](https://arxiv.org/abs/2610.36544v1) — arXiv; primary_research - **source-2026-10-28-017:** [arXiv preprint 2610.36441, original v1](https://arxiv.org/abs/2610.36441v1) — arXiv; primary_research - **source-2026-10-28-018:** [arXiv preprint 2610.36475, original v1](https://arxiv.org/abs/2610.36475v1) — arXiv; primary_research - **source-2026-10-28-019:** [arXiv preprint 2610.36525, original v1](https://arxiv.org/abs/2610.36525v1) — arXiv; primary_research - **source-2026-10-28-020:** [arXiv preprint 2610.36720, original v1](https://arxiv.org/abs/2610.36720v1) — arXiv; primary_research - **source-2026-10-28-021:** [arXiv preprint 2610.36732, original v1](https://arxiv.org/abs/2610.36732v1) — arXiv; primary_research