Pith. sign in

REVIEW 4 major objections 6 minor 35 references

Lucia: A Temporal Computing Platform for Contextual Intelligence

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Lucia claims to be the first wearable that pairs all-day comfort with continuously queryable memory.

desk verdict A reasonable hardware sketch for a wearable temporal-memory platform, but the 'all-day' uniqueness claim is undercut by the paper's own battery specs and an apples-to-oranges chart. read the letter →

arxiv 2411.12778 v1 pith:UBTGGIXF submitted 2024-11-19 cs.HC cs.AI

Weizhe Lin , Junxiao Shen This is my paper · ORCID
classification cs.HCcs.AI
keywords TemporalComputingwearabledevicecontextualmemorymultimodalsensingegocentricvideoaugmentationlargelanguagemodelsopen-sourceplatform
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Lucia is an open-source hardware and software platform for what the paper calls Temporal Computing: continuously recording a person's day and letting them query that record afterward. The paper's central claim is that a 44-gram wearable "pin" plus a separate computing "hub" achieves something no current device does—high comfort for all-day wear, high continuous perceptual capability, and direct real-time access to the recorded data. If true, this would make an externalized, searchable memory bank practical, with applications in activity analysis, mood tracking, personal safety, and natural-language questions like "Where did I leave my keys this morning?". The report presents the device's specifications, sensor suite, software stack, and example applications, and argues that decoupling sensing from computation is what resolves the wearability-versus-battery trade-off.

What carries the argument

The mechanism is a two-part hardware split: a wearable "temporal pin" carries the sensors—a 12-megapixel IMX681 camera, a BMI270 six-axis IMU, dual microphones, a proximity sensor, and an ambient light sensor—and records time-synchronized multimodal streams, while a separate "temporal hub" running Android 14 provides the processor, memory, storage, connectivity, and a larger 4000mAh battery. This decoupling lets the worn part stay light while heavy computation and power live elsewhere, and it is what the paper argues makes all-day wearability and all-day perceptual capability compatible. The companion software stack then turns raw recordings into queryable memory through natural-language question answering over stored experiences, with the option to run lightweight models on the hub's NPU or offload to a phone or cloud.

What would settle it

A controlled comparison would settle it: have participants wear Lucia and the devices named in the comparison chart for a full waking day, log actual continuous recording time on a single charge at the advertised settings, and have independent raters score comfort and obtrusiveness on a fixed protocol. If Lucia cannot record continuously through a normal day without a recharge, or if raters place it no better than existing devices on comfort, the claimed upper-right position collapses.

Watch

Extended reading notes

Core claim

The paper's central claim is that Lucia uniquely combines high all-day wearability and high all-day perceptual capability with direct real-time access to recorded data. It argues that existing devices trade off these properties: headsets are perceptually capable but not comfortable for all-day wear, action cameras and smartwatches lack continuous unobtrusive capture, and light data-sensing formats sacrifice on-device computation. By splitting the device into a 44-gram sensor pin and a separate computing hub, with wireless magnetic attachment and a claimed 160-minute pin battery and 400-minute hub battery, Lucia claims to escape this trade-off and enable continuous contextual memory that users can query in natural language.

Load-bearing premise

The whole uniqueness claim rests on the paper's own chart and its stated hardware numbers: the chart's two axes have no measurable scale or protocol, and the claim that 160/400-minute batteries and a 44-gram pin support "all-day" use is assumed, not demonstrated.

Editorial extensions

If this is right

  • If the specifications hold, researchers gain a reference wearable for collecting synchronized egocentric video, motion, audio, and proximity data over long sessions.
  • Natural-language querying over stored recordings would let users retrieve specific moments without manual annotation, moving memory augmentation from demonstration toward daily use.
  • The pin/hub split becomes a reusable design pattern: all-day wearables can keep the worn part minimal while pushing compute, battery, and interface into a separate module.
  • Because the platform is open-source and the hub runs Android 14, third-party developers can deploy custom on-device models using TensorFlow Lite or PyTorch Mobile and test contextual-intelligence applications.
  • Privacy controls—a physical privacy switch and a recording-indicator LED—are presented as load-bearing features that make continuous capture socially manageable.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper leaves implicit that the stated 160-minute pin and 400-minute hub battery life cannot cover a full waking day on one charge, so the "all-day" claim in practice depends on recharging, swapping, or recording intermittently; an actual daily-use test would determine which interpretation holds.
  • The hardware's value is contingent on the surrounding AI pipeline: without a working long-video question-answering model, the stored recordings are raw data rather than "temporal memory," so success will be decided as much by software as by the device.
  • A direct comparison against a smartphone camera worn in a chest pocket would strengthen the positioning argument, since the paper benchmarks against action cameras and smartwatches but not against the phone already in everyone's pocket.
  • The decoupling design suggests a privacy architecture worth testing: raw video stays on the pin or hub and only irreversible vector embeddings are transmitted to cloud models, which would let large-language-model reasoning run on personal data without uploading the video itself.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper presents Lucia, a two-part wearable consisting of a 44 g camera module (the 'temporal pin') and an Android-based 'temporal hub', intended for continuous egocentric capture and temporal memory. The central claim, stated in Sec. 1, is that Lucia uniquely combines high all-day wearability, high all-day perceptual capability, and direct real-time access to recorded data. The manuscript provides hardware specifications (Table 1), a qualitative comparison chart (Fig. 2), sample sensor traces (Figs. 4-6), software architecture, privacy considerations, and a set of applications explicitly described as potential. It reports no user studies, benchmarks, measurements, or implemented AI applications.

Significance. If the central claim were established, Lucia could be a useful open-source platform for egocentric perception and long-term memory research; the decoupled sensing/processing architecture and the effort to make sensor data available to researchers are genuine strengths. However, the central claim is neither measured nor internally consistent: the specification table gives 160/400 min operating times, and Fig. 2 is a self-positioned chart on undefined axes. As it stands, the paper cannot support its headline contribution.

major comments (4)
  1. [Sec. 1, Table 1] The paper's central claim that Lucia 'serves users all day without such inconveniences' is contradicted by Table 1: the camera module has 160 min of working time and the hub 400 min, so a 16-hour waking day would require roughly six camera recharges or 2.4 hub recharges. Section 1 defines 'all-day perceptual capability' as operating only 'assuming that wired charging or battery replacement is available,' which makes that axis incomparable with 'all-day wearability' (which excludes frequent charging) and directly conflicts with the Introduction's 'without such inconveniences' statement. Since the unique combination of these two properties is the central claim, this internal inconsistency is load-bearing rather than cosmetic.
  2. [Fig. 2] Figure 2 plots Lucia and competitors on two axes labeled 'All-day Wearability' and 'All-day Perceptual Capability' with no quantitative scale, measurement protocol, source of positions, or error bars. The paper provides no procedure by which a reader could reproduce any placement, and Lucia's upper-right position appears to be the authors' own judgment. The 'uniquely combines' claim therefore rests on an unverified, self-referential visualization rather than on benchmarked data.
  3. [Sec. 3.2] Section 3.2 asserts that the 44 g camera module is 'unobtrusive, allowing for all-day wear without causing user fatigue or discomfort,' but no comfort study, wear-time data, or objective measure of comfort is reported. Comfort and all-day wearability are central to the paper's positioning, so this assertion needs empirical support.
  4. [Sec. 7] The applications section explicitly labels its content as 'envisioned capabilities' and 'potential applications' using modal language throughout, yet the Introduction says the report demonstrates the device's 'necessity and efficacy' and the Conclusion calls it a 'breakthrough.' No implemented application, accuracy measurement, or latency measurement for the query/memory functionality is presented, so the demonstration claim is not supported by the manuscript's own evidence.
minor comments (6)
  1. [Sec. 1] The sentence beginning 'By harnessing the power of modern AI models and Temporal Computing principles' repeats the immediately preceding sentence; remove the duplication.
  2. [Fig. 2] The legend with checkmarks, crosses, and 'Streaming SDK Only' is not defined in the text; explain what qualifies as direct data access so the comparison is interpretable.
  3. [Sec. 2.2] The paper repeatedly calls the platform 'open-source' and 'publicly accessible' but gives no repository URL or license; adding these is necessary for the open-source claim to be checkable.
  4. [Table 1] Entries such as 'Front and Rear Camera Insertion: Pending' and 'HUB Weight /' are either missing or not explained; the table should distinguish unsupported features from unknown specifications.
  5. [Figs. 4-6] The sample sensor traces are presented without acquisition conditions, such as environment, mounting, or recording parameters; add this context so the samples are reproducible.
  6. [Sec. 6] The statement that personal data 'can be vectorized irreversibly' is not supported by any algorithmic description or security analysis; either provide details or soften the wording.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: Lucia's central claim is an unsupported qualitative assertion, not a definitional reduction.

full rationale

This paper is a hardware technical report rather than a derived quantitative result, and no circularity chain is present. The central uniqueness claim, stated in Section 1, rests on the qualitative comparison in Fig. 2, where the axes are defined by the authors and the devices are placed without a measurement protocol. That is an evidentiary weakness, but it is not circularity: no parameter is fitted and then renamed as a prediction, no equation defines the conclusion into existence, and no load-bearing step reduces to a self-citation. The self-citations in the paper (e.g., Lin et al. 2021, Shen et al. 2023, Zhang et al. 2020) support example applications in Section 7 only and do not justify the device's core hardware claims. The definition of 'all-day perceptual capability' in Section 1 assumes wired charging or battery replacement is available, while Table 1 reports 160-minute camera and 400-minute hub working times; this creates an internal-consistency concern about what 'all-day' means, but it is a validity or evidence issue, not a reduction of the conclusion to its own inputs. The paper's claimed uniqueness is asserted rather than derived, and the derivation chain is effectively absent; under the hard rules, an unsupported assertion without a self-referential reduction does not count as circularity. Therefore, no circular steps are identified and the circularity score is 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The paper offers no derivation and no fitted parameters. The claim depends on assumptions about the accuracy of the self-positioned comparison chart, the correctness of the specs, the acceptability of continuous recording, and the utility of a 1 TOPS NPU. No new theoretical entities are introduced.

assumptions (4)
  • domain assumption The qualitative device positions in Fig. 2 accurately reflect relative wearability and perceptual capability.
    The chart is the only evidence for the central uniqueness claim, but it has no defined scale or external measurement.
  • domain assumption Reported hardware specifications (44 g weight, 160/400 min working time, 1 TOPS NPU) are accurate under real-world use.
    The paper states these as facts without test data or conditions.
  • domain assumption Continuous all-day recording of a user's environment is socially acceptable and sufficiently privacy-preserving.
    The privacy switch and on-device processing are described but not evaluated; the applications sections assume users will accept wear and recording.
  • domain assumption A 1 TOPS NPU can run useful local language models for the described querying experience.
    The paper asserts the hub 'is capable of running small, quantized language models' but provides no latency, quality, or memory benchmarks.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lucia: A Temporal Computing Platform for Contextual Intelligence." pith.science (2026). https://pith.science/paper/UBTGGIXF

@misc{pith2026241112778,
  author       = {Pith},
  title        = {Pith review of: Lucia: A Temporal Computing Platform for Contextual Intelligence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UBTGGIXF}},
  note         = {Machine review of arXiv:2411.12778}
}
read the original abstract

The rapid evolution of artificial intelligence, especially through multi-modal large language models, has redefined user interactions, enabling responses that are contextually rich and human-like. As AI becomes an integral part of daily life, a new frontier has emerged: developing systems that not only understand spatial and sensory data but also interpret temporal contexts to build long-term, personalized memories. This report introduces Lucia, an open-source Temporal Computing Platform designed to enhance human cognition by capturing and utilizing continuous contextual memory. Lucia introduces a lightweight, wearable device that excels in both comfort and real-time data accessibility, distinguishing itself from existing devices that typically prioritize either wearability or perceptual capabilities alone. By recording and interpreting daily activities over time, Lucia enables users to access a robust temporal memory, enhancing cognitive processes such as decision-making and memory recall.

Figures

Figures reproduced from arXiv: 2411.12778 by the authors.

Figure 1
Figure 1. The illustration of the device. The device comprises two main components: a temporal pin designed for [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. The temporal pin of Lucia is designed for a [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Overview of sample data from the camera and the proximity sensor. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: Sample data from the IMU sensor. 0 5 10 15 20 25 30 35 40 Time (seconds) 1.00 0.75 0.50 0.25 0.00 0.25 0.50 0.75 Amplitude Sampled Waveform from the Microphone [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Sample data from the dual-microphone array. [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 19 canonical work pages

  1. [1]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774

    arXiv 2023

  2. [2]

    Peter Anderson, Qi Wu, Damien Teney, Jake Bruce, Mark Johnson, Niko S \"u nderhauf, Ian Reid, Stephen Gould, and Anton Van Den Hengel. 2018. Vision-and-language navigation: Interpreting visually-grounded navigation instructions in real environments. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3674--3683

    2018

  3. [3]

    AI Anthropic. 2024. https://www-cdn.anthropic.com/de8ba9b01c9ab7cbabf5c33b80b7bbc618857627/Model_Card_Claude_3.pdf The claude 3 model family: Opus, sonnet, haiku . Claude-3 Model Card

    work page 2024

  4. [4]

    Kirolos Ataallah, Xiaoqian Shen, Eslam Abdelrahman, Essam Sleiman, Mingchen Zhuge, Jian Ding, Deyao Zhu, Jürgen Schmidhuber, and Mohamed Elhoseiny. 2024. https://arxiv.org/abs/2407.12679 Goldfish: Vision-language understanding of arbitrarily long videos . Preprint, arXiv:2407.12679

    arXiv 2024

  5. [5]

    Carlos Bermejo, Tristan Braud, Ji Yang, Shayan Mirjafari, Bowen Shi, Yu Xiao, and Pan Hui. 2020. Vimes: A wearable memory assistance system for automatic information retrieval. In Proceedings of the 28th ACM International Conference on Multimedia, pages 3191--3200

    work page 2020

  6. [6]

    Tanvir Fatima Naik Bukht, Hameedur Rahman, Momina Shaheen, Asaad Algarni, Nouf Abdullah Almujally, and Ahmad Jalal. 2024. A review of video-based human activity recognition: Theory, methods and applications. Multimedia Tools and Applications, pages 1--47

    work page 2024

  7. [7]

    Sadil Chamishka, Ishara Madhavi, Rashmika Nawaratne, Damminda Alahakoon, Daswin De Silva, Naveen Chilamkurti, and Vishaka Nanayakkara. 2022. A voice-based real-time emotion detection technique using recurrent neural network empowered feature modelling. Multimedia Tools and Applications, 81(24):35173--35194

    work page 2022

  8. [8]

    Jakob Engel, Kiran Somasundaram, Michael Goesele, Albert Sun, Alexander Gamino, Andrew Turner, Arjang Talattof, Arnie Yuan, Bilal Souti, Brighid Meredith, et al. 2023. Project aria: A new tool for egocentric multi-modal ai research. arXiv preprint arXiv:2308.13561

    arXiv 2023

Show all 35 references
  1. [9]

    V \^a nia Figueira, Sandra Silva, In \^e s Costa, Bruna Campos, Jo \ a o Salgado, Liliana Pinho, Marta Freitas, Paulo Carvalho, Jo \ a o Marques, and Francisco Pinho. 2024. Wearables for monitoring and postural feedback in the work context: A scoping review. Sensors, 24(4):1341

    2024

  2. [10]

    Jing Gu, Eliana Stefani, Qi Wu, Jesse Thomason, and Xin Wang. 2022. https://doi.org/10.18653/v1/2022.acl-long.524 Vision-and-language navigation: A survey of tasks, methods, and future directions . In Proceedings of the 60th Annual Meeting of the Association for Computational ...

    2022 doi

  3. [11]

    Bo He, Hengduo Li, Young Kyun Jang, Menglin Jia, Xuefei Cao, Ashish Shah, Abhinav Shrivastava, and Ser-Nam Lim. 2024. Ma-lmm: Memory-augmented large multimodal model for long-term video understanding. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Rec...

    2024

  4. [12]

    Jonathan Karlsson, Fredrik Strand, Josef Bigun, Fernando Alonso-Fernandez, Kevin Hernandez-Diaz, and Felix Nilsson. 2022. Visual detection of personal protective equipment and safety gear on industry workers. arXiv preprint arXiv:2212.04794

    2022 arXiv

  5. [13]

    Weizhe Lin, Indigo Orton, Qingbiao Li, Gabriela Pavarini, and Marwa Mahmoud. 2021. Looking at the body: Automatic analysis of body gestures and self-adaptors in psychological distress. IEEE Transactions on Affective Computing, 14(2):1175--1187

    2021

  6. [14]

    Hao Lu, Xuesong Niu, Jiyao Wang, Yin Wang, Qingyong Hu, Jiaqi Tang, Yuting Zhang, Kaishen Yuan, Bin Huang, Zitong Yu, et al. 2024. Gpt as psychologist? preliminary evaluations for gpt-4v on visual affective computing. In Proceedings of the IEEE/CVF Conference on Computer Visio...

    2024

  7. [15]

    Ninad Mehendale. 2020. Facial emotion recognition using convolutional neural networks (ferc). SN Applied Sciences, 2(3):446

    2020

  8. [16]

    Preksha Pareek and Ankit Thakkar. 2021. A survey on video-based human action recognition: recent updates, datasets, challenges, and applications. Artificial Intelligence Review, 54(3):2259--2322

    2021

  9. [17]

    Immad A Shah and SukhDev Mishra. 2024. Artificial intelligence in advancing occupational health and safety: an encapsulation of developments. Journal of Occupational Health, 66(1):uiad017

    2024

  10. [18]

    Shaghayegh Shajari, Kirankumar Kuruvinashetti, Amin Komeili, and Uttandaraman Sundararaj. 2023. The emergence of ai-based wearable sensors for digital health technology: a review. Sensors, 23(23):9498

    2023

  11. [19]

    Junxiao Shen, John Dudley, and Per Ola Kristensson. 2023. Encode-store-retrieve: Enhancing memory augmentation through language-encoded egocentric perception. arXiv preprint arXiv:2308.05822

    2023 arXiv

  12. [20]

    Enxin Song, Wenhao Chai, Guanhong Wang, Yucheng Zhang, Haoyang Zhou, Feiyang Wu, Xun Guo, Tian Ye, Yan Lu, Jenq-Neng Hwang, et al. 2023. Moviechat: From dense token to sparse memory for long video understanding. arXiv preprint arXiv:2307.16449

    2023 arXiv

  13. [21]

    Enxin Song, Wenhao Chai, Tian Ye, Jenq-Neng Hwang, Xi Li, and Gaoang Wang. 2024. Moviechat+: Question-aware sparse memory for long video question answering. arXiv preprint arXiv:2404.17176

    2024 arXiv

  14. [22]

    Elena Stefana, Filippo Marciano, Diana Rossi, Paola Cocca, and Giuseppe Tomasoni. 2021. Wearable devices for ergonomics: A systematic literature review. Sensors, 21(3):777

    2021

  15. [23]

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Yonghui Wu, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, et al. 2023. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805

    2023 arXiv

  16. [24]

    Ferdews Tlili, Rim Haddad, Ridha Bouallegue, and Neila Mezghani. 2021. A real-time posture monitoring system towards bad posture detection. Wireless Personal Communications, 120(2):1207--1227

    2021

  17. [25]

    Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, et al. 2024. Qwen2-vl: Enhancing vision-language model's perception of the world at any resolution. arXiv preprint arXiv:2409.12191

    2024 arXiv

  18. [26]

    Wen Wu, Chao Zhang, and Philip C Woodland. 2023. Self-supervised representations in speech-based depression detection. In ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1--5. IEEE

    2023

  19. [27]

    Qu Yang, Mang Ye, and Bo Du. 2024. Emollm: Multimodal emotional understanding meets large language models. arXiv preprint arXiv:2406.16442

    2024 arXiv

  20. [28]

    Shukang Yin, Chaoyou Fu, Sirui Zhao, Ke Li, Xing Sun, Tong Xu, and Enhong Chen. 2023. A survey on multimodal large language models. arXiv preprint arXiv:2306.13549

    2023 arXiv

  21. [29]

    Shibo Zhang, Yaxuan Li, Shen Zhang, Farzad Shahabi, Stephen Xia, Yu Deng, and Nabil Alshurafa. 2022. Deep learning in human activity recognition with wearable sensors: A review on advances. Sensors, 22(4):1476

    2022

  22. [30]

    Yue Zhang, Ziqiao Ma, Jialu Li, Yanyuan Qiao, Zun Wang, Joyce Chai, Qi Wu, Mohit Bansal, and Parisa Kordjamshidi. 2024. Vision-and-language navigation today and tomorrow: A survey in the era of foundation models. arXiv preprint arXiv:2407.07035

    2024 arXiv

  23. [31]

    Ziheng Zhang, Weizhe Lin, Mingyu Liu, and Marwa Mahmoud. 2020. Multimodal deep learning framework for mental disorder recognition. In 2020 15th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2020), pages 344--350. IEEE

    2020

  24. [32]

    Wayne Xin Zhao, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang, Yupeng Hou, Yingqian Min, Beichen Zhang, Junjie Zhang, Zican Dong, et al. 2023. A survey of large language models. arXiv preprint arXiv:2303.18223

    2023 arXiv

  25. [33]

    Zhongna Zhou, Xi Chen, Yu-Chia Chung, Zhihai He, Tony X Han, and James M Keller. 2008. Activity analysis, summarization, and visualization for indoor human activity monitoring. IEEE transactions on circuits and systems for video technology, 18(11):1489--1498

    2008

  26. [34]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...

  27. [35]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.