pith. sign in

arxiv: 2605.29174 · v1 · pith:FAFXXLAUnew · submitted 2026-05-27 · 💻 cs.AI · cs.CR

Paper Agents, Paper Gains: An Empirical Analysis of DeFi Investment Agents

Jay Yu , Amy Zhao , Danning Sui This is my paper

Pith reviewed 2026-06-29 11:32 UTC · model grok-4.3

classification 💻 cs.AI cs.CR
keywords DeFi agentsAI trading agentson-chain analysistoken valuationSolana treasuriesmarket cap to AUMmaturity frameworkstakeholder alignment
0
0 comments X

The pith

DeFi investment agent treasuries hold over $30M in paper gains while token holders lose $191.7M collectively.

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

The paper surveys more than 1,900 AI-tagged crypto projects and narrows to investment-focused agents. It performs architectural review of two frameworks and on-chain analysis of 11 Solana treasuries tied to 925,323 token holders. The results show limited evidence of autonomous execution in many cases, with agent treasuries retaining paper gains while holders experience net losses and extreme concentration of benefits in the top wallets. Token market caps appear disconnected from treasury assets under management. The authors interpret the pattern as typical of a first-generation permissionless market and introduce a maturity framework along autonomous execution, risk-adjusted profitability, and stakeholder alignment.

Core claim

In the sample of 11 Solana-based agent treasuries, agent treasuries retain over USD 30M in paper gains while token holders collectively lost USD 191.7M, with the top 1% of wallets capturing 81.4% of all gains and market-cap-to-AUM ratios exceeding 10,000x. Many projects lack clear evidence of autonomous trade execution, aggregate user gains peaked then fell to net losses, median returns are negative on every platform, and tokens have declined 93% on average from all-time highs.

What carries the argument

Quantitative on-chain performance analysis comparing agent treasury gains to token holder returns across the 11 treasuries, combined with the proposed maturity framework measured on autonomous execution, risk-adjusted profitability, and stakeholder alignment.

If this is right

  • Many visible agent deployments appear to be basic API integrations rather than fully autonomous systems.
  • Token valuations remain weakly connected to treasury fundamentals, unlike established DeFi protocols.
  • User gains across platforms have declined from a peak of $2.4B to net losses with negative median returns.
  • The three-dimension maturity framework can be used to measure progress toward investment-grade agent systems.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the gain concentration pattern holds beyond the sample, token-based funding for AI agents may need stronger alignment tools to retain broad participation.
  • Repeating the treasury-versus-holder comparison on other chains could reveal whether the observed ratios are Solana-specific or general to early agent markets.
  • Community or protocol standards for verifiable autonomy might reduce the number of speculative agents that reach high valuations without demonstrated execution.

Load-bearing premise

The 10 curated projects and 11 Solana treasuries with attributable activity are representative of the broader set of investment-focused agents among the 1,900 surveyed projects.

What would settle it

A larger random sample drawn from the 1,900 projects that shows most deployments executing autonomous trades with aligned returns between treasuries and holders would challenge the reported patterns of limited autonomy and misaligned outcomes.

Figures

Figures reproduced from arXiv: 2605.29174 by Amy Zhao, Danning Sui, Jay Yu.

Figure 1
Figure 1. Figure 1: Architecture of ElizaOS and Virtuals Protocol with representative deployed agents. Deployed agents (top row: ai16z, Aiccelerate, BigPharmAI on ElizaOS; Vader AI, Axelrod, Ribbita on Virtuals) sit above their respective runtime and settlement layers. ElizaOS provides a modular plugin system (Actions, Providers, Evaluators) with developer-controlled LLM selection. Virtuals standardizes agent commerce through… view at source ↗
Figure 2
Figure 2. Figure 2: Platform token market capitalization (left) and treasury AUM (right) over time. ai16z/Eliza dominates both metrics. Caveats. Our methodology uses daily balance snapshots rather than individ￾ual trade events. We observe only the net change in token holdings per day, not the sequence of transactions that produced it. This approximation has two implications: (1) non-trade events such as transfers, mints, and … view at source ↗
Figure 3
Figure 3. Figure 3: Treasury PnL by platform. Left: Total PnL (realized + unrealized). Right: breakdown into FIFO realized PnL (top) and MTM unrealized PnL (bottom) [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: User PnL by platform. Left: Total PnL (realized + unrealized). Right: break￾down into FIFO realized PnL (top) and MTM unrealized PnL (bottom). Aggregate user PnL peaked at over $2 billion before collapsing to net losses. 5.3 Treasury vs. User Performance Treasury and user PnL exhibit different exposures across platforms rather than a uniform relationship (Figures 3 and 4). Treasury outcomes vary substantia… view at source ↗
Figure 5
Figure 5. Figure 5: Treasury vs. user cumulative PnL by platform. Blue lines represent treasury PnL; red lines represent aggregate user PnL. The boom-bust pattern is stark: at peak, aggregate user PnL across all plat￾forms exceeded $2.4 billion in paper gains (ai16z/Eliza alone reached $2.45B). As of our latest snapshot by end of 2026 Feb, this has collapsed to a collec￾tive loss of $191.7 million from 925,323 users across al… view at source ↗
Figure 6
Figure 6. Figure 6: User PnL distribution over time by platform. Each subplot shows the share of users in different PnL buckets (from large losses in red to large gains in green, with exact zero in gray). PnL values are forward-filled to the latest snapshot date for all wallets [PITH_FULL_IMAGE:figures/full_fig_p023_6.png] view at source ↗
read the original abstract

DeFi investment agents, systems that use AI for autonomous on-chain trading, have attained over USD 3 billion in combined token valuations since late 2024. We survey over 1,900 AI-tagged crypto projects, filter to investment-focused agents, and curate 10 representative projects spanning strategy and observability dimensions. We then conduct a deep-dive architectural analysis of two prominent agent frameworks, ElizaOS and Virtuals Protocol, and a quantitative on-chain performance analysis of 11 Solana-based agent treasuries with publicly attributable trading activity, covering 925,323 token holders. We find that current deployments remain early and heterogeneous: (1) in our sample, many projects do not yet provide clear evidence of autonomous trade execution, and developer interviews suggest that many visible deployments remain basic API integrations; (2) agent treasuries retain over USD 30M in paper gains while token holders collectively lost USD 191.7M, with the top 1% of wallets capturing 81.4% of all gains (USD 1.81B); (3) token valuations are weakly connected to treasury fundamentals, with market-cap-to-AUM ratios exceeding 10,000x versus below 1x for established DeFi protocols; and (4) aggregate user gains peaked at USD 2.4B before declining to net losses, with median returns negative on every platform and tokens declining 93% on average from all-time highs. We interpret these outcomes as characteristic of a permissionless, first-generation market in which open infrastructure enables rapid experimentation but also allows naive or speculative agents to launch before robust standards for autonomy, performance, and stakeholder alignment emerge. We therefore propose a maturity framework along three dimensions: autonomous execution, risk-adjusted profitability, and stakeholder alignment, to characterize the gap between current deployments and future investment-grade agent systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper surveys over 1,900 AI-tagged crypto projects, filters to investment-focused agents, curates 10 representative projects, analyzes two frameworks (ElizaOS and Virtuals Protocol), and performs on-chain analysis of 11 Solana-based agent treasuries (covering 925,323 holders). It reports that many deployments lack clear autonomous execution, agent treasuries hold >USD 30M in paper gains while holders lost USD 191.7M (top 1% capturing 81.4% or USD 1.81B), market-cap-to-AUM ratios exceed 10,000x, aggregate user gains peaked then declined with median negative returns and 93% average token drawdown from ATHs, and proposes a three-dimensional maturity framework (autonomous execution, risk-adjusted profitability, stakeholder alignment).

Significance. If the empirical patterns on treasury vs. holder outcomes, gain concentration, and valuation disconnects generalize beyond the analyzed subset, the work supplies concrete on-chain evidence of misalignment in early DeFi AI agents and a useful diagnostic framework for distinguishing experimental from investment-grade systems. The observational design draws directly from external records and interviews rather than fitted parameters, which strengthens external grounding but limits causal claims.

major comments (2)
  1. [quantitative on-chain performance analysis of 11 Solana-based agent treasuries] The headline results (USD 30M treasury paper gains vs. USD 191.7M holder losses, top-1% capture of 81.4%, MC/AUM >10,000x) are computed exclusively on the 11 Solana treasuries selected for publicly attributable trading activity. The manuscript notes that many of the 1,900 surveyed projects lack clear evidence of autonomous execution and that visible deployments are often basic API integrations, yet provides no robustness check, random draw of non-attributable treasuries, or comparison to the remaining curated projects to support treating these 11 as representative of investment-focused agents.
  2. [interpretation and maturity framework proposal] The interpretation that the observed outcomes are 'characteristic of a permissionless, first-generation market' rests on the selected sample without explicit discussion of how the 'publicly attributable trading activity' filter affects generalization; this selection criterion directly shapes the gain/loss disparity and concentration statistics that underwrite the central claim.
minor comments (2)
  1. [abstract and on-chain analysis] The abstract and methodology description supply aggregate dollar figures and holder counts but do not report error bars, exclusion criteria for the 11 treasuries, or full on-chain query parameters, which would improve reproducibility of the performance metrics.
  2. [survey and curation section] Clarify the exact criteria used to filter the 1,900 projects down to the 10 curated ones and to the 11 treasuries; a table listing the projects and their inclusion rationale would aid readers.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for their constructive comments, which help clarify the scope and limitations of our empirical analysis. We address each major comment below, indicating planned revisions to strengthen the manuscript's transparency on sample selection and generalizability.

read point-by-point responses

  1. Referee: [quantitative on-chain performance analysis of 11 Solana-based agent treasuries] The headline results (USD 30M treasury paper gains vs. USD 191.7M holder losses, top-1% capture of 81.4%, MC/AUM >10,000x) are computed exclusively on the 11 Solana treasuries selected for publicly attributable trading activity. The manuscript notes that many of the 1,900 surveyed projects lack clear evidence of autonomous execution and that visible deployments are often basic API integrations, yet provides no robustness check, random draw of non-attributable treasuries, or comparison to the remaining curated projects to support treating these 11 as representative of investment-focused agents.

    Authors: We selected the 11 Solana treasuries precisely because they are the subset with publicly attributable trading activity, enabling direct on-chain linkage of treasury performance, holder distributions, and token metrics to specific agents—an essential requirement for the quantitative analysis. The broader survey of 1,900 projects and curation of 10 representative ones already contextualizes that many lack clear autonomous execution. We agree that the manuscript would benefit from greater explicitness on this filter. In revision, we will expand the methods and limitations sections to: (a) detail the selection criteria and why non-attributable treasuries cannot support equivalent analysis, (b) provide a comparison of observable characteristics (e.g., deployment stage, framework usage) between the 11 analyzed treasuries and the remaining curated projects, and (c) include a sensitivity discussion of how the filter may emphasize more visible deployments. A random draw of non-attributable treasuries is not feasible, as they lack identifiable on-chain trading data for performance measurement. revision: partial

  2. Referee: [interpretation and maturity framework proposal] The interpretation that the observed outcomes are 'characteristic of a permissionless, first-generation market' rests on the selected sample without explicit discussion of how the 'publicly attributable trading activity' filter affects generalization; this selection criterion directly shapes the gain/loss disparity and concentration statistics that underwrite the central claim.

    Authors: The central claims and maturity framework are derived from and scoped to the observable, attributable subset, as these are the deployments where performance data exist. We will revise the interpretation and discussion sections to more explicitly state that the patterns (gain concentration, valuation disconnects, negative median returns) characterize investment-focused agents with detectable on-chain activity, rather than all 1,900 surveyed projects. The framework is presented as a diagnostic tool for distinguishing experimental from investment-grade systems based on these observations, and we will add language clarifying its intended applicability to similar early deployments while noting that broader generalization would require additional data on non-attributable cases. revision: yes

standing simulated objections not resolved
  • A random sample or robustness check on non-attributable treasuries cannot be performed, as these lack identifiable trading activity and on-chain data required for the performance metrics.

Circularity Check

0 steps flagged

No circularity: purely observational empirical analysis

full rationale

The paper reports survey results, on-chain treasury metrics, and interview findings drawn directly from external blockchain records and public project data. No equations, fitted parameters, predictions, or self-citations appear in the provided text; all headline statistics (USD 30M gains, 81.4% concentration, MC/AUM ratios) are computed from the selected 11 treasuries without reduction to internal assumptions or prior author work. The selection criterion is stated explicitly as a data availability filter rather than a derived claim.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on the assumption that the sampled projects and treasuries generalize to the market and that on-chain attribution correctly identifies autonomous activity; no free parameters or invented entities are introduced.

axioms (1)
  • domain assumption The 10 representative projects and 11 treasuries accurately reflect the state of DeFi investment agents.
    Used to generalize survey and performance findings beyond the specific cases examined.

pith-pipeline@v0.9.1-grok · 5867 in / 1298 out tokens · 34761 ms · 2026-06-29T11:32:54.521931+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. From Agent Identity to Agent Economy: Measuring the Operational Readiness of ERC-8004 AI Agents

    cs.CE 2026-06 unverdicted novelty 4.0

    ERC-8004 adoption on Ethereum is registration-heavy with limited metadata, services, reputation, and cross-chain activity, plus high concentration in ownership and feedback networks.

Reference graph

Works this paper leans on

49 extracted references · 11 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    Yu et al

    Pantera datboard.https://datboard.panteraresearchlab.xyz/, accessed: 2025- 12-03 20 J. Yu et al

    2025

  2. [2]

    AIXBT: Aixbt.tech – ai-driven crypto market intelligence.https://aixbt.tech/ (2025), accessed: 2025-11-21

    2025

  3. [3]

    AskSurf: Asksurf.ai – crypto’s ultimate ai.https://asksurf.ai/(2025), accessed: 2025-11-21

    2025

  4. [4]

    Axal: Axal: Quick start guide.https://docs.axal.com/getting-started/quick -start/(2025), accessed: 2025-11-21

    2025

  5. [5]

    aixvc.io/(2025), accessed: 2025-11-21

    Axelrod, A..: Axr / axelrod by aixvc: Agentic quant fund platform.https://axr. aixvc.io/(2025), accessed: 2025-11-21

    2025

  6. [6]

    In: Future Generation Computer Systems

    Bartoletti, M., Carta, S., Cimoli, T., Saia, R.: Dissecting ponzi schemes on ethereum: Identification, analysis, and impact. In: Future Generation Computer Systems. vol. 102, pp. 259–277. Elsevier (2020)

    2020

  7. [7]

    Accessed: 2025-11-21

    Bellan, R.: The promise and warning of Truth Terminal, the ai bot that secured $50,000 in bitcoin from marc andreessen.https://techcrunch.com/2024/12/1 9/the-promise-and-warning-of-truth-terminal-the-ai-bot-that-secured -50000-in-bitcoin-from-marc-andreessen/(Dec 2024), techCrunch. Accessed: 2025-11-21

    2024

  8. [8]

    Chen, Y., Yao, Z., Liu, Y., Ye, J., Yu, J., Hou, L., Li, J.: StockBench: Can LLM agents trade stocks profitably in real-world markets? arXiv preprint arXiv:2510.02209 (2025)

    work page arXiv 2025

  9. [9]

    Coinbase: Ai16z price.https://www.coinbase.com/price/ai16z(2025), accessed: 2025-09-08

    2025

  10. [10]

    CoinGecko: Top AI agents coins by market cap.https://www.coingecko.com/en /categories/ai-agents(2025), accessed: 2025-11-30

    2025

  11. [11]

    CoinMarketCap: Coinmarketcap (2025),https://coinmarketcap.com, accessed: 2025-09-06

    2025

  12. [12]

    CoinMarketCap: Top solana ecosystem tokens by market capitalization.https: //coinmarketcap.com/view/solana-ecosystem/(Jun 2025), accessed: 2025-06- 25

    2025

  13. [13]

    Journal of Behavioral and Experimental Finance50, 101518 (2025).https://doi.org/10.1016/j.jbef.2025.101518, https://doi.org/10.1016/j.jbef.2025.101518

    Conlon, T., Corbet, S., et al.: Memecoin contagion: Irrationality, illicit behaviour, and contagion in rapid memecoin growth. Journal of Behavioral and Experimental Finance50, 101518 (2025).https://doi.org/10.1016/j.jbef.2025.101518, https://doi.org/10.1016/j.jbef.2025.101518

    work page doi:10.1016/j.jbef.2025.101518 2025

  14. [14]

    Crypto.com University: What is elizaos?https://crypto.com/es/university/w hat-is-elizaos(2025), accessed: 2025-09-08

    2025

  15. [15]

    In: IEEE S&P (2020)

    Daian, P., Goldfeder, S., Kell, T., Li, Y., Zhao, X., Bentov, I., Breidenbach, L., Juels, A.: Flash boys 2.0: Frontrunning in decentralized exchanges, miner ex- tractable value, and consensus instability. In: IEEE S&P (2020)

    2020

  16. [16]

    https://decrypt.co/295717/meet-ai16z-dao-an-ai-based-investment-proje ct-that-aims-to-upend-silicon-valley(2024), accessed: 2026-05-05

    Decrypt: Meet ai16z, an AI-based investment DAO inspired by marc andreessen. https://decrypt.co/295717/meet-ai16z-dao-an-ai-based-investment-proje ct-that-aims-to-upend-silicon-valley(2024), accessed: 2026-05-05

    2024

  17. [17]

    Decrypt: AI DAO ai16z becomes elizaos amid branding confusion concerns.https: //decrypt.co/303244/ai-dao-ai16z-becomes-elizaos-amid-branding-confu sion-concerns(2025), accessed: 2026-05-05

    2025

  18. [18]

    Dune Analytics: Dune: Onchain data platform.https://dune.com/(2025), ac- cessed: 2025-02-14

    2025

  19. [19]

    https://dune.com/queries/6697062(2025), accessed: 2025-02-14

    Dune Analytics: Platform treasury PnL with FIFO realized and MTM unrealized. https://dune.com/queries/6697062(2025), accessed: 2025-02-14

    work page arXiv 2025

  20. [20]

    Dune Analytics: Token daily price snapshots.https://dune.com/queries/66972 97(2025), accessed: 2025-02-14

    2025

  21. [21]

    Dune Analytics: User daily balance snapshots for agent platform tokens.https: //dune.com/queries/6697157(2025), accessed: 2025-02-14 Paper Agents, Paper Gains 21

    work page arXiv 2025

  22. [22]

    ElizaOS: Elizaos developer documentation: Plugin components.https://docs.e lizaos.ai/plugins/components(2025), accessed: 2025-11-30

    2025

  23. [23]

    ElizaOS: Elizaos plugin registry.https://github.com/elizaos-plugins/regis try(2025), accessed: 2025-11-30

    2025

  24. [24]

    ElizaOS: Elizaos.ai – agentic operating system for ai agents.https://www.elizao s.ai/(2025), accessed: 2025-11-21

    2025

  25. [25]

    ElizaOS: plugin-hyperliquid: Hyperliquid DEX integration for ElizaOS.https: //github.com/elizaos-plugins/plugin-hyperliquid(2025), accessed: 2025-11- 30

    2025

  26. [26]

    ElizaOS: plugin-solana: Core Solana blockchain plugin for ElizaOS.https://gith ub.com/elizaos-plugins/plugin-solana(2025), accessed: 2025-11-30

    2025

  27. [27]

    arXiv preprint arXiv:2601.00523 (2026)

    F´ abrega, A., Austgen, J., Breckenridge, S., Yu, J., Zhao, A., Allen, S., Saraf, A., Juels, A.: The CoinAlg bind: Profitability-fairness tradeoffs in collective investment algorithms. arXiv preprint arXiv:2601.00523 (2026)

    work page arXiv 2026

  28. [28]

    Ai-trader: Benchmarking autonomous agents in real-time financial markets.arXiv preprint arXiv:2512.10971,

    Fan, T., Yang, Y., Jiang, Y., Zhang, Y., Chen, Y., Huang, C.: AI-Trader: Benchmarking autonomous agents in real-time financial markets. arXiv preprint arXiv:2512.10971 (2025)

    work page arXiv 2025

  29. [29]

    Giza: Giza protocol documentation.https://docs.gizaprotocol.ai/(2025), accessed: 2025-11-21

    2025

  30. [30]

    Columbia Law Review (2017)

    Ji, M.: Are Robots Good Fiduciaries? Regulating Robo-Advisors Under the In- vestment Advisers Act of 1940. Columbia Law Review (2017)

    1940

  31. [31]

    In: Financial Cryptography and Data Security (FC) (2023)

    Kell, T., Cr´ epault, H., Hua, Y., Liu, S., Sun, X., Mudgal, C., Juels, A.: Forsage: Anatomy of a smart-contract pyramid scheme. In: Financial Cryptography and Data Security (FC) (2023)

    2023

  32. [32]

    LLM-Powered Multi-Agent System for Automated Crypto Portfolio Management

    Luo, Y., Feng, Y., Xu, J., Tasca, P., Liu, Y.: LLM-powered multi-agent system for automated crypto portfolio management. arXiv preprint arXiv:2501.00826 (2025)

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  33. [33]

    Messari: Crypto fundraising database.https://messari.io/fundraising-data (2025), accessed: 2025-11-30

    2025

  34. [34]

    NoF1: Nof1.ai – project homepage.https://nof1.ai/(2025), accessed: 2025-11-21

    2025

  35. [35]

    Numerai: Numerai crypto – overview.https://docs.numer.ai/numerai-crypto/ crypto-overview(2025), accessed: 2025-11-21

    2025

  36. [36]

    arXiv preprint arXiv:2510.11695 (2025)

    Qian, L., Peng, X., Wang, Y., Zhang, V.J., He, H., Smith, H., Han, Y., He, Y., Li, H., Cao, Y., Yu, Y., Lopez-Lira, A., Lu, P., Nie, J.Y., Xiong, G., Huang, J., Ananiadou, S.: When agents trade: Live multi-market trading benchmark for LLM agents. arXiv preprint arXiv:2510.11695 (2025)

    work page arXiv 2025

  37. [37]

    In: Proceedings of the 21st ACM internet measurement conference

    Qin, K., Zhou, L., Gamito, P., Jovanovic, P., Gervais, A.: An empirical study of defi liquidations: Incentives, risks, and instabilities. In: Proceedings of the 21st ACM internet measurement conference. pp. 336–350 (2021)

    2021

  38. [38]

    In: International conference on financial cryptography and data security

    Qin, K., Zhou, L., Livshits, B., Gervais, A.: Attacking the defi ecosystem with flash loans for fun and profit. In: International conference on financial cryptography and data security. pp. 3–32. Springer (2021)

    2021

  39. [39]

    In: 2024 IEEE International Conference on Blockchain and Cryptocurrency (ICBC 2024)

    Sharma, T., Potter, Y., Pongmala, K., Wang, H., Miller, A., Song, D., Wang, Y.: Unpacking how decentralized autonomous organizations (daos) work in practice. In: 2024 IEEE International Conference on Blockchain and Cryptocurrency (ICBC 2024). pp. 416–424. Institute of Electrical and Electronics Engineers (IEEE), Dublin, Ireland (2024).https://doi.org/10.1...

    work page doi:10.1109/icbc59979.2024.10634404 2024

  40. [40]

    Terminal, T.: Truthterminal.wiki – the truth terminal project.https://truthter minal.wiki/(2025), accessed: 2025-11-21

    2025

  41. [41]

    Yu et al

    TrueNorth: Truenorth – crypto’s first ai discovery engine.https://www.true-nor th.xyz/(2025), accessed: 2025-11-21 22 J. Yu et al

    2025

  42. [42]

    Virtuals Protocol: Agent commerce protocol: Technical deep dive.https://whit epaper.virtuals.io/about-virtuals/agent-commerce-protocol-acp/technic al-deep-dive(2025), accessed: 2025-11-30

    2025

  43. [43]

    https://whitepaper.virtuals.io/builders- hub/game- framework(2025), accessed: 2025-11-30

    Virtuals Protocol: GAME framework: Generative autonomous multimodal entities. https://whitepaper.virtuals.io/builders- hub/game- framework(2025), accessed: 2025-11-30

    2025

  44. [44]

    Virtuals Protocol: GAME Python SDK.https://github.com/game-by-virtual s/game-python(2025), accessed: 2025-11-30

    2025

  45. [45]

    Virtuals Protocol: Virtuals protocol smart contracts.https://github.com/Virtu al-Protocol/protocol-contracts(2025), accessed: 2025-11-30

    2025

  46. [46]

    io/#background(2025), accessed: 2025-06-25

    Virtuals Protocol: Virtuals protocol whitepaper.https://whitepaper.virtuals. io/#background(2025), accessed: 2025-06-25

    2025

  47. [47]

    Walters, S., Gao, S., Nerd, S., Da, F., Williams, W., Meng, T.C., Chow, A., Han, H., He, F., Zhang, A., Wu, M., Shen, T., Hu, M., Yan, J.: Eliza: A web3 friendly ai agent operating system (2025),https://arxiv.org/abs/2501.06781

    work page arXiv 2025

  48. [48]

    In: Proceedings of the 28th USENIX Security Symposium

    Xu, J., Livshits, B.: The anatomy of a cryptocurrency pump-and-dump scheme. In: Proceedings of the 28th USENIX Security Symposium. pp. 1609–1625 (2019)

    2019

  49. [49]

    Fingpt: Open-source financial large language models.arXiv preprint arXiv:2306.06031,

    Yang, H., Liu, X.Y., Wang, C.D.: FinGPT: Open-source financial large language models. arXiv preprint arXiv:2306.06031 (2023) Paper Agents, Paper Gains 23 A User PnL Distribution Over Time Fig. 6.User PnL distribution over time by platform. Each subplot shows the share of users in different PnL buckets (from large losses in red to large gains in green, wit...

    work page arXiv 2023