Pith. sign in

REVIEW 4 major objections 4 minor 1 cited by

Demystifying AI Agents: The Final Generation of Intelligence

T0 review · 4 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Paper calls AI agents the final generation of intelligence

desk verdict A readable AI-milestones survey whose central '5.9-month doubling' claim is unsupported and contradicted by the paper's own Table VII; more position paper than research. read the letter →

arxiv 2505.09932 v1 pith:WZBKIN6R submitted 2025-05-15 cs.AI cs.ETcs.LGcs.MA

Kevin J McNamara , Rhea Pritham Marpu This is my paper · ORCID
classification cs.AIcs.ETcs.LGcs.MA
keywords AIagentslargelanguagemodelschain-of-thoughtpromptingreinforcementlearningfromhumanfeedbackretrieval-augmentedgenerationintelligencedoublingTransformerarchitectureartificialgeneral
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

This paper argues that today's AI agents—large language models wrapped with tools and human-feedback training—are a culminating "final generation" of intelligence, not just another step in a long line of systems. It reads the history from the Logic Theorist and ELIZA through GPT-3, GPT-4, and tool-using agents as a convergence of prompting, training, hardware, architecture, and plug-in integration. The quantitative claim at the center is that AI performance on major benchmarks doubles roughly every six months, and that this pace could put human-surpassing performance within a decade. A sympathetic reader would take the paper's contribution as a synthesis and a warning: the social stakes follow from current capability and its growth rate, so planning should start now.

What carries the argument

The load-bearing object is the modern AI agent, defined as a Transformer-based language model equipped with chain-of-thought prompting, RLHF alignment, retrieval-augmented generation or live tools, and enough hardware to scale. The paper treats this stack as the mechanism that turns pretraining into goal-directed action, and it uses the stack's benchmark scores as the evidence that intelligence is doubling. The second piece of machinery is the doubling estimate itself: an extrapolation of benchmark gains on a roughly 5.9-month cycle, which converts the historical record into a forward-looking prediction of human-level AI within a decade.

What would settle it

Watch the same benchmarks over the next two years: if MMLU or GSM8K scores plateau, slow below the 5.9-month doubling rate, or fail to rise when new models are released, the central claim is contradicted; saturation near 100 percent alone is enough to falsify a continued exponential.

Watch

Extended reading notes

Core claim

The central discovery, on the paper's own terms, is that the separate advances of the last decade have converged into a single new kind of system: an agent that can reason step by step, retrieve current information, call external tools, and act on goals. The paper calls this the final generation of intelligence "as we currently conceive it," and backs the label with benchmark data: MMLU from roughly 25 percent to above 86 percent in under three years, GSM8K from 17.7 percent to 92 percent with chain-of-thought, and an estimated doubling of capability every 5.9 months. If the paper is right, current systems are not an intermediate stage but a plateau-or-singularity boundary, and the next questions are ethical and managerial rather than primarily architectural.

Load-bearing premise

The load-bearing premise is that score gains on benchmarks like MMLU, GSM8K, and HumanEval measure one underlying quantity called "intelligence," and that the recent doubling rate will continue into the future unchanged.

Editorial extensions

If this is right

  • If the six-month doubling holds, systems scoring at or below human average on broad tests today should match or exceed human performance on many cognitive benchmarks within roughly ten years.
  • The "final generation" framing implies that the main remaining work is integration, safety, and governance of existing agent capabilities rather than invention of new core architectures.
  • Planning that assumes linear progress will underestimate capability growth; the paper's compounding rate implies a capability roughly four times larger each year.
  • The paper's own caveats tie the benefits to governance: whether agents cure diseases and personalize education or deepen inequality depends on deployment choices, not on the technology alone.

Reading between the lines

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

  • The paper's doubling rate blends compute growth, algorithmic efficiency, and benchmark design; a capability metric separated from hardware scaling could double more slowly or faster, so the six-month figure should be read as a composite rather than a clock.
  • The "final generation" label implies a plateau that the same tool-using agents could undermine: agents that design experiments or write code could accelerate progress beyond the historical doubling, making the label self-limiting.
  • A testable extension the paper does not run is to hold one agent fixed on a private, unpublished task set across several model releases, separating genuine capability growth from benchmark saturation.
  • The societal concerns listed in the paper—job displacement, bias, energy use, and accountability—do not actually depend on the doubling claim; they would be urgent even if progress flatlined tomorrow.
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 / 4 minor

Summary. This manuscript is a position/survey paper that argues that recent AI systems, particularly large language model agents with tool integration, represent the 'final generation of intelligence' before a possible singularity or plateau. Its central evidence is an acceleration claim: benchmark performance such as MMLU allegedly rose from 25% to 86% in under three years, and 'intelligence metrics' are said to double approximately every 5.9 months, leading to a projection that AI will surpass human intelligence within a decade. The paper surveys prompting techniques, training methods, hardware evolution, architecture, and tool use as converging factors behind this acceleration.

Significance. If substantiated, the acceleration thesis would be a claim of major societal and scientific importance. The paper does provide a readable chronology of AI agent technologies and correctly identifies several real trends, such as the growth of context windows and the role of RLHF. However, the central quantitative argument is not supported: the 5.9-month doubling period is contradicted by the paper's own Table VII, the cited source for it reports compute-doubling rather than benchmark-doubling rates, and many specific empirical figures are uncited or misattributed. As a result, the paper's headline claims, including the prediction of human-surpassing AI within a decade, are not credible on the evidence presented.

major comments (4)
  1. [Section IX, Table VII] The text claims that AI performance is 'doubling approximately every 5.9 months' and implies a fourfold annual increase, but Table VII shows benchmark gains of only 2–3x for MMLU, 5.2x for GSM8K, and 2.3x for HumanEval over the 2020–2023 period. A true 5.9-month doubling across 36 months would imply a factor of roughly 2^(36/5.9) ≈ 68x, which is more than an order of magnitude larger than any row in the table. Moreover, MMLU and GSM8K are percentage-bounded benchmarks, so a literal doubling rate cannot persist beyond a few doublings; the paper never defines the unbounded 'intelligence metric' that is supposed to double. The projection that AI will surpass human intelligence within a decade is therefore not derivable from the paper's own evidence.
  2. [Section IX, reference [55]] The '5.9-month doubling' statement is attributed to Epoch AI via reference [55], which is the paper 'Compute Trends Across Three Eras of Machine Learning' by Sevilla et al. That paper analyzes growth in training compute, not doubling of benchmark performance or of any 'intelligence' measure. No other source or derivation is provided for a benchmark-doubling rate. Since the entire forward-looking conclusion rests on this number, the central extrapolation is unsupported by the cited literature.
  3. [Sections II-B, III, VII-A; Tables I-III] Many quantitative claims are presented without sufficient sourcing, and several citations do not support the asserted numbers. For example: Section II-B states that Self-Consistency reduced hallucination rates from 21% to 15.8% and cites [13], but Wang et al. 2022 does not report such hallucination-rate outcomes; the '63% to 82%' arithmetic-reasoning improvement is attributed to [14], a paper about training verifiers, not structured prompts; Table I reports 'Legal reasoning ∼40%/∼80%' and other accuracies with no study name or confidence interval; Table II reports zero-shot/few-shot gains for BERT, LLaMA, and CodeGen without sources containing those exact numbers; and Table VII gives no citations at all for its benchmark values. Because the paper's conclusion claims 'the evidence is stark,' the unreliability of this evidence is a load-bearing problem.
  4. [Section X] The core concepts of the thesis—'intelligence', 'final generation of intelligence', and 'surpassing human intelligence'—are never operationally defined. The paper does not state what measurements would confirm or falsify the claim that intelligence doubles every six months, nor what would count as having reached a 'final generation'. This makes the central claim unfalsifiable as presented, and it cannot be evaluated scientifically without such definitions.
minor comments (4)
  1. [Figures 1-4] The manuscript contains captions for Fig. 1, Fig. 2, Fig. 3, and Fig. 4, but the actual figure images are not included in the text, so the reader cannot inspect the 'Evolution of AI Capabilities' or the other visualized content.
  2. [Table IV] The TFLOPS values for the V100, A100, and H100 are listed without specifying the precision (FP16, FP32, sparse vs. dense) or the exact configuration, which makes the comparison ambiguous; for example, the H100's '1000+ TFLOPS' typically refers to sparse FP8 rather than dense throughput.
  3. [Section II-B, reference [13]] The text says 'Stanford researchers demonstrated' a self-consistency result, but reference [13] lists authors from multiple institutions including Google Research, not predominantly Stanford; the attribution is inaccurate.
  4. [Section VII-B, reference [55]] The bullet about error management states that tool integration failures can increase task error rates by 10% and cites [55], but reference [55] is the compute-trends paper and contains no such claim about tool integration error rates.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper performs no fitted derivation and its doubling claim is cited to external Epoch AI work, not constructed from its own inputs.

full rationale

The paper does not derive its central quantitative claim from its own fitted parameters or definitions. The claim that AI benchmark performance is 'effectively doubling approximately every 5.9 months' is explicitly attributed to Epoch AI via reference [55], and Table VII is presented as illustrative evidence rather than as the source from which the doubling rate is fit. There are no equations linking inputs to outputs, no parameters fitted to a subset of data and then renamed as a prediction, and no self-citations or author-imported uniqueness theorems. The projection in Section IX is conditional ('If capabilities continue to double every six months, we may see AI systems surpassing human intelligence within the next decade'), so the conclusion is not forced by construction. The mismatch between Table VII's 2–5x gains over roughly three years and the 5.9-month doubling rate, as well as the undefined meaning of 'intelligence' being doubled, are substantive correctness and support concerns, but they are not circularity: the paper's acceleration claim rests on an external citation rather than on a derivation that reduces to its own inputs. Accordingly, the circularity burden is minimal and the honest finding is no significant circularity.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The central claim depends on the assumptions that benchmark gains map to a single quantity called intelligence, that the observed doubling rate persists, and that a few commercial systems fairly represent all AI agents. The only quantitative parameter, the 5.9-month doubling period, is imported from an external source without uncertainty. No invented entities are introduced.

free parameters (1)
  • capability doubling period = ~5.9 months
    The paper asserts that AI benchmark performance doubles every 5.9 months, citing Epoch AI [55], and uses this as the basis for the 'intelligence doubling' claim. No confidence interval or methodology is provided in this paper.
assumptions (3)
  • domain assumption Benchmark scores measure general intelligence.
    The paper equates performance on MMLU, GSM8K, HumanEval, etc., with 'intelligence metrics' and uses their doubling as evidence that AI is approaching human-level intelligence (Section IX).
  • domain assumption The historical doubling rate will continue into the future.
    The projection that 'we may see AI systems surpassing human intelligence within the next decade' assumes the 5.9-month doubling rate does not plateau or slow (Section IX).
  • domain assumption The cited vendor systems are representative of a general class of agents.
    The paper generalizes from specific examples (ChatGPT with plugins, Grok) to a universal 'AI agent' category without defining boundaries or evaluating selection bias (Section VIII).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Demystifying AI Agents: The Final Generation of Intelligence." pith.science (2026). https://pith.science/paper/WZBKIN6R

@misc{pith2026250509932,
  author       = {Pith},
  title        = {Pith review of: Demystifying AI Agents: The Final Generation of Intelligence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WZBKIN6R}},
  note         = {Machine review of arXiv:2505.09932}
}
read the original abstract

The trajectory of artificial intelligence (AI) has been one of relentless acceleration, evolving from rudimentary rule-based systems to sophisticated, autonomous agents capable of complex reasoning and interaction. This whitepaper chronicles this remarkable journey, charting the key technological milestones--advancements in prompting, training methodologies, hardware capabilities, and architectural innovations--that have converged to create the AI agents of today. We argue that these agents, exemplified by systems like OpenAI's ChatGPT with plugins and xAI's Grok, represent a culminating phase in AI development, potentially constituting the "final generation" of intelligence as we currently conceive it. We explore the capabilities and underlying technologies of these agents, grounded in practical examples, while also examining the profound societal implications and the unprecedented pace of progress that suggests intelligence is now doubling approximately every six months. The paper concludes by underscoring the critical need for wisdom and foresight in navigating the opportunities and challenges presented by this powerful new era of intelligence.

Figures

Figures reproduced from arXiv: 2505.09932 by the authors.

Figure 1
Figure 1. Evolution of AI Capabilities and the integration of external tools granting AI agency in the digital world. Together, these threads weave the story of how AI grew from programmed logic into agents that increasingly mirror, and in some cases surpass, human ingenuity, setting the stage for an era defined by intelligent machines. II. EARLY STAGES OF AI MATURITY: THE FIRST STIRRINGS OF UNDERSTANDING Every technological … view at source ↗
Figure 2
Figure 2. Chain-of-Thought Reasoning Process received a shipment of 60 cartons. How many cartons of eggs do they have now? Let’s think step by step: 1) The store starts with 150 cartons. 2) They sold 35 cartons on Monday, so they have 150 - 35 = 115 cartons left. 3) They received 60 cartons on Tuesday, so they now have 115 + 60 = 175 cartons. 4) Therefore, the final answer is 175.” By prompting the model to articulate the int… view at source ↗
Figure 3
Figure 3. Context Window Size Evolution Visualization [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: AI Agent Architecture • Tool Selection: Effective use requires sophisticated plan￾ning capabilities Tool integration represents a critical advancement in AI’s evolution from passive text generation to active digital agency. It establishes the foundation for versatile A…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Rethinking Testing for LLM Applications: Characteristics, Challenges, and a Lightweight Interaction Protocol

    cs.SE 2025-08 conditional novelty 5.0 of 10

    This position paper classifies testing methods for LLM applications into three layers and proposes AICL, a structured protocol for testable agent communication; neither the framework nor the protocol is empirically validated.

Reference graph

Works this paper leans on

58 extracted references · 33 canonical work pages · cited by 1 Pith paper

  1. [55]

    Compute Trends Across Three Eras of Machine Learning,

    J. Sevilla et al., “Compute Trends Across Three Eras of Machine Learning,” Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society (AIES ’22) , pp. 749–762, Aug. 2022. DOI: 10.1145/3514094.3534166. [Online]. Available: https://epochai.org/

    arXiv 2022

  2. [50]

    Claude 2,

    Anthropic, “Claude 2,” Anthropic Blog, Jul. 11, 2023. [Online]. Avail- able: https://www.anthropic.com/index/claude-2

    work page 2023

  3. [22]

    RankCSE: Unsupervised Sentence Representations via Learning to Rank,

    K. Krishna et al., “RankCSE: Unsupervised Sentence Representations via Learning to Rank,” Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (ACL 2023) , pp. 8876–8890, Jul. 2023. DOI: 10.18653/v1/2023.acl-long.492

    work page doi:10.18653/v1/2023.acl-long.492 2023

  4. [13]

    Self-Consistency Improves Chain of Thought Rea- soning in Language Models,

    X. Wang et al., “Self-Consistency Improves Chain of Thought Rea- soning in Language Models,” arXiv:2203.11171, Mar. 2022. [Online]. Available: https://arxiv.org/abs/2203.11171

    arXiv 2022

  5. [14]

    Training Verifiers to Solve Math Word Problems,

    K. Cobbe et al., “Training Verifiers to Solve Math Word Problems,” arXiv:2110.14168, Oct. 2021. [Online]. Available: https://arxiv.org/abs/ 2110.14168

    arXiv 2021

  6. [1]

    Empirical explorations with the logic theory machine: A case study in heuristics,

    A. Newell, J. C. Shaw, and H. A. Simon, “Empirical explorations with the logic theory machine: A case study in heuristics,” Proceedings of the Western Joint Computer Conference , pp. 218–239, 1957. DOI: 10.1145/1455292.1455311

    work page arXiv 1957

  7. [2]

    ELIZA–a computer program for the study of nat- ural language communication between man and machine,

    J. Weizenbaum, “ELIZA–a computer program for the study of nat- ural language communication between man and machine,” Commu- nications of the ACM , vol. 9, no. 1, pp. 36–45, Jan. 1966. DOI: 10.1145/365153.365168

    arXiv 1966

  8. [3]

    GPT-4 Technical Report,

    OpenAI, “GPT-4 Technical Report,” arXiv:2303.08774, Mar. 2023. [Online]. Available: https://arxiv.org/abs/2303.08774

    arXiv 2023

Show all 58 references
  1. [4]

    Cramming more components onto integrated circuits,

    G. E. Moore, “Cramming more components onto integrated circuits,” Electronics, vol. 38, no. 8, pp. 114–117, Apr. 1965

    1965

  2. [5]

    Attention is All You Need,

    A. Vaswani et al., “Attention is All You Need,” Advances in Neu- ral Information Processing Systems 30 (NIPS 2017) , pp. 5998– 6008, 2017. [Online]. Available: https://papers.nips.cc/paper/2017/hash/ 3f5ee243547dee91fbd053c1c4a845aa-Abstract.html

    2017

  3. [6]

    Training language models to follow instruc- tions with human feedback,

    L. Ouyang et al., “Training language models to follow instruc- tions with human feedback,” Advances in Neural Information Pro- cessing Systems 35 (NeurIPS 2022) , pp. 27730–27744, 2022. [On- line]. Available: https://proceedings.neurips.cc/paper files/paper/2022/ hash/b1efde53...

    2022

  4. [7]

    Deep Blue,

    M. Campbell, A. J. Hoane Jr., and F.-H. Hsu, “Deep Blue,” Artificial In- telligence, vol. 134, no. 1–2, pp. 57–83, Jan. 2002. DOI: 10.1016/S0004- 3702(01)00129-1

    2002 doi

  5. [8]

    Language Models are Few-Shot Learners,

    T. Brown et al., “Language Models are Few-Shot Learners,” Advances in Neural Information Processing Systems 33 (NeurIPS 2020) , pp. 1877– 1901, 2020. [Online]. Available: https://papers.nips.cc/paper/2020/hash/ 1457c0d6bfcb4967418bfb8ac142f64a-Abstract.html

    2020

  6. [9]

    Prompting Large Language Models for Legal Analysis Tasks,

    L. Reynolds and K. D. Ashley, “Prompting Large Language Models for Legal Analysis Tasks,”Proceedings of the 19th International Conference on Artificial Intelligence and Law (ICAIL 2023), pp. 278–287, Jun. 2023. DOI: 10.1145/3594536.3595132

    2023

  7. [10]

    Evaluating Large Language Models for Medical Diagno- sis,

    X. Liu et al., “Evaluating Large Language Models for Medical Diagno- sis,” Stanford HAI & Stanford Medicine Research Update , Feb. 2022

    2022

  8. [11]

    Learning Transferable Visual Models From Natural Language Supervision,

    A. Radford et al., “Learning Transferable Visual Models From Natural Language Supervision,” Proceedings of the 38th International Confer- ence on Machine Learning (ICML 2021) , PMLR vol. 139, pp. 8748– 8763, Jul. 2021. [Online]. Available: https://proceedings.mlr.press/v139/ r...

    2021

  9. [12]

    GitHub Copilot: Flying Ahead with Speed and Accu- racy,

    A. Ziegler, “GitHub Copilot: Flying Ahead with Speed and Accu- racy,” GitHub Blog, Jun. 2022. [Online]. Available: https://github.blog/ 2022-06-21-github-copilot-is-generally-available-to-all-developers/

    2022

  10. [15]

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding,

    J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding,” Pro- ceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologie...

    2019 doi

  11. [16]

    Improving language models by retrieving from trillions of tokens,

    S. Borgeaud et al., “Improving language models by retrieving from trillions of tokens,” Proceedings of the 39th International Confer- ence on Machine Learning (ICML 2022) , PMLR vol. 162, pp. 2206– 2240, Jul. 2022. [Online]. Available: https://proceedings.mlr.press/v162/ borge...

    2022

  12. [17]

    Llama 2: Open Foundation and Fine-Tuned Chat Models,

    H. Touvron et al., “Llama 2: Open Foundation and Fine-Tuned Chat Models,” arXiv:2307.09288, Jul. 2023. [Online]. Available: https://arxiv. org/abs/2307.09288

    2023 arXiv

  13. [18]

    CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis,

    E. Nijkamp et al., “CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis,” arXiv:2203.13474, Mar. 2022. [Online]. Available: https://arxiv.org/abs/2203.13474

    2022 arXiv

  14. [19]

    Chain-of-Thought Prompting Elicits Reasoning in Large Language Models,

    J. Wei et al., “Chain-of-Thought Prompting Elicits Reasoning in Large Language Models,” Advances in Neural Information Pro- cessing Systems 35 (NeurIPS 2022) , pp. 24824–24837, 2022. [On- line]. Available: https://proceedings.neurips.cc/paper files/paper/2022/ hash/9d560961352...

    2022

  15. [20]

    Large Language Models are Zero-Shot Reasoners,

    T. Kojima et al., “Large Language Models are Zero-Shot Reasoners,” Advances in Neural Information Processing Systems 35 (NeurIPS 2022) , pp. 22199–22213, 2022. [Online]. Available: https://proceedings.neurips.cc/paper files/paper/2022/hash/ 8bb0d291acd4acf06ef112099c16f326-Abs...

    2022

  16. [21]

    Building Watson: An Overview of the DeepQA Project,

    D. Ferrucci et al., “Building Watson: An Overview of the DeepQA Project,” AI Magazine , vol. 31, no. 3, pp. 59–79, Fall 2010. DOI: 10.1609/aimag.v31i3.2303

    2010 doi

  17. [23]

    TruthfulQA: Measuring How Models Mimic Human Falsehoods,

    S. Lin, J. Hilton, and O. Evans, “TruthfulQA: Measuring How Models Mimic Human Falsehoods,” Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (ACL 2022) , pp. 3214– 3252, May 2022. DOI: 10.18653/v1/2022.acl-long.229

    2022 doi

  18. [24]

    Deep reinforcement learning from human pref- erences,

    P. F. Christiano et al., “Deep reinforcement learning from human pref- erences,” Advances in Neural Information Processing Systems 30 (NIPS 2017), pp. 4299–4307, 2017. [Online]. Available: https://papers.nips.cc/ paper/2017/hash/d5e2c0adad503c91f91df240d0cd4e49-Abstract.html

    2017

  19. [25]

    ChatGPT Sets Record for Fastest-Growing User Base,

    K. Hu, “ChatGPT Sets Record for Fastest-Growing User Base,” Reuters, Feb. 2, 2023. [Online]. Available: https://www.reuters.com/technology/ chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/

    2023

  20. [26]

    Constitutional AI: Harmlessness from AI Feedback,

    Y . Bai et al., “Constitutional AI: Harmlessness from AI Feedback,” arXiv:2212.08073, Dec. 2022. [Online]. Available: https://arxiv.org/abs/ 2212.08073

    2022 arXiv

  21. [27]

    BioBERT: a pre-trained biomedical language representa- tion model for biomedical text mining,

    J. Lee et al., “BioBERT: a pre-trained biomedical language representa- tion model for biomedical text mining,” Bioinformatics, vol. 36, no. 4, pp. 1234–1240, Feb. 2020. DOI: 10.1093/bioinformatics/btz682

    2020 doi

  22. [28]

    Mastering the game of Go with deep neural networks and tree search,

    D. Silver et al., “Mastering the game of Go with deep neural networks and tree search,” Nature, vol. 529, no. 7587, pp. 484–489, Jan. 2016. DOI: 10.1038/nature16961

    2016 doi

  23. [29]

    Retrieval-Augmented Generation for Knowledge- Intensive NLP Tasks,

    P. Lewis et al., “Retrieval-Augmented Generation for Knowledge- Intensive NLP Tasks,” Advances in Neural Information Processing Systems 33 (NeurIPS 2020) , pp. 9459–9474,

    2020

  24. [30]

    DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter,

    V . Sanh, L. Debut, J. Chaumond, and T. Wolf, “DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter,” arXiv:1910.01108, Oct. 2019. [Online]. Available: https://arxiv.org/abs/1910.01108

    1910 arXiv

  25. [31]

    MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices,

    Z. Sun et al., “MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices,” Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL 2020) , pp. 2158– 2170, Jul. 2020. DOI: 10.18653/v1/2020.acl-main.194

    2020 doi

  26. [32]

    NVIDIA A100 Tensor Core GPU Architecture,

    NVIDIA, “NVIDIA A100 Tensor Core GPU Architecture,” NVIDIA Technical Brief , May 2020. [Online]. Available: https://www.nvidia.com/content/dam/en-zz/Solutions/Data-Center/ nvidia-ampere-architecture-whitepaper.pdf

    2020

  27. [33]

    NVIDIA H100 Tensor Core GPU Architecture,

    NVIDIA, “NVIDIA H100 Tensor Core GPU Architecture,” NVIDIA Technical Brief, Mar. 2022. [Online]. Available: https://resources.nvidia. com/en-us-tensor-core/nvidia-hopper-architecture-whitepaper

    2022

  28. [34]

    Energy and Policy Consid- erations for Deep Learning in NLP,

    E. Strubell, A. Ganesh, and A. McCallum, “Energy and Policy Consid- erations for Deep Learning in NLP,” Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (ACL 2019) , pp. 3645–3650, Jul. 2019. DOI: 10.18653/v1/P19-1355

    2019 doi

  29. [35]

    PaLM: Scaling Language Modeling with Path- ways,

    A. Chowdhery et al., “PaLM: Scaling Language Modeling with Path- ways,” Journal of Machine Learning Research (JMLR), vol. 24, no. 246, pp. 1-113, 2023. [Online]. Available: http://jmlr.org/papers/v24/22-1144. html

    2023

  30. [36]

    TPU v4: An Optically Reconfigurable Super- computer for Machine Learning with Hardware Support for Spar- sity,

    N. P. Jouppi et al., “TPU v4: An Optically Reconfigurable Super- computer for Machine Learning with Hardware Support for Spar- sity,” Proceedings of the 49th Annual International Symposium on Computer Architecture (ISCA ’22) , pp. 174–188, Jun. 2022. DOI: 10.1145/3470496.3527403

    2022

  31. [37]

    Scaling Deep Learning Workloads Beyond the Limits of Today’s Hardware,

    Cerebras Systems, “Scaling Deep Learning Workloads Beyond the Limits of Today’s Hardware,” Cerebras White Paper, Aug. 2021. [On- line]. Available: https://www.cerebras.net/wp-content/uploads/2021/08/ Scaling Deep Learning Workloads White Paper.pdf

    2021

  32. [38]

    EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,

    M. Tan and Q. V . Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” Proceedings of the 36th Interna- tional Conference on Machine Learning (ICML 2019) , PMLR vol. 97, pp. 6105–6114, Jun. 2019. [Online]. Available: https://proceedings.mlr. press/v9...

    2019

  33. [39]

    Tesla AI Day 2022,

    Tesla, “Tesla AI Day 2022,” Event Presentation, Sep. 30, 2022. [Online]. Available: https://www.tesla.com/AI

    2022

  34. [40]

    A million spiking-neuron integrated circuit with a scalable communication network and interface,

    P. A. Merolla et al., “A million spiking-neuron integrated circuit with a scalable communication network and interface,” Science, vol. 345, no. 6197, pp. 668–673, Aug. 2014. DOI: 10.1126/science.1254642

    2014 doi

  35. [41]

    Apple unveils iPhone 14 Pro and iPhone 14 Pro Max,

    Apple, “Apple unveils iPhone 14 Pro and iPhone 14 Pro Max,” Apple Newsroom , Sep. 7, 2022. [Online]. Available: https://www.apple.com/newsroom/2022/09/ apple-unveils-iphone-14-pro-and-iphone-14-pro-max/

    2022

  36. [42]

    Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity,

    W. Fedus, B. Zoph, and N. Shazeer, “Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity,” Journal of Machine Learning Research (JMLR), vol. 23, no. 120, pp. 1–39, 2022. [Online]. Available: http://jmlr.org/papers/v23/21-0998.html

    2022

  37. [43]

    Introducing Claude 2.1,

    Anthropic, “Introducing Claude 2.1,” Anthropic Blog , Nov. 21, 2023. [Online]. Available: https://www.anthropic.com/index/claude-2-1

    2023

  38. [44]

    Introducing Gemini: our largest and most capable AI model,

    D. Hassabis and E. Sheleff, “Introducing Gemini: our largest and most capable AI model,” Google DeepMind Blog , Dec. 6, 2023. [Online]. Available: https://deepmind.google/technologies/gemini/

    2023

  39. [45]

    Gemini 1.5: Our next-generation model, break- through long context,

    Google DeepMind, “Gemini 1.5: Our next-generation model, break- through long context,” Google DeepMind Blog, Feb. 15, 2024. [Online]. Available: https://deepmind.google/technologies/gemini/gemini-1-5/

    2024

  40. [46]

    Lost in the Middle: How Language Models Use Long Contexts,

    N. F. Liu, K. Lin, J. Hewitt, A. Singh, P. Liang, and P. S. H. E. Chen, “Lost in the Middle: How Language Models Use Long Contexts,” arXiv:2307.03172, Jul. 2023. [Online]. Available: https://arxiv.org/abs/ 2307.03172

    2023 arXiv

  41. [47]

    WebGPT: Browser-assisted question-answering with human feedback,

    R. Nakano et al., “WebGPT: Browser-assisted question-answering with human feedback,” arXiv:2112.09332, Dec. 2021. [Online]. Available: https://arxiv.org/abs/2112.09332

    2021 arXiv

  42. [48]

    Auto-GPT: An Autonomous GPT-4 Experiment,

    T. Grantham, “Auto-GPT: An Autonomous GPT-4 Experiment,” GitHub Repository , Mar. 2023. [Online]. Available: https://github.com/ Significant-Gravitas/Auto-GPT

    2023

  43. [49]

    Experimental evidence on the productivity effects of generative artificial intelligence,

    S. Noy and W. J. Zhang, “Experimental evidence on the productivity effects of generative artificial intelligence,” Science, vol. 381, no. 6654, pp. 187-192, Jul. 2023. DOI: 10.1126/science.adh1806

    2023 doi

  44. [51]

    Introducing Adept: Useful General Intelligence That Enables Humans And Computers To Work Together Creatively,

    Adept AI, “Introducing Adept: Useful General Intelligence That Enables Humans And Computers To Work Together Creatively,” Adept Blog ,

  45. [52]

    Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification,

    J. Buolamwini and T. Gebru, “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification,” Proceedings of Ma- chine Learning Research, vol. 81, pp. 77–91, 2018. [Online]. Available: http://proceedings.mlr.press/v81/buolamwini18a.html

    2018

  46. [53]

    The future of employment: How susceptible are jobs to computerisation?

    C. B. Frey and M. A. Osborne, “The future of employment: How susceptible are jobs to computerisation?” Technological Forecast- ing and Social Change , vol. 114, pp. 254–280, Jan. 2017. DOI: 10.1016/j.techfore.2016.08.019

    2017 doi

  47. [54]

    Killer Robots,

    R. Sparrow, “Killer Robots,” Journal of Applied Philosophy, vol. 24, no. 1, pp. 62–77, Feb. 2007. DOI: 10.1111/j.1468-5930.2007.00346.x

    2007

  48. [56]

    Neural Architecture Search with Reinforcement Learning,

    B. Zoph and Q. V . Le, “Neural Architecture Search with Reinforcement Learning,” Proceedings of the 5th International Conference on Learning Representations (ICLR 2017) , Apr. 2017. [Online]. Available: https:// arxiv.org/abs/1611.01578

    2017 arXiv

  49. [2020]

    Available: https://papers.nips.cc/paper/2020/hash/ 6b493230205f780e1bc26945df7481e5-Abstract.html

    [Online]. Available: https://papers.nips.cc/paper/2020/hash/ 6b493230205f780e1bc26945df7481e5-Abstract.html

    2020

  50. [2022]

    Available: https://www.adept.ai/blog/introducing-adept

    [Online]. Available: https://www.adept.ai/blog/introducing-adept

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.