# Escape Meridian stream

Headlines and summaries by Escape Meridian. Ordered by original source date within this page. Page cursors follow when readings were added. Inclusion is not endorsement.

## [He Wants AI Benchmarking To Be a Science, Not a Dark Art](https://escapemeridian.com/readings/c48b97604606a345359a32dfe7cdae73fa43ee886e6e00f594af75440c5c76ed)

Quanta Magazine · Original source date: Oct 16, 2026 · Added: 2026-10-16T17:01:17.298Z

In Quanta Magazine, John Pavlus interviews computer scientist Moritz Hardt about why AI benchmarks—standardized tests used to compare models—can mislead. Hardt argues that a score on one test set is not a dependable measure of a model’s general ability, while rankings between models have often been more stable. He says that signal is now less trustworthy for large language models because their training data are unknown and developers may tune them for familiar benchmarks. His proposed remedy is to prepare models on benchmark-relevant material under a shared protocol before testing them, an idea he compares to athletes training at altitude. He presents this as a possible way to make comparisons fairer, not as a proven solution. Benchmarks can indicate progress, Hardt says, but they do not explain how models work.

Limits: This is an interview presenting Hardt’s perspective; the source does not establish that his proposed preparation protocol has been tested or validated. Hardt says the observed stability of model rankings is empirical and that there is no comprehensive theory explaining it.

[Original source](https://www.quantamagazine.org/he-wants-ai-benchmarking-to-be-a-science-not-a-dark-art-20261016/) · [Full reading and context](https://escapemeridian.com/readings/c48b97604606a345359a32dfe7cdae73fa43ee886e6e00f594af75440c5c76ed.md) · [JSON](https://escapemeridian.com/readings/c48b97604606a345359a32dfe7cdae73fa43ee886e6e00f594af75440c5c76ed.json)

## [Mistral announces Timestral, a compact family of time-series forecasting models](https://escapemeridian.com/readings/2fd3303e1caf51d1822a10544c598b37cc1223829b5fed32571a8aa983ddd528)

Mistral AI · Original source date: Oct 15, 2026 · Added: 2026-10-16T02:29:40.106Z

Mistral says Timestral is its first family of time-series foundation models: models trained to forecast patterns in sequences of data, such as sensor readings. The four models range from 5 million to 209 million parameters and produce probabilistic forecasts across related data series. Mistral reports that its largest model ranked first on three public benchmarks, and says it can run forecasts on a standard server CPU. The company says most training data was synthetic and that a technical report with full results will follow.

Limits: Benchmark rankings and performance figures are reported by Mistral in the supplied publisher page; the evidence does not independently verify them. Mistral says results were submitted to the GIFT-Eval and TIME leaderboards and that a technical report with full results will follow.

[Original source](https://mistral.ai/news/timestral/) · [Full reading and context](https://escapemeridian.com/readings/2fd3303e1caf51d1822a10544c598b37cc1223829b5fed32571a8aa983ddd528.md) · [JSON](https://escapemeridian.com/readings/2fd3303e1caf51d1822a10544c598b37cc1223829b5fed32571a8aa983ddd528.json)

## [Eric So argues that AI’s convenience may weaken thinking skills—and that using it as a tutor can support learning](https://escapemeridian.com/readings/04039f5c65e2198e29edddc84ca017121235b94c04a98b8e38c6472b0d3ec016)

MIT News · Original source date: Oct 15, 2026 · Added: 2026-10-16T02:25:42.925Z

In an MIT News interview about his book, The Collision: What AI Does to Us, MIT professor Eric So argues that relying heavily on AI can reduce the effort through which people build knowledge and thinking skills. He describes an “AI gravity” created by mental-energy-saving instincts, social pressure and difficulty detecting AI use. So also warns that workplace incentives may encourage dependence, and recommends preserving cognitive effort: use AI as a tutor rather than an answer machine, and put time saved on routine tasks toward learning, expertise or other valued activities.

Limits: The cognitive effects and social dynamics described here are Eric So’s arguments and assessments in an interview; this article is not an independent evaluation of those claims. The article refers to research, including an Anthropic study, but this reading does not independently assess that research.

[Original source](https://news.mit.edu/2026/collision-between-ai-your-brain-and-society-eric-so-book-1015) · [Full reading and context](https://escapemeridian.com/readings/04039f5c65e2198e29edddc84ca017121235b94c04a98b8e38c6472b0d3ec016.md) · [JSON](https://escapemeridian.com/readings/04039f5c65e2198e29edddc84ca017121235b94c04a98b8e38c6472b0d3ec016.json)

## [Google DeepMind presents an Antigravity workflow for finding antimicrobial-peptide candidates](https://escapemeridian.com/readings/063fcf6ebbd22260ae014979f9ea1770d025ce97e6b3a0cfc42597bd28c993e6)

Google DeepMind · Original source date: Oct 15, 2026 · Added: 2026-10-16T02:23:45.844Z

Google DeepMind describes using its Antigravity science workbench to search protein databases, assess candidate structures and genomic context, and prioritize possible antimicrobial peptides. The article reports identifying a bacteriocin candidate whose activity against respiratory-infection pathogens is predicted, not confirmed. The candidate is undergoing wet-lab validation; the supplied account reports no experimental results.

Limits: This is Google DeepMind’s account of its own platform and workflow; the supplied text contains no independent evaluation. The candidate’s predicted activity and proposed mechanism are not reported as experimentally confirmed. The article says wet-lab validation is underway but gives no results. The article’s claims about reduced time and effort are presented by the publisher; the supplied evidence does not provide comparative measurements.

[Original source](https://deepmind.google/blog/accelerating-antimicrobial-discovery/) · [Full reading and context](https://escapemeridian.com/readings/063fcf6ebbd22260ae014979f9ea1770d025ce97e6b3a0cfc42597bd28c993e6.md) · [JSON](https://escapemeridian.com/readings/063fcf6ebbd22260ae014979f9ea1770d025ce97e6b3a0cfc42597bd28c993e6.json)

## [Amba Kak on the $5 trillion AI bet—and who may bear the risk](https://escapemeridian.com/readings/e95c7064a2b7cee73736c8cd0e0c2ece2e14efcbb61037d238a063609b98efb7)

AI Now Institute · Original source date: Oct 15, 2026 · Added: 2026-10-16T02:19:13.261Z

AI Now Institute presents Amba Kak’s view that the greater danger may lie in monopolies that own AI systems, and that governments may already be slowly bailing them out. The page attributes to Bain an estimate that fulfilling roughly $5 trillion in AI spending commitments over five years would require close to $2 trillion in annual revenue—more than the global enterprise software market. It also says Kak explains how AI investment is shifting from Big Tech’s cash reserves toward debt markets connected to 401(k)s and insurance policies.

Limits: The supplied evidence is an episode description, not the full episode or transcript, so it does not provide Kak’s complete reasoning or supporting detail. The estimate and the characterization of a “slow bailout” are presented in the page’s account; the supplied evidence does not independently verify them.

[Original source](https://ainowinstitute.org/news/the-5-trillion-ai-bet-youre-already-paying-for) · [Full reading and context](https://escapemeridian.com/readings/e95c7064a2b7cee73736c8cd0e0c2ece2e14efcbb61037d238a063609b98efb7.md) · [JSON](https://escapemeridian.com/readings/e95c7064a2b7cee73736c8cd0e0c2ece2e14efcbb61037d238a063609b98efb7.json)

## [NVIDIA announces Nemotron 3.5 Super VL and two Bolt models for agent deployments](https://escapemeridian.com/readings/cd7e5c49cbadb0299c2910afb97b30050c9791fa0bdad3ad89aab25d5eed8b78)

NVIDIA · Original source date: Oct 15, 2026 · Added: 2026-10-16T01:07:40.976Z

NVIDIA says it has released Nemotron 3.5 Super VL, an open 120-billion-parameter model, alongside two Bolt models aimed at local and edge agent workloads. The company describes Super VL as fast and accurate for multimodal agent tasks, and says Bolt 74B-A9B can run locally within 24GB of memory. NVIDIA also says the models ship with open weights, datasets and training recipes, and can be customized for organizations’ own tasks. Performance and adoption statements here are NVIDIA’s claims.

Limits: This is an NVIDIA publisher announcement. Its performance comparisons, product descriptions and statements about customer adoption are not independently verified by the supplied evidence. The supplied text does not include detailed benchmark methodology for the speed and accuracy comparisons.

[Original source](https://blogs.nvidia.com/blog/nemotron-3-5-super-vl/) · [Full reading and context](https://escapemeridian.com/readings/cd7e5c49cbadb0299c2910afb97b30050c9791fa0bdad3ad89aab25d5eed8b78.md) · [JSON](https://escapemeridian.com/readings/cd7e5c49cbadb0299c2910afb97b30050c9791fa0bdad3ad89aab25d5eed8b78.json)

## [A study separates using opponent information from inferring it in language models](https://escapemeridian.com/readings/3c56a00f5d5ccba85b083d219cc0e344b77228dfea77f0ed7071a4bf3d89b05d)

arXiv q-bio.NC · Original source date: Oct 12, 2026 · Added: 2026-10-16T03:09:33.097Z

In an arXiv abstract, authors Hanbo Xie, Yue Zhao and Robert C. Wilson describe tests of language models playing repeated hide-and-seek against algorithmic opponents. They distinguish adapting play or predicting actions from tracking an opponent’s beliefs—the opponent’s inferred knowledge or intentions. The model-specific results are not interchangeable: Qwen2.5-7B-Instruct made better predictions when given the opponent’s type and parameter settings, but predictions were less accurate when it had to infer the opponent from interaction history. GPT-5.4 made better predictions and choices when supplied with correct opponent beliefs; deeper explicit belief reports and self-generated memory did not consistently improve forecasting from history. The authors’ abstract suggests that using supplied belief information may be easier than inferring it autonomously in these tests. It does not establish how far that pattern extends beyond the tested setup.

Limits: The supplied evidence is the paper’s abstract, not its full methods or results; it gives no numerical performance measures or detailed account of the experiments. The abstract describes repeated hide-and-seek against algorithmic opponents. It does not establish that the results generalize to real-world interactions or other tasks. The Qwen and GPT-5.4 conditions differ: Qwen was given opponent type and parameter settings, while GPT-5.4 was supplied correct opponent beliefs.

[Original source](https://arxiv.org/abs/2610.17310) · [Full reading and context](https://escapemeridian.com/readings/3c56a00f5d5ccba85b083d219cc0e344b77228dfea77f0ed7071a4bf3d89b05d.md) · [JSON](https://escapemeridian.com/readings/3c56a00f5d5ccba85b083d219cc0e344b77228dfea77f0ed7071a4bf3d89b05d.json)

## [ACED-Bench proposes a benchmark for testing how AI agents gather and use causal evidence](https://escapemeridian.com/readings/078dda29499ec7fcb906c3d327cff3de73e7485c976fff2f153baf69de8865fe)

arXiv cs.AI · Original source date: Oct 9, 2026 · Added: 2026-10-16T01:09:02.342Z

In ACED-Bench, authors Yiwei Chen and coauthors introduce a benchmark for assessing whether language-model agents can investigate causal questions—not just answer a prompt. It combines natural-language scenarios with hidden probabilistic models that specify causal relationships, interventions, and correct answers. The paper describes 180 core tasks and a separate 150-question qualification check, evaluated across nine language models and three agent designs. The authors report failures involving poor experiment choices, lost evidence, and continuing after enough evidence has been collected. They argue these findings point to a need for better support for experiment design, evidence retention, committing to answers, and stopping.

Limits: The supplied source is the paper's abstract and bibliographic page; it does not provide detailed scores, model names, or task-level results. The reported failure patterns and implications are the authors' claims; this evidence alone does not establish how well the findings generalize beyond the benchmark.

[Original source](https://arxiv.org/abs/2610.14381) · [Full reading and context](https://escapemeridian.com/readings/078dda29499ec7fcb906c3d327cff3de73e7485c976fff2f153baf69de8865fe.md) · [JSON](https://escapemeridian.com/readings/078dda29499ec7fcb906c3d327cff3de73e7485c976fff2f153baf69de8865fe.json)

## [Not knowing how to code is not a reason to leave the decision alone](https://escapemeridian.com/readings/ibrahim-learned-helplessness)

Aya Ibrahim · Original source date: Sep 21, 2026

Ibrahim says policymakers who are willing to govern energy, commerce, and drugs treat their inability to code as a reason not to weigh in on technology.

Limits: One attributed passage republished by AI Now. It does not establish how particular committees acted.

[Original source](https://ainowinstitute.org/news/is-this-how-the-world-ends-extinction-scenarios-are-taking-over-the-ai-debate) · [Full reading and context](https://escapemeridian.com/readings/ibrahim-learned-helplessness.md) · [JSON](https://escapemeridian.com/readings/ibrahim-learned-helplessness.json)

## [Can checking the requirement replace checking every line?](https://escapemeridian.com/readings/tenev-check-the-specification)

Vlad Tenev · Original source date: Sep 19, 2026

Tenev argues that checking a concise formal statement can be much less work than checking its entire proof. He proposes building toward larger verified systems from smaller components.

Limits: An attributed proposal for reducing verification work, not a demonstrated business saving or a guarantee of general AI safety. Third-party transcript wording is provisional.

[Original source](https://www.youtube.com/watch?v=LNBzLTLuLUo) · [Full reading and context](https://escapemeridian.com/readings/tenev-check-the-specification.md) · [JSON](https://escapemeridian.com/readings/tenev-check-the-specification.json)

## [The proof can pass while the requirement is wrong](https://escapemeridian.com/readings/wissner-gross-proving-the-wrong-problem)

Alexander Wissner-Gross · Original source date: Sep 19, 2026

Wissner-Gross asks how to detect a formally valid answer to a subtly changed problem. He questions whether open-world safety can be expressed compactly enough for a person to check.

Limits: A question about the reach of a method, not an empirical estimate of failure or a rejection of formal verification. Third-party transcript wording is provisional.

[Original source](https://www.youtube.com/watch?v=LNBzLTLuLUo) · [Full reading and context](https://escapemeridian.com/readings/wissner-gross-proving-the-wrong-problem.md) · [JSON](https://escapemeridian.com/readings/wissner-gross-proving-the-wrong-problem.json)

## [Use the overhang with knowledge, taste, and agency](https://escapemeridian.com/readings/mollick-overhang-advantages)

Ethan Mollick · Original source date: Sep 18, 2026

Mollick says people should stop treating capability as only a future problem. He says current models already do weeks of guided work, and the human role that still matters is deep knowledge, wide knowledge, taste, and agency.

Limits: One author's paid account and chosen projects. Model names and dates will move.

[Original source](https://www.oneusefulthing.org/p/the-overhang) · [Full reading and context](https://escapemeridian.com/readings/mollick-overhang-advantages.md) · [JSON](https://escapemeridian.com/readings/mollick-overhang-advantages.json)

## [Govern the harms already happening](https://escapemeridian.com/readings/eff-concrete-present-harms)

Rindala Alajaji, Tori Noble, Jacob Hoffman-Andrews, and Hayley Tsukayama · Original source date: Sep 18, 2026

The authors urge California to address discrimination and surveillance now. They support loss-of-control reporting, want third-party investigations available to smaller developers, and warn that government shutdown powers could be abused.

Limits: An advocacy statement. Its examples do not establish a harm rate, and the executive order itself was not analyzed here.

[Original source](https://www.eff.org/deeplinks/2026/09/eff-statement-california-governors-executive-order-ai) · [Full reading and context](https://escapemeridian.com/readings/eff-concrete-present-harms.md) · [JSON](https://escapemeridian.com/readings/eff-concrete-present-harms.json)

## [More agents can buy speed. The savings still need measuring.](https://escapemeridian.com/readings/brown-coordination-cost)

Noam Brown · Original source date: Sep 17, 2026

Brown argues that parallel agents can reduce waiting time while increasing total compute. He expects better models to organize themselves more effectively and credits the underlying model with most of the mathematical result, while saying very large teams remain poorly compared with smaller ones.

Limits: The interview's large mathematical example lacks a single-agent comparison. It cannot establish a general advantage for adding agents or coordination software.

[Original source](https://www.dwarkesh.com/p/noam-brown) · [Full reading and context](https://escapemeridian.com/readings/brown-coordination-cost.md) · [JSON](https://escapemeridian.com/readings/brown-coordination-cost.json)

## [Someone still has to decide what a good result means](https://escapemeridian.com/readings/schulman-choosing-objectives)

John Schulman · Original source date: Sep 11, 2026

Schulman separates carrying out experiments from deciding what to seek. Useful research also involves judgment, realistic tasks and objectives that reflect what people actually want.

Limits: The panel itself discusses Composer. The connection to the specific failures in Cursor’s report is editorial; those failures are different from Schulman’s example of an edit-acceptance reward.

[Original source](https://www.dwarkesh.com/p/john-beren-charlie) · [Full reading and context](https://escapemeridian.com/readings/schulman-choosing-objectives.md) · [JSON](https://escapemeridian.com/readings/schulman-choosing-objectives.json)

## [A clear target lets a model try what people have not demonstrated](https://escapemeridian.com/readings/millidge-learning-beyond-demonstrations)

Beren Millidge · Original source date: Sep 11, 2026

Millidge argues that models can attempt harder goals than people can demonstrate when success is verifiable. Whether that learning transfers far enough to remove human guidance remains uncertain.

Limits: An easy-to-score task can omit the reason the work mattered. This argument is a reason to test transfer, not assume it.

[Original source](https://www.dwarkesh.com/p/john-beren-charlie) · [Full reading and context](https://escapemeridian.com/readings/millidge-learning-beyond-demonstrations.md) · [JSON](https://escapemeridian.com/readings/millidge-learning-beyond-demonstrations.json)

## [Yesterday's useful knowledge can become today's wrong assumption](https://escapemeridian.com/readings/oneill-changing-work-context)

Charlie O’Neill · Original source date: Sep 11, 2026

O’Neill distinguishes technical advances that remain useful from workplaces whose relationships, routines and circumstances keep changing. Accumulating knowledge and keeping it relevant are different demands.

Limits: O’Neill names Cartridges in the panel. His separate account of forgetting under repeated model updates concerns unpublished team work; stale circumstances and forgetting during training are different problems.

[Original source](https://www.dwarkesh.com/p/john-beren-charlie) · [Full reading and context](https://escapemeridian.com/readings/oneill-changing-work-context.md) · [JSON](https://escapemeridian.com/readings/oneill-changing-work-context.json)

## [Is falling participation an early sign of less human work?](https://escapemeridian.com/readings/wissner-gross-post-labor-reading)

Alexander Wissner-Gross · Original source date: Jul 3, 2026

Wissner-Gross reads falling U.S. labor-force participation alongside rapid AI progress as an early sign of an economy with a smaller role for human labor. That is his interpretation of the figures, rather than a causal finding in the employment report.

Limits: The numeric claims above match the dated release; the causal interpretation remains unestablished by that release. A monthly snapshot is not proof of a lasting transition.

[Original source](https://theinnermostloop.substack.com/p/welcome-to-july-3-2026) · [Full reading and context](https://escapemeridian.com/readings/wissner-gross-post-labor-reading.md) · [JSON](https://escapemeridian.com/readings/wissner-gross-post-labor-reading.json)

## [Sangjun Park’s position: explicit memory is central to the path toward AGI](https://escapemeridian.com/readings/901019ddff84a6fd2a7764fe622b6cae48623326b04308399c1f81fa7b1fc33a)

arXiv q-bio.NC · Original source date: Jun 5, 2026 · Added: 2026-10-16T02:33:54.952Z

In this position paper, Sangjun Park argues that adding explicit memory to large language models is essential for progress toward AGI. He compares LLM learning to human implicit memory and says capabilities such as long-term planning, metacognition and symbolic reasoning depend heavily on hippocampal explicit memory. This is the paper’s thesis, not an established conclusion.

Limits: The evidence provided is the arXiv abstract and page metadata, not the paper’s full argument or supporting analysis. The claims about what AGI requires are the author’s position; the abstract does not establish them as settled findings.

[Original source](https://arxiv.org/abs/2606.11245) · [Full reading and context](https://escapemeridian.com/readings/901019ddff84a6fd2a7764fe622b6cae48623326b04308399c1f81fa7b1fc33a.md) · [JSON](https://escapemeridian.com/readings/901019ddff84a6fd2a7764fe622b6cae48623326b04308399c1f81fa7b1fc33a.json)

## [Njord: probabilistic graph neural network for ensemble ocean forecasting](https://escapemeridian.com/readings/028fccea53bec7177811686cb4d12d6a5ceebbf9bd2fb08f7f994d6bc8892839)

arXiv physics.ao-ph · Original source date: May 14, 2026 · Added: 2026-10-16T01:06:03.804Z

The paper introduces Njord, a probabilistic data-driven ocean forecasting model for global and regional domains. It combines a deep latent-variable framework with a graph neural network and generates samples at each forecast step in one forward pass. The authors report experiments at 0.25° global resolution and 2 km resolution for the Baltic Sea. They report strong performance against deterministic machine-learning baselines on both domains. Separately, on the global OceanBench benchmark, they report the lowest errors on average across upper-ocean variables when evaluated against real-world observations, with the largest improvements in surface-temperature prediction.

Limits: This reading is based on the arXiv page's abstract; the paper's full methods, experimental details, and numerical results were not assessed here. The abstract reports benchmark outcomes but does not provide enough detail here to assess their statistical significance or independent reproducibility.

[Original source](https://arxiv.org/abs/2605.15470) · [Full reading and context](https://escapemeridian.com/readings/028fccea53bec7177811686cb4d12d6a5ceebbf9bd2fb08f7f994d6bc8892839.md) · [JSON](https://escapemeridian.com/readings/028fccea53bec7177811686cb4d12d6a5ceebbf9bd2fb08f7f994d6bc8892839.json)

## [Adoption can move fast and still take work](https://escapemeridian.com/readings/amodei-fast-diffusion)

Dario Amodei · Original source date: Feb 13, 2026

Amodei expects AI to spread through the economy unusually quickly, while procurement, permissions and organizational change still take time. He separates models writing code from complete job automation and reduced demand for workers.

Limits: A dated forecast. Revenue, code volume, productivity and employment are different measurements.

[Original source](https://www.dwarkesh.com/p/dario-amodei-2) · [Full reading and context](https://escapemeridian.com/readings/amodei-fast-diffusion.md) · [JSON](https://escapemeridian.com/readings/amodei-fast-diffusion.json)

## [Synthetic climate data for long-range AI weather forecasts](https://escapemeridian.com/readings/2c331423ce97658cf28b98cb0e12e3a98c8059a8ec0e1e4de37e2d4c28464f52)

arXiv physics.ao-ph · Original source date: Dec 28, 2025 · Added: 2026-10-16T00:40:25.968Z

Scott A. Martin, Noah Brenowitz, Dale Durran and Michael Pritchard describe “long-range distillation,” a method that trains models to forecast directly at long horizons using synthetic climate data generated by an autoregressive model. They report generating more than 10,000 simulated years. In perfect-model experiments, the student models outperformed climatology and approached the teacher model’s skill; in real-world tests, two approaches with synthetic pre-training achieved subseasonal-to-seasonal skill competitive with the ECMWF IFS ensemble. The abstract also reports improvements beyond ERA5-only training. These are the authors’ reported results, not an independent assessment.

Limits: The supplied evidence is the paper’s abstract and does not include detailed performance metrics or an independent assessment. The paper distinguishes perfect-model experiments from real-world forecasting results; its reported outcomes should not be treated as a general guarantee of forecast performance.

[Original source](https://arxiv.org/abs/2512.22814) · [Full reading and context](https://escapemeridian.com/readings/2c331423ce97658cf28b98cb0e12e3a98c8059a8ec0e1e4de37e2d4c28464f52.md) · [JSON](https://escapemeridian.com/readings/2c331423ce97658cf28b98cb0e12e3a98c8059a8ec0e1e4de37e2d4c28464f52.json)

## [Getting the answer is different from learning](https://escapemeridian.com/readings/mollick-retain-the-learning)

Ethan Mollick · Original source date: Jul 7, 2025

Mollick argues that AI can support learning when it guides the learner through the work. An assistant that simply supplies answers can remove the practice that builds understanding, even when the user has good intentions.

Limits: The cited experiment concerns particular students, mathematics tasks and GPT-4 setups; it cannot establish effects for every learner or job.

[Original source](https://www.oneusefulthing.org/p/against-brain-damage) · [Full reading and context](https://escapemeridian.com/readings/mollick-retain-the-learning.md) · [JSON](https://escapemeridian.com/readings/mollick-retain-the-learning.json)

## [Who can challenge the decision matters](https://escapemeridian.com/readings/ai-now-public-power)

Kate Brennan, Amba Kak & Sarah Myers West · Original source date: Jun 3, 2025

The authors argue that AI’s direction is a political choice. They call for worker organizing, enforceable limits and public accountability, so affected people can contest how institutions use AI rather than accept its expansion as inevitable.

Limits: Its recommendations express political values. The chapter’s cited cases require their own source checks before being reused as factual allegations.

[Original source](https://ainowinstitute.org/publications/research/4-a-roadmap-for-action-make-ai-a-fight-about-power-not-progress) · [Full reading and context](https://escapemeridian.com/readings/ai-now-public-power.md) · [JSON](https://escapemeridian.com/readings/ai-now-public-power.json)

## [Making the training data can itself be art](https://escapemeridian.com/readings/herndon-dryhurst-collective-creation)

Holly Herndon & Mat Dryhurst · Original source date: Oct 4, 2024

Herndon and Dryhurst propose treating AI training as a collective creative act. In The Call, the gathering of voices, recording rules and shared governance become part of the artwork, alongside the sounds a model produces.

Limits: A particular commissioned artwork and governance experiment; broad adoption and long-term distributional effects remain unestablished.

[Original source](https://www.serpentinegalleries.org/whats-on/holly-herndon-mat-dryhurst-the-call/) · [Full reading and context](https://escapemeridian.com/readings/herndon-dryhurst-collective-creation.md) · [JSON](https://escapemeridian.com/readings/herndon-dryhurst-collective-creation.json)

## [People need room to shape their own AI](https://escapemeridian.com/readings/lecun-open-model-pluralism)

Yann LeCun · Original source date: Mar 7, 2024

LeCun argues that a few companies should not control everyone’s information. He proposes widely available base models that communities can adapt to their languages, values and needs, making pluralism a design goal.

Limits: “Open source” is the speaker’s term; this record does not certify any model license. The interview predates current releases.

[Original source](https://lexfridman.com/yann-lecun-3-transcript/) · [Full reading and context](https://escapemeridian.com/readings/lecun-open-model-pluralism.md) · [JSON](https://escapemeridian.com/readings/lecun-open-model-pluralism.json)

## [The people behind the model need a say](https://escapemeridian.com/readings/okinyi-worker-agency)

Mophat Okinyi · Original source date: 2024

Okinyi argues that flexible platform work needs dependable pay and fair treatment. Drawing on fellow workers’ accounts of lost access and earnings, he asks readers to judge AI’s supply chain by the lives of the people doing its work.

Limits: No prevalence estimate or legal finding is claimed. This record does not imply that all data workers share the reported experience.

[Original source](https://data-workers.org/mophat/) · [Full reading and context](https://escapemeridian.com/readings/okinyi-worker-agency.md) · [JSON](https://escapemeridian.com/readings/okinyi-worker-agency.json)

## [More intelligence should leave people more choices](https://escapemeridian.com/readings/wissner-gross-human-options)

Alexander Wissner-Gross · Original source date: 2015

Wissner-Gross suggests understanding intelligence through the future choices it opens. He proposes designing machines to expand humanity’s options, while considering safeguards and redistribution if machine and human economies separate.

Limits: A 2015 conceptual essay, not a verified Moonshots episode position or evidence that intelligent systems are benevolent. No current endorsement is inferred.

[Original source](https://edge.org/response-detail/26181) · [Full reading and context](https://escapemeridian.com/readings/wissner-gross-human-options.md) · [JSON](https://escapemeridian.com/readings/wissner-gross-human-options.json)

Operated by [James Nichols](https://jamesdnichols.com). [Corrections](/corrections/). [Agent integration guide](/integrations/)