# Bright — See what AI is making possible

AI is doing real work across an astonishing range of the world — from the rain over a subcontinent to the shape of a single protein. Fall through the scales, try the work, and check the evidence behind every claim.

Canonical: https://brightaifuture.com/
Format: page
Source publication: Not established
Bright publication: Not established
Substantive update: None recorded

## Bright at every scale

AI is doing real work across an astonishing range of the world — from the rain over a subcontinent to the shape of a single protein. 14 scales, from about 10²⁵ meters to 10⁻⁸, each with a source-checked record. They are a curated edition; every record is in the Index.

## 10²⁵ m — a ripple from space

10²⁵ m · a ripple from space · contextual scale (how far the ripple traveled — the AI work itself happened at LIGO’s 4-km detector).

Two black holes merged about 1.3 billion years ago. On 14 September 2015, the ripple reached Earth.

Evidence: Demonstrated — Quieting the mirrors that listen to space Limit: The projected benefit of hundreds more events per year depends on applying the method across all LIGO mirror-control loops.

## 12,700 km — the whole atmosphere

12,700 km · the whole atmosphere · size of the subject.

0.3% of the computing its own supercomputer forecast needs. That is what NOAA reports for a single 16-day global forecast from its AI model, which finishes in about 40 minutes.

Evidence: Deployed — A world-class forecast now fits on one machine. Limit: Every one of these models is trained on a reanalysis built from the observing network. No model can supply a measurement that was never made.

Evidence: Deployed — From research to the forecast desk. Limit: Deployment is not a guarantee of superiority for every variable, place, or event.

## 2,000 km — a subcontinent’s rains

2,000 km · a subcontinent’s rains · size of the setting.

38,845,214 farmers across India received an AI-based forecast of the monsoon’s arrival — by text message, in five regional languages.

Evidence: Deployed — More time to decide when to plant. Limit: The reported decision changes are self-reported survey results from Bihar and Madhya Pradesh, not measured yield or income impacts.

## 1 km — a computing campus

1 km · a computing campus · size of the setting.

0.7% of Google’s worldwide computing, recovered continuously for more than a year — by a scheduling rule an AI system discovered.

Evidence: Deployed (company-reported) — Making room in a data center Limit: This is a company-reported, Google-specific operational result.

## 100 m — a field

100 m · a field · size of the setting.

One plant at a time.

Evidence: Deployed — Lasers matched the herbicides. Limit: The peer-reviewed trial used an autonomous demonstration unit, not a machine a farm can buy, and collected no economic data. Its authors thank Carbon Robotics for donating the unit, though the study was publicly funded and declares no conflicts.

## 10 m — a classroom

10 m · a classroom · size of the setting.

AI helped students learn. Without teachers’ guardrails, it also got in the way.

Evidence: Demonstrated — A teacher and an AI tutor after school. Limit: The intervention was short and teacher-supported, so it does not establish effects of unsupervised chatbot use.

Evidence: Demonstrated — When the answer arrives too soon. Limit: The study examined short-term exam outcomes in one high-school mathematics setting.

## 1 m — two people talking

1 m · two people talking · size of the setting.



Evidence: Emerging — A voice in everyday life. Limit: One participant with an implanted investigational device. Caregivers still helped set up the system. This is not general clinical availability or independent replication.

## 30 cm — the laptop in front of you

30 cm · the laptop in front of you · size of the setting.

Some capable AI now runs on the computer in front of you — and a few come with the whole recipe.

Evidence: Deployed — Reasoning weights people can run themselves. Limit: A stated 16 GB memory requirement does not guarantee equally accessible hardware, energy use, or easy setup.

## 12 cm — a human heart

12 cm · a human heart · size of the subject.

In a double-blind randomized trial, an AI that finds where to treat an irregular heartbeat helped more patients stay free of atrial fibrillation a year later.

Evidence: Demonstrated — Personalizing an ablation for persistent AF Limit: There was no significant between-arm difference in freedom from any atrial arrhythmia after one ablation.

## 8 cm — an item in a robot’s grip

8 cm · an item in a robot’s grip · size of the setting.

A robot that knows when to stop squeezing.

Evidence: Deployed (company-reported) — A warehouse robot that knows when to stop squeezing. Limit: Amazon's capability, ergonomics, and worker-development statements are company-reported; it provides no independent injury, error, pace, or job-quality results.

## 3 cm — a carved letter

3 cm · a carved letter · size of the subject.

Historians working with an AI restored damaged Latin inscriptions better than either could alone.

Evidence: Demonstrated — Finding a missing line's ancient neighbors. Limit: A generated restoration or attribution is a hypothesis, not recovered ground truth.

## 12 mm — what imaging finds

12 mm · what imaging finds · size of the subject.

12 mm is the median size of a breast cancer found by imaging. Found by a hand — the woman’s own, or a clinician’s — the median is 21 mm. That is most of a stage.

Evidence: Emerging (company-reported deployment) — Twelve millimetres, or three centimetres. Limit: The screening numbers and any account of lives saved come from the participating companies. No independent evaluation, audit or named patient exists, and early detection is not the same as a life saved.

## 10 µm — a bacterial micro-colony

10 µm · a bacterial micro-colony · size of the subject.

Finding bacteria among the crumbs.

Evidence: Experimental — Finding bacteria among the crumbs. Limit: This is laboratory validation with selected bacteria and food matrices, not evidence of regulatory or commercial deployment.

## 10 nm — one protein

10 nm · one protein · size of the subject.

p53 helps stop damaged cells from becoming cancer. AlphaFold predicted its shape. Here, light is the model’s confidence.

Evidence: Demonstrated (a prediction) — Seeing the shape of life. Limit: Predictions carry uncertainty. A predicted structure does not establish function or prove a treatment works.

## Limitations

Scale placements are editorial, order-of-magnitude placements, not measurements. Records with no honest physical size are off the axis.

Several stations show context photographs or Bright illustrations rather than the work itself; each is labeled.

## Provenance and history

{}

## Original sources

- [Using AI to perceive the universe in greater depth](https://deepmind.google/blog/using-ai-to-perceive-the-universe-in-greater-depth/)
- [NOAA deploys new generation of AI-driven global weather models](https://www.noaa.gov/news-release/noaa-deploys-new-generation-of-ai-driven-global-weather-models)
- [ECMWF’s AI forecasts become operational](https://www.ecmwf.int/en/about/media-centre/news/2025/ecmwfs-ai-forecasts-become-operational)
- [Indian Ministry of Agriculture and Farmers Welfare parliamentary response on an AI-based monsoon-onset pilot](https://eparlib.sansad.in/bitstream/123456789/3015227/1/lsd_18_VI_02-12-2025.pdf)
- [AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms](https://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/)
- [AlphaEvolve: How our Gemini-powered coding agent is scaling impact across fields](https://deepmind.google/blog/alphaevolve-impact/)
- [Deep learning-based laser weed control compared to conventional herbicide application across three vegetable production systems · Pest Management Science](https://scijournals.onlinelibrary.wiley.com/doi/10.1002/ps.8912)
- [From Chalkboards to Chatbots: Evaluating the Impact of Generative AI on Learning Outcomes in Nigeria](https://documents.worldbank.org/en/publication/documents-reports/documentdetail/099548105192529324)
- [Generative AI without guardrails can harm learning: Evidence from high school mathematics](https://doi.org/10.1073/pnas.2422633122)
- [Long-term independent use of an intracortical brain–computer interface for speech and cursor control · Nature Medicine (PubMed)](https://pubmed.ncbi.nlm.nih.gov/42297978/)
- [An Accurate and Rapidly Calibrating Speech Neuroprosthesis · NEJM](https://www.nejm.org/doi/full/10.1056/NEJMoa2314132)
- [Introducing gpt-oss](https://openai.com/index/introducing-gpt-oss/)
- [OLMo 2 32B: First fully open model to outperform GPT 3.5 and GPT 4o mini](https://allenai.org/blog/olmo2-32b)
- [New NVIDIA Nemotron 3 Super Delivers 5x Higher Throughput for Agentic AI](https://blogs.nvidia.com/blog/nemotron-3-super-agentic-ai/)
- [Qwen3: Think Deeper, Act Faster](https://qwenlm.github.io/blog/qwen3/)
- [Artificial intelligence for individualized treatment of persistent atrial fibrillation: a randomized controlled trial](https://www.nature.com/articles/s41591-025-03517-w)
- [Introducing Vulcan: Amazon's first robot with a sense of touch](https://www.aboutamazon.com/news/operations/amazon-vulcan-robot-pick-stow-touch)
- [Contextualizing ancient texts with generative neural networks](https://www.nature.com/articles/s41586-025-09292-5)
- [FDA 510(k) clearance K252482 — CogNet AI-MT+ (MedCognetics)](https://www.accessdata.fda.gov/cdrh_docs/pdf25/K252482.pdf)
- [Deep learning enabled rapid detection of live bacteria in the presence of food debris](https://www.nature.com/articles/s41538-025-00636-z)
- [Highly accurate protein structure prediction with AlphaFold · Nature](https://www.nature.com/articles/s41586-021-03819-2)
- [Observation of Gravitational Waves from a Binary Black Hole Merger (Phys. Rev. Lett. 116, 061102)](https://arxiv.org/abs/1602.03837)
- [Gravitational Waves Detected 100 Years After Einstein’s Prediction](https://www.ligo.caltech.edu/page/press-release-gw150914)
- [LIGO facts](https://www.ligo.caltech.edu/page/facts)
- [Science summary: Observation of gravitational waves (GW150914)](https://ligo.org/science-summaries/GW150914/)
- [GW150914 tutorial and event data](https://gwosc.org/s/events/GW150914/GW150914_tutorial.html)
- [CODATA 2022 recommended values: proton rms charge radius](https://physics.nist.gov/cuu/Constants/Table/allascii.txt)
- [Government Conducts AI-Based Pilot for Local Monsoon Forecasting to Support Kharif Sowing Decisions](https://www.pib.gov.in/PressReleasePage.aspx?PRID=2241412)
- [New brain-computer interface allows man with ALS to ‘speak’ again](https://health.ucdavis.edu/news/headlines/new-brain-computer-interface-allows-man-with-als-to-speak-again/2024/08)
- [OpenAI open-weight models (gpt-oss)](https://help.openai.com/en/articles/11870455)
- [How to run gpt-oss locally with Ollama](https://developers.openai.com/cookbook/articles/gpt-oss/run-locally-ollama)
- [TP53 gene](https://medlineplus.gov/genetics/gene/tp53/)
- [Mutational landscape and significance across 12 major cancer types](https://www.nature.com/articles/nature12634)
- [Mutant p53: one name, many proteins](https://genesdev.cshlp.org/content/26/12/1268.full.pdf)
- [UniProtKB P04637 (P53_HUMAN)](https://rest.uniprot.org/uniprotkb/P04637)
- [AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models](https://academic.oup.com/nar/article/50/D1/D439/6430488)
- [Highly accurate protein structure prediction for the human proteome](https://www.nature.com/articles/s41586-021-03828-1)
- [Deep learning-based laser weed control compared to conventional herbicide application across three vegetable production systems](https://scijournals.onlinelibrary.wiley.com/doi/10.1002/ps.8912)
- [Carbon Robotics LaserWeeder case study](https://wga.s3.us-west-1.amazonaws.com/cit/2024/cit_case-study_carbon-robotics.pdf)
- [California Code of Regulations, Title 8, §3456, Hand-Held Tools (hand weeding)](https://www.dir.ca.gov/title8/3456.html)
- [The International Herbicide-Resistant Weed Database](https://www.weedscience.org/)
- [Laser Weeding Technology in Cropping Systems: A Comprehensive Review](https://www.mdpi.com/2073-4395/14/10/2253)
- [ECMWF’s AI forecasts become operational](https://www.ecmwf.int/en/about/media-centre/news/2025/ecmwfs-ai-forecasts-become-operational)
- [NOAA deploys new generation of AI-driven global weather models](https://www.noaa.gov/news-release/noaa-deploys-new-generation-of-ai-driven-global-weather-models)
- [Closing the gaps in the observing network](https://wmo.int/media/news/closing-gaps-observing-network)
- [Systematic Observations Financing Facility](https://un-soff.org/)
- [Metastatic presentation of breast cancer in India: evidence from the national cancer registry (2009–2020)](https://pmc.ncbi.nlm.nih.gov/articles/PMC13092702/)
- [Tumor size and detection in breast cancer: self-examination and clinical breast examination are at their limit](https://doi.org/10.1016/j.cdp.2008.04.002)
- [Female cancer screening in India: uptake and inequality, National Family Health Survey-5](https://pmc.ncbi.nlm.nih.gov/articles/PMC11610822/)
- [FDA 510(k) clearance K252482 — CogNet AI-MT+ (MedCognetics)](https://www.accessdata.fda.gov/cdrh_docs/pdf25/K252482.pdf)
- [Community-based cancer screening at primary care level in Varanasi district, northern India](https://pmc.ncbi.nlm.nih.gov/articles/PMC10729271/)
- [Artificial intelligence-supported screen reading versus standard double reading in the MASAI trial](https://www.thelancet.com/journals/landig/article/PIIS2589-7500(24)00267-X/fulltext)

## Continue exploring

- [Quieting the mirrors that listen to space](https://brightaifuture.com/discoveries/deep-loop-shaping-ligo)
- [A world-class forecast now fits on one machine.](https://brightaifuture.com/discoveries/nvidia-earth-2)
- [From research to the forecast desk.](https://brightaifuture.com/discoveries/aifs)
- [More time to decide when to plant.](https://brightaifuture.com/discoveries/monsoon-farmers)
- [Making room in a data center](https://brightaifuture.com/discoveries/alphaevolve-data-centers)
- [Finding more workable grid plans](https://brightaifuture.com/discoveries/alphaevolve-grid-feasibility)
- [Lasers matched the herbicides.](https://brightaifuture.com/discoveries/carbon-laserweeder)
- [A teacher and an AI tutor after school.](https://brightaifuture.com/discoveries/nigeria-ai-tutor)
- [When the answer arrives too soon.](https://brightaifuture.com/discoveries/guarded-learning)
- [A voice in everyday life.](https://brightaifuture.com/discoveries/speech-home)
- [A way back to conversation.](https://brightaifuture.com/discoveries/speech)
- [Reasoning weights people can run themselves.](https://brightaifuture.com/discoveries/gpt-oss)
- [A model whose full recipe is visible.](https://brightaifuture.com/discoveries/olmo-2-32b)
- [An open reasoning model with its recipe beside it.](https://brightaifuture.com/discoveries/nemotron-3-super)
- [Choosing when a model thinks longer.](https://brightaifuture.com/discoveries/qwen3)
- [Personalizing an ablation for persistent AF](https://brightaifuture.com/discoveries/tailored-af-ai-guided-ablation)
- [A warehouse robot that knows when to stop squeezing.](https://brightaifuture.com/discoveries/amazon-vulcan)
- [Finding a missing line's ancient neighbors.](https://brightaifuture.com/discoveries/aeneas-latin-inscriptions)
- [Twelve millimetres, or three centimetres.](https://brightaifuture.com/discoveries/medcognetics-mammography-workflow)
- [Finding bacteria among the crumbs.](https://brightaifuture.com/discoveries/food-bacteria)
- [Seeing the shape of life.](https://brightaifuture.com/discoveries/alphafold)
- [Index — every record](https://brightaifuture.com/records)
- [Now — two clocks](https://brightaifuture.com/now)
