No products in the cart.
Blog
AI's carbon footprint and a DNA nanomotor — the week in infographics – Nature.com
Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Advertisement
You have full access to this article via your institution.
The carbon footprint associated with artificial intelligence (AI) has been revealed in a study that calculated the carbon cost of training a range of models at various cloud-computing data centres. The results show that training BERT, a common machine-learning language model, at data centres in the central United States or Germany emitted 22–28 kilograms of carbon dioxide, depending on the time of year. This was more than double the emissions generated by doing the same experiment in Norway, which gets most of its electricity from hydroelectric power, or in France, which relies mostly on nuclear power. The time of day at which experiments run also matters. For example, training the AI in Washington during the night, when the state’s electricity comes from hydroelectric power alone, led to lower emissions than doing so during the day, when power also comes from gas-fired stations.
Source: Dodge, J. et al. Preprint at https://arxiv.org/abs/2206.05229 (2022).
A rethink is under way at the world’s largest laser-fusion facility. On 8 August 2021, the National Ignition Facility (NIF) at Lawrence Livermore National Laboratory in California announced a record-breaking result: it had produced, albeit for a fraction of a second, an energetic fusion reaction of the kind that powers stars and thermonuclear weapons. But attempts to replicate that result have failed, yielding at best 50% of the energy seen last year. As a result, the programme’s leadership has decided to halt replication experiments and to focus on next steps that could push the NIF well beyond the fusion threshold and into an entirely new — and more predictable — regime, where yields are significantly larger than in the August experiment.
Source: Lawrence Livermore National Laboratory
A paper in Nature reports a self-assembling nanoscale motor made from DNA. The motor makes use of the ‘origami’ technique, in which a long DNA strand is folded into a complex shape by dozens of short DNA strands acting as staples. The nanomotor consists of a pedestal, a platform and a rotor (a), all formed by the this method. These components self-assemble in solution to create the motor, which docks to a glass surface (b). Obstacles are incorporated on the edges of the platform to create a ratchet that promotes movement of the rotor in one direction. An alternating voltage applied to electrodes on either side of the motor generates an electric field that spins the rotor.
doi: https://doi.org/10.1038/d41586-022-02064-5
News
Outlook
News
News
Research Highlight
Correspondence
Article
Article
Technology Feature
Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V. (ISAS)
Dortmund, Germany
Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V. (ISAS)
Dortmund, Germany
Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V. (ISAS)
Dortmund, Germany
Oklahoma Medical Research Foundation (OMRF)
Oklahoma City, United States
You have full access to this article via your institution.
An essential round-up of science news, opinion and analysis, delivered to your inbox every weekday.
Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.
Advanced search
© 2022 Springer Nature Limited