Fix dangling figure

sustainable_ai.qmd

@@ -149,6 +149,8 @@ Without action, this exponential demand growth risks ratcheting up the carbon fo
 
 The concept of a 'carbon footprint' has emerged as a key metric. This term refers to the total amount of greenhouse gasses, particularly carbon dioxide, that are emitted directly or indirectly by an individual, organization, event, or product. These emissions significantly contribute to the greenhouse effect, which in turn accelerates global warming and climate change. The carbon footprint is measured in terms of carbon dioxide equivalents (CO2e), allowing for a comprehensive account that includes various greenhouse gasses and their relative impact on the environment.
 
+![The carbon footprint of large-scale machine learning tasks. It includes various models running at Meta. The emissions are categorized by offline training, online training, and inference. Source: @wu2022sustainable.](images/sustainable_ai/model_carbonfootprint.png)
+
 The consideration of the carbon footprint is especially important in the field of artificial intelligence (AI). AI's rapid advancement and integration into various sectors have brought its environmental impact into sharp focus. AI systems, particularly those involving intensive computations like deep learning and large-scale data processing, are known for their substantial energy demands. This energy, often drawn from power grids, may still predominantly rely on fossil fuels, leading to significant greenhouse gas emissions.
 
 Take, for example, the training of large AI models such as GPT-3 or complex neural networks. These processes require immense computational power, typically provided by data centers. The energy consumption associated with operating these centers, particularly for such high-intensity tasks, results in notable greenhouse gas emissions. Studies have highlighted that training a single AI model can generate carbon emissions comparable to that of the lifetime emissions of multiple cars, shedding light on the environmental cost of developing advanced AI technologies [@dayarathna2015data]. @fig-carboncars shows a comparison from lowest to highest carbon footprints, starting with a roundtrip flight between NY and SF, human life average per year, American life average per year, US car including fuel over a lifetime, and a Transformer model with neural architecture search, which has the highest footprint.
@@ -268,8 +270,6 @@ The life cycle of an AI system can be divided into four key phases as shown in @
 
 * **Design Phase:**This includes the energy and resources used in the research and development of AI technologies. It encompasses the computational resources used for algorithm development and testing contributing to carbon emissions.
 
-![Bar chart comparing the carbon footprint of large-scale machine learning tasks. It includes various models, such as 'LM', 'RM', and 'BERT-NAS', showing their CO2 equivalent emissions in millions of kilograms. The emissions are categorized by offline training, online training, and inference, with 'RM1' having the highest offline training footprint.](images/sustainable_ai/model_carbonfootprint.png)
-
 * **Manufacture Phase:**This stage involves producing hardware components such as graphics cards, processors, and other computing devices necessary for running AI algorithms. Manufacturing these components often involves significant energy use for material extraction, processing, and greenhouse gas emissions.
 
 * **Use Phase:**The next most energy-intensive phase involves the operational use of AI systems. It includes the electricity consumed in data centers for training and running neural networks and powering end-user applications. This is arguably one of the most carbon-intensive stages.