RESTATEMENT NOTAION

Consequentially, no one within our solar system fully grasps the inner workings of LLMs. Researchers are working to better understand these models, though this is a slow process. One that may take years or even decades to complete.

CONNECTION ANNOTATIONS

Just as in the Trott and Lee LLM article, word vectors are described as a matchmaking site for words. AI generated photos could be viewed in a similar way. Not every date will result in a soul mate, therefore every match is not ideal. I often find myself frustrated when typing and my cell phone will decide that it is smarter than me and knows exactly what I am going to say next. I become irate when I hit send before, I realize those changes have been made and I sound as though I only completed my elementary school education. The same can be send for these photos, when an AI model acquires many photos to compile one master image, thought to be exactly what you are looking for, only to discover that AI is so generous, that they have provided the subject of the photo with an extra arm.

1. Similar to how Trott and Lee LLM article phrased AI as being similar to "turning the water on, making sure it's coming out of the right faucet, and when it's not, running around to tighten and loosen all the valves until it does", AI generated photos (7:20) can create biases that then we must run around to resolve. The original information inputted to generate a photo is like the water. Once it is turned on and a photo is generated, but it's not the photo we were expecting, we then have to tweak and edit the information until we receive a desired outcome.

VISUAL ANNOTATIONS

1. (3:55) Code Carbon helps to provide a visual for the emissions produced in a certain time period. Much like the nutrition information on the outside of a cereal box, though it is helpful to know how many calories or grams of sugar are contained in one serving, that information would be useless without knowing what a serving size is. Even more so if we are unaware of how many calories we should consume in a given time frame. This information allows us to make informed decisions about what we should consume.

2. (5:04) This visual of the photos that are generated when she inputs her own information provides a bit of shock factor to the audience. Not only are most of those images of someone who is not her, but they are quite illicit photos of someone who is not her. Taken out of context and unknown to be AI generated, photos such as these could be detrimental to someone's career or personal life.

REACTION ANOTAION:

I never paused to consider that AI models are not contained within the digital world and exist boldly in our physical world. I figured that, since AI is so easily available on a cell phone, it couldn't possibly require so much to sustain it.

RESTATEMENT ANNOTAIONS

1. (0:37) Though at times AI can be extremely helpful, as we often hear in the media, there is also a dark side to these functions. For instance, one chatbot suggested that the solution to one man's problems was to file for divorce. Another suggested a sure-to-please recipe for a large gathering, that contained chlorine gas among its list of ingredients.

2. (1:19) AI does not exist in solitude. Its affects have left the digital world and have begun to affect our physical world. AI models have been noted to contribute to climate change, use artist and authors work without their consent, and discriminate against entire groups of people.

3. (1:56) In regard to sustainability, the figurative cloud we imagine AI resides upon is not something only contained within the digital world. We often fail to acknowledge the large quantity of materials needed to build these models as well as the sizeable amount of energy required to sustain them.

TRACKING ANNOTATION

One word can have many different meanings, depending on the context it is used in.

For example, the next time I "find myself in a jam", I seriously hope I am not literally "in a jam", fighting my way out of a Smucker's jar.

"Word vectors"

TRACKING ANNOTATION

By simply recalling the definition of "vetors" I am able to infer that this portion of the article will provide an explanation of how words are substituted for numbers in determining the relativity of two point to one another. Typically, a vector is a function represented by numbers, as it is used to determine the quantity of the inputs.

When watching the video "How ChatGPT Works Technically | ChatGPT Architecture" I found it fascinating to learn that words are represented by numbers, as they are easier for the model to process. This gives cause to question just how reliable these models are, as one slight misspelling can skew the results completely. For instance, if I were chatting with a friend via text about my plans for the holidays and they told me they were "going home to visit their parents" and I responded "Yes. I think I will go home to visit my pants too." They would easily be able to deduce my intended statement by referencing the context of our conversation. AI models fail to offer fluid thinking in these situations.

I would like to learn more about what the "constraints" of an AI model mean. When looking at the word constraint from my own personal experience with constraints in manufacturing that represent where we are falling short or what may be holding us back. Does this mean the same thing in AI or does the word simply mean the rules or conditions within which the AI model exists?