Diary 2023-03-21
I was thinking of stopping because the diary page would be blank and making it a weekly journal, but if there are so many things to write every day that I can't make it to "a page with an appropriate title" in time, why not write in the diary?
After all, it is more convenient to have a "pinned page where you can write anything you want".
I think the problem was that "when a day goes by without anything being put on that page, it is not deleted".
Fork the pin-diary script...
Probably a change in Twitter's API has made it "a bit of a hassle" to republish Twitter to Scrapbox.
Maybe it's important to record the current episode.
It's silly to say "AI can't do that" without evidence, and it's silly to say that future AI can't do that based on "AI at the moment can't do that", but aside from that, it's also silly to say something that shouldn't be possible as currently structured as if it could.
GPT-4 that's great.... I'm already aware that I can get credit for the classes in the middle if I bring in AI (or I will be able to in 6 months).
I think college classes will be more difficult to evaluate.
・Return to the origin of paper examinations
・Provide a thesis level report (honestly, this could be done quickly too)
Is there any better way other than that?
What I am afraid of is that if we use AI systems as a prerequisite level of evaluation, those who can't use them may get terrible grades, so we will have to focus separately on education to enable them to use them. However, I don't want to educate tools because it will become obsolete & it's not essential / I will have a hard time teaching the essential.
I wonder if there will be strong opposition from professors to teaching how to use AI as a university course. I wonder if it will be in the form of a study group where senior volunteers teach junior students.
The format of "senior to junior" cannot be the basis for mandatory education. (Although when I was in school, that was the tradition, Chinese or otherwise.)
Even if there wasn't much of a backlash (in fact, if it wasn't my burden, I'd be less likely to get backlash on average), I think the bigger issue would be how to do it in practice.
motoso > When you try to create a character or a human being with AI, if there is an ethical filter, you can't create a strong and human character. It is not possible to make them hold dangerous things in their arms. yubais > When people can make apps in natural language, I think more people will understand that it wasn't because they couldn't program that they couldn't make apps. > yubais > I too thought "the reason I can't draw comics is because I can't draw" until I tried the original comic book work > motoso I'm afraid that there will be a generation in the future that will have GPT when they start all kinds of learning. Learning efficiency is too different.
motoso The time is there that we don't need comments on the code or anything, we can put it in GPT and have it explained to us. cicada3301_kig I think Pascal once said, and I think it's an astute point, that if you're not smart, you have to choose between timidity and boldness. If they are less intelligent, they can't put the world into a fine classifier, so they can only choose to take a wide or narrow risk margin depending on their personality, and choose their actions accordingly. They can only choose to take a wide or narrow risk margin depending on their personality, and act accordingly. cicada3301_kig Essential tremor in epilepsy practice would be a clear example. Low intelligence is interpreted as shaking hands = epilepsy, and no matter how much explanation is given, they can only react in a uniform way to the tremor. They either freak out and make a big deal out of every tremor, or they are indifferent and leave it alone. They don't see "this is epilepsy, that is not." cicada3301_kig The life history of people with low intelligence is usually optimistic in childhood, and then there is a major event in adulthood or old age, such as an illness, and all at once they turn to anxiety. The world is not divided into thin cases. The world is not seen in a thin and segmented way, so as soon as something critical is crossed, the whole world begins to look ominous. ajico_k I just realized, this is what the future looks like after AI has fully evolved. https://pbs.twimg.com/media/Froc2SWaIAA74_A?format=jpg&name=900x900#.png https://pbs.twimg.com/media/Froc5UHacAEzKBp?format=jpg&name=900x900#.png https://pbs.twimg.com/media/Froc7LJaEAAoyYK?format=jpg&name=900x900#.png
ajico_k This is included in the third volume of our coterie magazine "This is a Bad Internet". It is available for kindle unlimited. https://pbs.twimg.com/card_img/1635518539271667712/BibTtkYn?format=png&name=small#.png
https://pbs.twimg.com/media/FrlmeIiaYAIYWux?format=jpg&name=900x900#.png https://pbs.twimg.com/media/FrlmeIdaEAItFZy?format=jpg&name=900x900#.png https://pbs.twimg.com/media/FrlmeIbaUAUdJvg?format=jpg&name=900x900#.png
https://pbs.twimg.com/media/FrlmgIXaQAETcq4?format=jpg&name=900x900#.png https://pbs.twimg.com/media/FrlmgJlaYAMw0kM?format=jpg&name=900x900#.png https://pbs.twimg.com/media/FrlmgKgaUAQIVQQ?format=jpg&name=900x900#.png https://pbs.twimg.com/media/FrlmgKmaMAAbRbw?format=jpg&name=900x900#.png
---
This page is auto-translated from /nishio/日記2023-03-21 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.