ChatGPT is bullsh*t

They had to rewrite all their lesson plans, and adjust the curriculum. Students get timed writing assignment in the classroom so they have the time to apply their writing and organizational skills in the class. They can ask the teacher clarifying questions in the class. This builds a positive learning environment in the class.
For homework the students can get a reading assignment etc.

 

Those AI detectors are much more fxxxing insane!

We feed AI with the best human art, then it turns out that the fewer grammatical errors you make, the more decent wording you select, and the more likely your homework will be considered AI-generated!

 

I can barely imagine what writers of essays and other prose will write like to keep their work from looking as if ChatGPT did it. Same with visual artists and their work, to keep it from looking as if Midjourney were involved.

 

When feeding prompts in to ChatGPT, you can ask the ChatGPT to suggest a better question or prompt. Then use if you like a refined question etc.
There are very useful patterns in ChatGPT.
Tell the large language model to create the question that's better, but rather than requiring us to go and cut and paste it or tell the large language model
use the question that you just generated, we just have it say do you want to use this question instead. If we say yes, it uses that question,
if we say no, it should use our original question.
You can tell ChatGPT to write only original work, and see the output.

As to writers and actors, artists all are affected and many protest, asking for action to be taken.
Here is a development in the state of CA.

https://www.msn.com/en-us/news/us/california-passes-ai-laws-to-curb-election-deepfakes-protect-actors/ar-AA1qLaAT

California passes AI laws to curb election deepfakes, protect actors.

"California Gov. Gavin Newsom (D) signed into law a raft of artificial intelligence bills Tuesday, aimed at curbing the effects of deepfakes during elections and protecting Hollywood performers from their likenesses being replicated by AI without their consent.

There is growing worry about deepfakes circulating during the 2024 campaign, and concerns over Hollywood’s use of artificial intelligence were a prominent part of last year’s historic actors strike. California is home to “32 of the world’s 50 leading AI companies, high-impact research and education institutions,” according to Newsom’s office, forcing his government to balance the public’s welfare with the ambitions of a rapidly evolving industry."

 

I love ChatGPT and use it almost everyday! I've used it for coursework, math content, Instagram posts, and cover letters for jobs. I've used other free AI tools including but not limited to Microsoft Copilot and Google Gemini. They all have errors from time to time though.

 

This message was brought to you by ChatGPT.

 

We've heard over and over in the news that these large bank language models make all kinds of mistakes and can lie to you, and they'll never admit their mistakes in the first place nail to keep telling you they're right and produce weird outputs. Of course, this fundamentally misses the point of the tool.
These aren't meant to be tools for answering questions necessarily. Now sometimes they can answer questions and do it very accurately. Sometimes they produce things that look like answers but aren't actually accurate.
As long as we understand they're really for generating text and sometimes the texts they generate can answer things for us correctly and sometimes it's going to make mistakes.
As long as we realized that we can use these tools effectively.
Now one thing we can do is we can use prompts to support us in our identification of what are the facts that the tool is putting into its output and which of those facts really matter and we can go in then follow up on them and check them.
Now one of the things that happens sometimes when we're working with output from these tools, is that because the output looks so realistic, so humans so convincing, we assume that it's true and when it's producing large volumes of text, it can be hard to spot all the different assumptions and facts that are baked into the text that we really need to follow up on and back check. Anything you get out of one of these models you need to take ownership of and make sure that you've checked it thoroughly and really believe it.

We can put ChatGPT in to Fact-Checking Mode. We can instruct it to use external Fact Checking sources, be it scientific publications or other trust wordy accredited sources of information.

 

OpenAI has now raised more than $20 billion with a fresh megaround led by Thrive Capital( previous investor).
Existing investor Microsoft also reportedly joined the new $6.5 billion round, while Apple dropped out.
That values OpenAI at $157 billion post-money.
Microsoft, Nvidia, SoftBank, Khosla Ventures, Altimeter Capital, Fidelity, and MGX also participated in the fundraising.
OpenAI was already the world’s best-funded AI startup. But the mammoth new tranche puts the San Francisco company in a category all its own.