11 replies to this topic

[BSLMikeLowe]

This has gained a lot of traction the past couple months, and sparked a lot of debate. For better or worse, AI that can realistically replace humans has arrived. For those in fields like coding and journalism, this is probably going to be a game-changer. Take Buzzfeed's announcement today.

 

https://www.reuters....eal-2023-01-26/

 

Today's announcement of their plans to create content via ChatGPT/OpenAI comes just a month after they announced layoffs of 12% of their workforce. And investors seem to love it (as evidenced by a 157% jump in Buzzfeed's share price).

 

Again, this genie isn't going back in the bottle. For those in certain fields, they probably want to learn what "prompt engineering" is, and learn it both well and quickly.

[BobPhelan]

For the record, it takes a human touch to write a three up three down article during the Orioles season…

[BSLMikeLowe]

For the record, it takes a human touch to write a three up three down article during the Orioles season…

 

That's where the prompt engineering comes in. The better the input you give the chatbot as to what you want, even the style and tone, the better the end product will be. If you tell it to be dramatic or humorous, it will do so. Maybe not exactly as you would have written it, but enough that you can tell it understood what you wanted. It's not like it's almost entirely effortless or doesn't take some knowledge on the human side (yet).

 

And for now, while it's in testing mode, it is only aware of events that occurred through 2021. So if you ask it to write a 1,000 word essay on the Orioles 2022 season, you will get some unexpected results that will probably be useless, or at least require major edits on your part. Just for shiggles, a couple weeks ago I asked it to write me a summary of the Georgia-TCU CFP title game. It didn't know the score or stats, and had Jake Fromm and Deandre Swift as Georgia's QB and RB. Even when I prompted it with the actual final score, the rest came out largely the same. But that will eventually change when it's officially launched at some point.

[BobPhelan]

That's where the prompt engineering comes in. The better the input you give the chatbot as to what you want, even the style and tone, the better the end product will be. If you tell it to be dramatic or humorous, it will do so. Maybe not exactly as you would have written it, but enough that you can tell it understood what you wanted. It's not like it's almost entirely effortless or doesn't take some knowledge on the human side (yet).

And for now, while it's in testing mode, it is only aware of events that occurred through 2021. So if you ask it to write a 1,000 word essay on the Orioles 2022 season, you will get some unexpected results that will probably be useless, or at least require major edits on your part. Just for shiggles, a couple weeks ago I asked it to write me a summary of the Georgia-TCU CFP title game. It didn't know the score or stats, and had Jake Fromm and Deandre Swift as Georgia's QB and RB. Even when I prompted it with the actual final score, the rest came out largely the same. But that will eventually change when it's officially launched at some point.

I was kidding haha but it’s all interesting, hopefully more good comes from it than bad.

[BSLMikeLowe]

I was kidding haha but it’s all interesting, hopefully more good comes from it than bad.

 

Ha, missed that. But in a way, it still requires a human touch. The interesting thing is that eventually it's supposed to be able to curate your own previous writing samples (past three up three downs) and produce results that emulate your individual style.

 

I imagine it will be both good and bad. It'll enhance our productivity as individuals, but inevitably will cost some people their jobs.

[BSLChrisStoner]

The Ringer: The AI Revolution Could Be Bigger and Weirder Than We Can Imagine

https://www.theringe...-we-can-imagine

[You Play to Win the Game]

Once again ushering in new technology without any sort of conscience or guard rails. Should be fun.

[BSLChrisStoner]

The Ringer: AI Programs Have Started Making Pop Music. Will Anybody Listen?

https://www.theringe...al-intelligence

[BSLChrisStoner]

The Ringer: OpenAI Tried to Fire Sam Altman. It Only Made Him More Powerful.

[Mike in STL]

I thought AI would be the beginning of the end. James Cameron warned us in The Terminator and here we are.

 

After further review. I didn't even try to to trick AI. I simply typed in "Lamar Jackson" and this is what AI generated.

 

 

What the hell is that?

 

Google image search is more dangerous than AI apparently. I'm less scared than I was 20 minutes ago. 

[BSLMikeLowe]

Man Lamar faked out that defender so bad I doubt his right leg will ever heal.

[Mike in STL]

Man Lamar faked out that defender so bad I doubt his right leg will ever heal.

And his arm fell off!