How much money will I lose trading in?

[Ogre]

Last year my optometrist cousin was telling me he is already 70% obsolete and mainly is around so a human can take the blame in case something goes wrong.

since we already hijacked the thread, here is what this guy thinks on the subject (long and not much new):

I’ve started using ChatGPT extensively for a lot of things. I needed to automate a process I do weekly (export spreadsheet to Excel, send an email to a specific address). Nothing huge, but I just asked CharGPT to do it and it had 80% of the leg work done in 20 seconds.

I also use it extensively for scaffolding and boilerplate. It’s almost always easier/ faster to ask ChatGPT to do something then adapt what it cranks out then it is to start from scratch. Been working with some old code in PHP lately and it’s been years since I’ve dug into this stuff. I get the 80% done with automation and focus on the finer details myself.

This is with just the basic stupid web interface. Going to start experimenting with the API soon.

It is real. It is improving fast and it is often entirely wrong about stuff. Kind of scary because I know how much people are going to be leaning on it shortly.

Sponsored

 

[charliemagpie]

Yesterday I asked for a simple shop lease, I couldn't believe what I was seeing.

I purposefully just asked in simple English...
ie.. write me a one page shop lease Rent $50,000 per year,start date is 1/6/2023 , length 5 yrs, yearly increases per CPI.

And that was a test. !!!

When this catches on... the hundreds of simple tasks which professionals provide will be decimated .. this is bread and butter stuff which helps pay for office overheads. And its here... it's just a start. Wow

 

[Diehard]

It is real. It is improving fast and it is often entirely wrong about stuff. Kind of scary because I know how much people are going to be leaning on it shortly.

It is good that it is wrong. It trains people to think of themselves as the responsible party when they use it. But as it becomes right more often, people will let go of steering wheel and that is when we should start worrying about whose black box we are using. I started using it as I would Google search engine by just asking questions and as you said, I got wrong answers. When I asked for reference, I got dead links. Even when it is wrong it sounds confident. Then I learned learning to ask for the right thing in the right way and adjusting the request is the trick in this partnership.

 

[cvalue13]

It is real. It is improving fast

If you haven’t yet, check out AutoGPT - available on GitHub so far.

Two weeks ago one guy wrote in the ability for several models to “talk” to one-another. So they are now prompt-engineering themselves: ask it a complex task, and it begins assigning workflows to the other models ‘in the room’, and they begin recursively prompting one-another until completed.

Basically an early glimpse at language model autonomy.

 

[Ogre]

It is good that it is wrong. It trains people to think of themselves as the responsible party when they use it. But as it becomes right more often, people will let go of steering wheel and that is when we should start worrying about whose black box we are using. I started using it as I would Google search engine by just asking questions and as you said, I got wrong answers. When I asked for reference, I got dead links. Even when it is wrong it sounds confident. Then I learned learning to ask for the right thing in the right way and adjusting the request is the trick in this partnership.

‘Unfortunately I think many people don’t take that lesson away and just use what’s output. I’ve seen it too often in conversation.

 

[Diehard]

If you haven’t yet, check out AutoGPT - available on GitHub so far.

Two weeks ago one guy wrote in the ability for several models to “talk” to one-another. So they are now prompt-engineering themselves: ask it a complex task, and it begins assigning workflows to the other models ‘in the room’, and they begin recursively prompting one-another until completed.

Basically an early glimpse at language model autonomy.

That is like a lot of team sports. If everyone involved is capable, that is where the magic happens. Once you add an incompetent idiot to the mix, the whole game is ruined. I just hope humans (the incompetent idiots) are at least allowed to watch the game.

 

[Ogre]

The ChatGPT layoffs are already coming fast and heavy. TONS of industries rely on people writing fairly formulaic stuff.

In the programming world, we spent years outsourcing to India. Very Very soon, companies will be insourcing to ChatGPT. Most of the sort of projects suitable for outsourcing are equally suitable for AI to develop with a babysitter. That seems to be very much what my job is transforming into already.

There will very soon be tools specific to the specific industries.

 

[CyberGus]

In the programming world, we spent years outsourcing to India. Very Very soon, companies will be insourcing to ChatGPT.

I discussed with my co-workers about using GPT for writing code, but the idea was rejected. ChatGPT does not provide footnotes, so there's no way to tell where code comes from. If Chat GPT just snips something out of GitHub and then I sell it with my commercial product, I've violated the GPL.

Of course, people will try. Once, while debugging, I searched Google for the error message to check if anyone else had complained, and was shocked to find the exact message in a code sample of a book on programming. The entirely of the code was taken from that example, spelling mistakes and all. Needless to say we hastily re-wrote it from scratch.

 

[Ogre]

I discussed with my co-workers about using GPT for writing code, but the idea was rejected. ChatGPT does not provide footnotes, so there's no way to tell where code comes from. If Chat GPT just snips something out of GitHub and then I sell it with my commercial product, I've violated the GPL.

Of course, people will try. Once, while debugging, I searched Google for the error message to check if anyone else had complained, and was shocked to find the exact message in a code sample of a book on programming. The entirely of the code was taken from that example, spelling mistakes and all. Needless to say we hastily re-wrote it from scratch.

It doesn’t replace anything critical. Scaffolding and simple stuff I don’t want to spend time on. Particularly this new project where I’m working on a tool to migrate data on an ancient PHP code base to ours. This is all internal, not-for distribution so no copyright issues.

But there are huge copyright issues with ChatGPT other projects I’ve worked on it is pretty clear it’s swallowed monstrous amounts of copyrighted material and is perfectly willing to paraphrase it quite closely. To the point where I’m pretty sure there would be legal actions if this was published.

 

[DMC-81]

 

[CyberTruckeeTheOne]

My daughter and son-in-law said they already suffered (paper) loss because Tesla lowered the prices of their X and S Plaid that they hold off selling/trading.

Foundation Cybertruck may however increase in value as it maybe a collector's item if Musk can be true to his words that it will only be produced in limited quanity, 1,000 as announced but it seems Tesla already extended orders to I don't know how many more.

 

[dinizchipon]

With how things are now, depreciation on EVs like Teslas isn’t as bad as most gas cars, but it’s still a gamble, especially if Tesla tweaks pricing again.
If it were me, I’d keep an eye on how much value similar cars are holding onto after a year or so. Also, if you're into minimizing losses or making the most of your money, tools like https://forexflexea.com have been a lifesaver for me in other areas—automating trades while I focus on bigger stuff.

Sponsored