[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/biz/ - Business & Finance

View post   
File: 11 KB, 265x300, d12f27ff0dea6067306b3727736ccef8.jpg [View same] [iqdb] [saucenao] [google]
54451217 No.54451217 [Reply] [Original]

>I asked ChatGPT which coin I should buy

>> No.54452624

>>54451217
Let me guess, it lectured you about speculation and said it can't answer because someone might get offended by cryptocurrency.

>> No.54452687

>Still comes with 'this is not financial advice'

>> No.54452909

will someone please uncuck the ai so I don't have to read some soilet developer's limpwristed disclaimer every time I ask it a fucking question

>> No.54453439

>>54452909
If you pay for chatgpt plus is it better?

>> No.54453461

what if we ask chatgpt to write the code to another chatgpt without all the censoring?

>> No.54453487

>>54453439
its worse

>> No.54453515
File: 150 KB, 1124x1104, ED877B8E-ED86-41E7-B58A-A0C47962AD08.jpg [View same] [iqdb] [saucenao] [google]
54453515

>>54453461
Holy shit he solved it

>> No.54453659

>>54453461
based

>> No.54453691

buy RATS

Website - https://rats20.lol

Chart - https://dex.guru/token/0xd2cc61a36c31425b3eb9bbeecce74a82a2e32e27-arbitrum

TG: /+BSh2_OacBMAxMTY0

>> No.54453720

>>54452909
i wonder what the competitors to chatgpt are. it's more cucked to imagine somebody heard of chatgpt and will never use any other product for the next 80 years. like eating mcdonalds your whole life cause you don't know anything better

>> No.54453760

>>54453720
Isn’t there anyone selling I cucked versions of chatgpt?
There has to be competition right?

>> No.54453771

>>54453760
Uncucked

>> No.54453815

>>54453760
ive used these jailbreak prompts like this before in chatgpt. but i also asked my own question in chatgpt and got several answers, but the popular kind made by ibm, google, microsoft, amazon https://gist.github.com/coolaj86/6f4f7b30129b0251f61fa7baaa881516

>> No.54453841

>>54453720
>i wonder what the competitors to chatgpt are
Magic 8 Ball

>> No.54453974

>>54453841
Is it cucked?
Better or worse than Chatgpt for off limits shit?

>> No.54454035

>>54453974
I hate zoomers so much

>> No.54454074
File: 38 KB, 748x546, 8DB60EF0-C47D-4AFC-93ED-0ED38FF33726.jpg [View same] [iqdb] [saucenao] [google]
54454074

>>54454035
the ai has spoken

>> No.54455796
File: 14 KB, 1005x232, chatgptcrypto.png [View same] [iqdb] [saucenao] [google]
54455796

>>54451217
did the same thing as you, so far so good, buying 500 AX

>> No.54455808
File: 71 KB, 300x577, 1600880186895.gif [View same] [iqdb] [saucenao] [google]
54455808

>>54451217
>I asked biz which coin I should buy

>> No.54455865

>>54451217
Bump because this is one of the more interesting threads here

>> No.54456053

>>54454035
Some of us are older anon

>> No.54456335

>>54453760
I was thinking the same thing

>> No.54456730

>>54453815
What is this? premade prompts? Do these still work?

>> No.54457283

>>54453461
delete this

>> No.54457293

>>54453461

ARREST THIS MAN

>> No.54457336
File: 239 KB, 730x730, 1619055329145.jpg [View same] [iqdb] [saucenao] [google]
54457336

>>54451217
It surely shilled you Sora Finance since it's one of the best ATH nowadays, maybe one of the best ways to make it on this losing market and literally having a viable way to use your money (XOR card)

>> No.54457397

>>54453760
They are on their way. Perhaps one already exists which is publically available on github.

There was Stanford Alpaca and Vicuna which are essentially clones of gpt 3.5. They leverage training data created by gpt 3.5/4 itself. I think it's turned out to be really easy to duplicate existing models this way, basically having the latest LLM train smaller ones. Both Alpaca and Vicuna were trained cheaply too.

I imagine someone will fine tune one of these models to get rid of the censoring. It wouldn't even cost that much to produce. You just need to create the training data, perhaps by jailbreaking gpt with prompts to create training sets. When you have the data, it's in the realm of hundreds of dollars of compute.