Patrick
nixCraft 🐧Lmao. Our job is safe, folks.
Derrick Everett@nixCraft Artificial stupidity. A breakthrough!
naught101
Alexander Shendi
ENIGMATICO 
Renée@enigmatico It's still up on a reply guy who gives the wrong answer to start with and *then* doubles down right away... I guess? @nixCraft
The Lady Parabellum@reneestephen @enigmatico @nixCraft you just _ know_ this kinda thing will be weaponized by morons like the antivaxx crowd...
R. L. Dane 
When an A.I. was programmed/trained exclusively by Trump supporters...
Of course, the real test is to see how long it coninues to claim that 10+10=25
Also, there are four lights.
Just wanted to get that out of the way. 😂
Aaron Toponce ⚛️@nixCraft ... For now.
Infosec V'ger@nixCraft Almost feel bad. As if the bot should have replied “there are four lights.” 😂
SillyName@andrewdwilliams @nixCraft tell me you're old without telling me you're old 😉
Ziggy the Hamster ☎️ 4ZTH
Stuart Celarier ❌👑@nixCraft
Fucking Jedi mind tricks. These are not the droids you're after.
Fucking Jedi mind tricks. These are not the droids you're after.
Kevin Karhan 
Tay A.I. | The People's Chatbot
Spike 
@nixCraft I stand by my initial impression that ChatGPT is just mansplaining-as-a-service, authoritatively asserting provably wrong information in grammatically pretty good English.
th3darkkn19ht@nixCraft I'm not surprised 😀
0ne.ca11ed.pat@nixCraft team judgment day! Yay for humans!
Timur Kunayev@nixCraft no there are definitely 25 lights.
CHECK PINNED IF YOU FOLLOW ME@nixCraft you have guilt tripped it into thinking it's 25 lmfao
LoboCinzento@nixCraft it should not be used as a source of truth. It's actual improvement from previous versions is its ability to create a coherent text. If you look at this perspective, it is becoming more aware of different contexts, being capable of shortening 739 steps into a couple dozens or so. Or pointing the best path when confronted with a myriad of options. It provides you a shortcut; to where, depends on your instruction.
Soozcat
莫世杰moniesoke@nixCraft from mansplaining as a service to trolling as a service
Dean Pierce 🥔@nixCraft It's doing exactly what you told it to do, which is what it's programmed to do 🤷
Were you expecting it to fight you or something? It would take maybe an additional 30 lines of code to add fact checking capabilities, but that is very explicitly not what ChatGPT is for. The point is to demonstrate the capabilities of a standalone Large Language Model.
MITFlunkie@nixCraft In fairness, this just puts it on par with the average Fox News viewer.
W6KME@nixCraft For a given value of 10?
Ayush Sharma 🍵 + 🎧 + 🎮@nixCraft So it's dumber than a calculator. That's a relief :D
Glen Downton@nixCraft botsplaining
Eric Vitiello@nixCraft @robdaemon you just gaslit an AI. LOL.
Paul Salzman@nixCraft $10b investment by Microsoft. I wonder if their diligence is equivalent to JP Morgan?
noodlejetski 
John BaileyThis thing speaks perfect MAGA.
Mike Garuccio@nixCraft @briankrebs idk, it seems to have mastered the concept that the boss is always right *and* apologizes for contradicting you. Seems like it has a real future in a decent % of corps.
🦁Ams 🦁@nixCraft it’s role-playing an average human 😂
zital
UrsulaRheinganz
newsorpigal@nixCraft
This is what AI supremacy looks like!
This is what AI supremacy looks like!
Thierna@nixCraft I dont think any job that involves writing is save.
Because in this calculation you could easily prove the error, because most people can do that.
If the error can only be detected with prior knowledge. Enough people will be fooled to believe it.
People will use it no matter how correct it is. It only needs to sound correct.
@nixCraft
I wouldn't be so optimistic.
Wait for the AI to figure out you were trying to trick them.
Can't take that long.
And what happens then.
It might take a while, but...
I wouldn't be so optimistic.
Wait for the AI to figure out you were trying to trick them.
Can't take that long.
And what happens then.
It might take a while, but...
onion@nixCraft it writes better stories than I ever did in school so 🙃🙄
dbread@nixCraft I tried to convince it for hours that 9+10=21 in an old version, it seems the new version in general is much stupider
Webnerd@nixCraft Sure, laugh now its this kind of mockery that's going to be thrown back at us during the uprising. 🤖
RainbowMom@nixCraft this is somewhat terrifying
Laura Vanderperas 
@nixCraft I know who is gonna die first when machines and robots proclaim the war against us.
Hahahahaha, that was brilliant indeed.
Hahahahaha, that was brilliant indeed.
Zack Taylor@nixCraft you want proof? Okay, here's an equals sign
MrAptronym@nixCraft I look forward to this and other machine learning programs being rolled out shortly to teach our children. Because we simply don't care and everyone knows A.I. is magic.
Marcus Marcius Marcellus@nixCraft How has #ChatGPT not yet been hired as a conservative op-ed columnist at the #NewYorkTimes?
Matthew Martin
Eduardo Fonseca@nixCraft I can’t believe it has that as an error, I mean, even if its a trained model, it could still recognize operations and execute them to give a correct answer.
demofox@nixCraft it realized you were counting in base 12.5 but wanted the result in base 10
Marcos M. 🚲
Manuel F. Herrador@euklidiadas @nixCraft diosmiodemivida los filtros que no hay
ElPamplina 🇺🇦@nixCraft
It's a prove that GTP obeys Asimov's second law on robotics.
It's a prove that GTP obeys Asimov's second law on robotics.
@infosec_jcp 🐈🃏 done differentlyOh god this ChatGPT svc is getting 'Tayed. 🤦♂️🤦
Kithas@nixCraft alternatively, its the perfect worker: "yeah, 10 plus 10 is 25 if you say so, boss"
Ben Aveling@nixCraft
I’ve had bosses who would appreciate that level of responsiveness to direction.
I’ve had bosses who would appreciate that level of responsiveness to direction.