The case is Mata v. Avianca, Inc. involving Mata suffering a knee injury when struck by an airplane drinks cart of an airline that subsequently entered and left bankruptcy, its details are not important.
Here is where the trouble starts as plaintiff's counsel Peter LoDuca opposes a motion to dismiss. There are some very well-formed citations like "Shaboon v. Egyptair, 2013 IL App (1st) 111279-U (Ill. App. Ct. 2013)" and "Varghese v. China Southern Airlines Co. Ltd., 925 F.3d 1339 (11th Cir. 2019)".
Defendant attempts to find these cases to write a response:
Plaintiff cites to “Varghese v. China Southern Airlines Co., Ltd., 925 F.3d 1339 (11th Cir. 2019).” The undersigned has not been able to locate this case by caption or citation, nor any case bearing any resemblance to it. Plaintiff offers lengthy quotations purportedly from the “Varghese” case [...] The undersigned has not been able to locate this quotation, nor anything like it any case. The quotation purports to cite to “Zicherman v. Korean Air Lines Co., Ltd., 516 F.3d 1237, 1254 (11th Cir. 2008).” The undersigned has not been able to locate this case; although there was a Supreme Court case captioned Zicherman v. Korean Air Lines Co., Ltd., that case was decided in 1996, it originated in the Southern District of New York and was appealed to the Second Circuit, and it did not address the limitations period set forth in the Warsaw Convention. 516 U.S. 217 (1996).
The Court orders plaintiff's counsel LoDuca to submit copies of these cases.
LoDuca sends in some cases, noting that some of the cases seem to be incomplete copies. Also notice this filing was notarized by one Steven Schwartz, with an incorrect date. Purported cases are attached like the Varghese and Shaboon decisions.
Defense "respectfully submits that the authenticity of many of these cases is questionable".
Now the Court is pissed and LoDuca is up for sanctions for citing/submitting false cases:
The Court is presented with an unprecedented circumstance. A submission filed by plaintiff’s counsel in opposition to a motion to dismiss is replete with citations to non-existent cases. (ECF 21.) When the circumstance was called to the Court’s attention by opposing counsel (ECF 24), the Court issued Orders requiring plaintiff’s counsel to provide an affidavit annexing copies of certain judicial opinions of courts of record cited in his submission, and he has complied. (ECF 25, 27, 29.) Six of the submitted cases appear to be bogus judicial decisions with bogus quotes and bogus internal citations. Set forth below is an Order to show cause why plaintiff’s counsel ought not be sanctioned.
- LoDuca responds that it wasn't actually him doing all this work, it was Steven Schwartz. Schwartz:
the citations and opinions in question were provided by Chat GPT which also provided its legal source and assured the reliability of its content. Excerpts from the queries presented and responses provided are attached hereto.
- Now the Court is more pissed and LoDuca, Schwartz, and their whole firm are up for sanctions for citing/submitting false cases and fraudulent notarization.
In conclusion I hope this was a bullshit suit because if this guy Mata had a serious complaint I don't think he's seeing any justice anytime soon
Jump in the discussion.
No email address required.
Why would you stake your entire very expensive career on this rslurration? They are the most qualified people to know how wrong it will go
Jump in the discussion.
No email address required.
At the very least he should have double checked the citations. Of course chatgpt is going to tell you they're totally good bro. Chatgpt suffers from occasional delusions and makes shit up. Aka pizzashilling.
Jump in the discussion.
No email address required.
As a codecel I once implemented a C version of Cliff Click's lock-free concurrent hashmap, so I figured that was something high-level I knew a lot about, it was something pretty obscure but I was able to learn from reading the Internet, a good test for ChatGPT. I said "Describe Cliff Click's concurrent hashmap."
OK, except "HighwayHash" is completely unrelated.
Lock-free programming could be considered the finest-grained locking ("lock-free" is a bit of misnomer, it uses CPU-level locks), I guess. "Optimistic locking" and "lock striping"? Not involved here.
Cliff Click's hashmap does use open addressing, but far from being a "key innovation" it's a standard choice you learn in baby's first data structures class. It does not use linear probing.
It does support dynamic resizing, but "incremental resizing" is not quite right, and the sentence needs to say "copy all the keys at once" to make sense.
Just kind of BS and I have no idea how popular Cliff Click's hashmap is. The whole point isn't large amounts of data, this kind of thing doesn't generally even hold the actual data instead of pointers/references, it's large numbers of threads without blocking.
Jump in the discussion.
No email address required.
More options
Context
What helps me is I ask gpt to give me links for it's citations and it usually gives me some
Jump in the discussion.
No email address required.
It will make those up, too
Jump in the discussion.
No email address required.
Not if they're clickable
Jump in the discussion.
No email address required.
Huh? It can make up links that are “clickable”.
Jump in the discussion.
No email address required.
To actual webpages of reputable sites with articles?
Jump in the discussion.
No email address required.
It’s still a clickable link, r-slur. It’ll happily make those up. You can’t tell it “cite your sources but make sure the urls resolve”. And even if you can, it’ll just link you to nonsense.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
I code with its help knowing that it's a liar. It's great for giving me ideas but it's harder to get anything usable out of it. It's directionally right
Jump in the discussion.
No email address required.
It's really really great if you hate starting at step 0. I'm much better off changing and correcting and going from there than starting from scratch and gets my ball rolling.
Jump in the discussion.
No email address required.
More options
Context
I just use it to help with vector math lol
Jump in the discussion.
No email address required.
More options
Context
More options
Context
He did check the sources by asking chatgpt if it was telling the truth.
Jump in the discussion.
No email address required.
Asking chatgpt "are these legit" "yeah totally my guy" is not checking the source.
Jump in the discussion.
No email address required.
It literally is, dumbass
Jump in the discussion.
No email address required.
It literally isn't if it turns out the sources aren't valid you tier frick.
Jump in the discussion.
No email address required.
Chatgpt is the source. Checking with chatgpt is checking the source you illiterate tard
Jump in the discussion.
No email address required.
We're not talking about chatgpt we're talking about factchecking for real cases.
Its no wonder you're a house femboy founder if you have c*m for brains.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
Very dramaphilic
Jump in the discussion.
No email address required.
More options
Context
neighbor this article is AI written lmao
Jump in the discussion.
No email address required.
Your mom is ai written.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
It makes more sense the less you know about AI lol
For the layman, you basically smash two gigantic arrays together doing matrix multiplications on some input and then whatever numbers come out is the text output
Highly reductive, but essentially anyone claiming AI is actually thinking or reasoning doesn't understand this much
Jump in the discussion.
No email address required.
Daily reminder that neural nets are literally designed to work like brains do and that anyone who thinks AIs can't think must also believe that humans can't think.
Jump in the discussion.
No email address required.
Neighbor this site is proof humans cant think
Jump in the discussion.
No email address required.
More options
Context
Yes a neural net is a complete model of the brain and not just the simple part of the brain
Jump in the discussion.
No email address required.
Let me know when we find scientific proof of the soul, chud
Jump in the discussion.
No email address required.
I think sapience is gonna be quantum related and I don't think the way we build PCB or computers will allow us to get there personally
Jump in the discussion.
No email address required.
So you're in denial. You think there's some invisible force of exceptionality that means only humans have the capacity to be intelligent. Like those morons who stop and stare at the smoke plume of an exploding volcano whilst the sentient hightail it.
Jump in the discussion.
No email address required.
No r-slur im a literal sapient biocomputer, im saying that electricity is obviously not enough to create sapience so clearly there's something happening we don't yet understand, and the physics we don't fully understand is subatomic and quantum in nature
Jump in the discussion.
No email address required.
@Borpa fight me loser what's your theory of mind
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
HUMANS ARE SUPER SPECIAL MAGIC QUANTUM BRAINS!!
At what point does a human become not sentient? Down syndrome? If so, then we can scientifically CHUD determine if your brain is too primitive to be sentient.
Ur an r-slur and stop believing in god and souls an shit
Jump in the discussion.
No email address required.
Quantum isn't magic r-slur, its how sub-particles behave, you're telling me you're confident we can achieve sapience using only 1 of over 17 elementary particles alone?
And how many light bulbs before the lamp achieves sapience too?
I can calculate chatgpt's inputs and outputs on paper given a billion years, is my paper fricking sapient?
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Souls don't exist r-slur
Jump in the discussion.
No email address required.
More options
Context
1. All things come to be from their opposite states: for example, something that comes to be “larger” must necessarily have been “smaller” before (70e-71a).
2. Between every pair of opposite states there are two opposite processes: for example, between the pair “smaller” and “larger” there are the processes “increase” and “decrease” (71b).
3. If the two opposite processes did not balance each other out, everything would eventually be in the same state: for example, if increase did not balance out decrease, everything would keep becoming smaller and smaller (72b).
4. Since “being alive” and “being dead” are opposite states, and “dying” and “coming-to-life” are the two opposite processes between these states, coming-to-life must balance out dying (71c-e).
5. Therefore, everything that dies must come back to life again (72a).
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
t. Chinese room
Jump in the discussion.
No email address required.
Point is, matrix multiplication is sufficient to emulate a brain. So you can't say "lmao these idiots think ai is sentient, don't they realize its just matrix multiplication??"
Jump in the discussion.
No email address required.
There's literally 0 reason to believe that
Jump in the discussion.
No email address required.
You have to have faith
Jump in the discussion.
No email address required.
Exactly. And until they completely map out the human brain and are able to reduce every single thought and emotion that anyone ever felt to a causal chain of neurons firing, I'm not buying that human thought can be replicated.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
There's also zero reason to think that we need to fully emulate a human brain in order for a model to be superior to a human brain
Jump in the discussion.
No email address required.
Of course not, I didn't claim otherwise. The argument we were having is not about superiority of the model but rather what constitutes as it 'thinking' in the same way like humans
Jump in the discussion.
No email address required.
Chinese room is a meme, and 'sentience' is a useless metric. I don't even know if you are thinking in the same way as a human - are you even conscious? It's not provable. We've had 3000+ years to figure this shit out, and we're still stuck on even defining it. Experiential qualia might just be an evolutionary quirk of our particular neural architecture, or it might be a common emergent feature of any substantially complex system. It's impossible to measure, and ultimately doesn't provide any good insight into a system.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
neighbor is you r-slurred?
Jump in the discussion.
No email address required.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Not particularly intelligent beings can fool humans of their sentience. See: cats and dogs
Jump in the discussion.
No email address required.
Dogs and Cats are self-aware and experience qualia, though, they are conscious. Even if people don’t correctly interpret their behavior and put it in a human framework.
GPT is just an algorithm that experiences nothing but is a remarkable parrot.
Jump in the discussion.
No email address required.
NOOO YOU CAN'T TAKE OUR JOBS AND REPLACE US YOU'RE NOT EVEN CONSCIOUS YOU DON'T EXPERIENCE QUALIA NOOOO
beep boop
Jump in the discussion.
No email address required.
I’m not opposed to having AI do shit for us, I think it’s a great took.
People seem to think that makework is necessary for people when there are other solutions to AI replacing menial jobs. Gibs might be a necessity at some point, though. This is going to be a necessary side effect of moving to a post scarcity economy.
Now if AGI ever comes around, we might have to review the ethics of forcing a conscious being into servitude.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
Citation needed, neighbor
Jump in the discussion.
No email address required.
Hang on let me just consult with chatGPT
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
NO. Neural nets are an attempt by humans to simulate the brain. We don't even understand how the brain fully works. What makes you think we got the full simulation down?
Jump in the discussion.
No email address required.
More options
Context
A cargo-cult version of how AI enthusiasts think a human brain works isn’t the same as literally works like a human brain except in the sense that a perpetual motion machine is also literally designed to produce endless energy based on an r-slured model.
Connectionist models are also lacking, Openworm couldn’t even properly emulate the 302 neurons of C. elegans despite mountains and mountains of delicious cope.
Jump in the discussion.
No email address required.
As of when?
Jump in the discussion.
No email address required.
More options
Context
More options
Context
Watch me unplug that b-word and see whose still thinking.
Jump in the discussion.
No email address required.
More options
Context
I'm so happy in the Congogerudo 11mo ago #4233506 Edited None spent 0 currency on pingsAI can think but humans can't.
Jump in the discussion.
No email address required.
More options
Context
Can you honestly blame them, after reading this post?
Jump in the discussion.
No email address required.
More options
Context
R-slur
Jump in the discussion.
No email address required.
More options
Context
More options
Context
But it can generate solutions for medium difficulty Java leetcode questions and write out lengthy paragraphs about generic topics you would have found on google.
White collar workers are finished.
Jump in the discussion.
No email address required.
Unironically though, just because they're not sapient doesn't mean they're not gonna replace us in under 5 years
You only need a single ai to be as good as a human worker for every ai to be as good as a human worker
Jump in the discussion.
No email address required.
Currently ChatGPT4 can barely scrap together a generic Java code for a super generic situation. Any kind of non-template task and it just gives you back something that often won’t even compile and def won’t work. It’s like an r-slured intern that won’t even learn from his mistakes and the accessible dataset has been drained already.
I think we will be fine. Remember when the first advanced IDEs came out and people were malding that you won’t need juniors anymore, because what else would they possibly do if you can just generate getters/setters and other monkey tasks? I member.
Jump in the discussion.
No email address required.
Its difficult for me to explain how wrong you are, its like looking at a mechanical 1 year old and saying look at this fricking r-slur there's no way it could ever do my job
We will have more effecient models, we will develop more powerful and specialized hardware, we will develop better techniques, we will generate better and better synthetic data
I know because this is already what we're doing, you dont have chatgpt4 output an answer you have chatgpt4 control chatgpt4 if you want to do anything cool rn
Jump in the discussion.
No email address required.
I’m sure that will work well.
Jump in the discussion.
No email address required.
You need to understand that this is rslurration of a magnitude and speed the likes of which this world has never seen but only aped at with man
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Skill issue for the most part. If you are shit at prompting it will give you a vague answer, but if you know it's limitations you can get it chugging away on the grunt work pretty easily. Making things multithreaded, prompting a situation for possible algorithms to use and the pros and cons of each. Of course if you go "write code for X" it will just fill in the gaps with bullshit.
Jump in the discussion.
No email address required.
So you’re saying it can’t actually write code other than boilerplate it’s seen 1000 times? You know what else can write that boilerplate? Some dumb cli, you don’t need “ai” for that.
Jump in the discussion.
No email address required.
I think identifying the most efficient algorithm to use is a big deal. If you come across a problem in what you're writing, are able to isolate the specific problem and prompt gpt to identify different ways and compare them, that's something it can do.
You still need the knowledge, but it does speed things up significantly.
Jump in the discussion.
No email address required.
If you’re not competent enough to come up with solutions and compare them, you’re not competent enough to implement them either.
You probably used to just search stackoverflow and copy/paste the first answer with a checkmark next it.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Yeah it's basically good for reminding you of things you already understand
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
I’d be surprised if GPT-4 didn’t have some spaghetti thinking ability. It is too good at remembering what it’s been doing and what it’s been asked to do before. It correctly infers which reply I’m talking about in a discussion involving multiple subjects. Even people don’t do that well. It is aware of what it’s said and what it “intended” to imply. It’s certainly not sentient but it definitely has some simple ability to reason, or I just can’t tell that it doesn’t.
Jump in the discussion.
No email address required.
ChatGPT4 has exactly no memory, what you think is memory is context, go look at the openai chat completion api reference and think deeply about how you need to change so that you aren't so easily tricked into believing in magic and fairy tales
Jump in the discussion.
No email address required.
What I need is more experience. Explain the difference between “memory” and “context” to me.
Jump in the discussion.
No email address required.
Memory implies its storing data at rest, what its actually doing is taking your entire convsation and using it as an input each time up to a context character limit essentially, once the context window is exceeded they need to start using tricks to compress or look up information from earlier in the conversation to use as input
What you think you're prompting it with is actually just a very small part of what its actually being fed as input each time
Jump in the discussion.
No email address required.
That is tantamount to memory, but I do actually see what you mean about it being different.
Jump in the discussion.
No email address required.
Yeah, when I say
I mean inherent to the model itself, we can and do emulate "memory" by storing data that can be conditionally accessed and passed in depending on the situation, but this is handled by traditional software, not the actual language model
Jump in the discussion.
No email address required.
More options
Context
More options
Context
This immediately exhausted me when I noticed it, you'd have to pay me millions to put up with the amount of essay writing these tools require.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
there's autogpt too
Jump in the discussion.
No email address required.
Does the “g” in “gpt” stand for garbage?
Jump in the discussion.
No email address required.
More options
Context
AutoGPT is still in "deathloop for anything slightly intricate" mode... like finding the best-value bike in an online shop. See
Jump in the discussion.
No email address required.
AutoGPT makes me think of Deep Fried Memes
Jump in the discussion.
No email address required.
More options
Context
More options
Context
AutoGPT does the same thing (its chatgpt in a loop) just with parsing responses for specific tokens and taking actions based on those tokens
Jump in the discussion.
No email address required.
More options
Context
More options
Context
@GabrielMartinelli lets talk about why you believe in the wizard of oz
Jump in the discussion.
No email address required.
Jump in the discussion.
No email address required.
You're the kind of person who ends up sacrificing their kid to aztec death gods because it seemed logical
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
Let all these r-slurs cope, ChatGPT is more sentient than 60% of the population and 90% of rDrama
Jump in the discussion.
No email address required.
More options
Context
More options
Context
This isn't actually an argument until we know that the brain is in fact not doing something just like that. "Neurons are just gates that only let through signals above a certain threshhold. Obviously humans can't be intelligent/ must have a soul to do the thinking"
Jump in the discussion.
No email address required.
I don't buy it
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Lawyercels seething because they can't half-butt their work like a lot of people
Jump in the discussion.
No email address required.
More options
Context
Arrogance is a common affliction in the legal profession.
Jump in the discussion.
No email address required.
More options
Context
Why do airline pilots always show up drunk?
Jump in the discussion.
No email address required.
Why do dogs lick their balls?
Jump in the discussion.
No email address required.
Because they can
Jump in the discussion.
No email address required.
More options
Context
Probably feels good
Jump in the discussion.
No email address required.
More options
Context
More options
Context
What else are you supposed to do stuck in an airport all night?
Airport lounge and hit on the flight attendants. That's it
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context