Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
fart simpson
Jul 2, 2005

DEATH TO AMERICA
:xickos:

wasnt outsourcing programming to india etc supposed to do something similar? did it?

Adbot
ADBOT LOVES YOU

git apologist
Jun 4, 2003

that’s where the big money is, you see.

nurse, teacher, cleaner, doing work critical to society? get paid poo poo

fart sniffing computer arm waver like me? get paid 2x-4x the above. And what I do is worthless

Beeftweeter
Jun 28, 2005

a medium-format picture of beeftweeter staring silently at the camera, a quizzical expression on his face
no, i don't think it'll make programmers obsolete either, but it's all pretty similar: words, graphics, code

because the output is based upon prior work i'm not convinced it can come up with something truly original. in that sense looking at generated text is pretty instructive. you can see that, for example, in this style of writing this sequence of words is generally used with this lexical structure and these subjects usually go with these terms, etc.

you can abstract the same sort of idea to art or code, it's in essence the same thing. is that "good enough" for most use cases? it probably is. but if there's something that needs to be completely invented from the ground up, i'm not so sure it'd be capable of that, ever, at least in the models we're using today. it'd be foolish to expect that to remain the same as well but the same problems will persist

e: lol grammar

Beeftweeter fucked around with this message at 07:50 on Jan 10, 2023

Beeftweeter
Jun 28, 2005

a medium-format picture of beeftweeter staring silently at the camera, a quizzical expression on his face

fart simpson posted:

wasnt outsourcing programming to india etc supposed to do something similar? did it?

yes and no. i think anyone that's dealt with outsourced code will tell you that the quality will vary widely, which isn't really a surprise

will those jobs be obsoleted though? probably, and i'd wager that they'd be the first to be replaced

fart simpson
Jul 2, 2005

DEATH TO AMERICA
:xickos:

i mean honestly, how often are things completely invented from the ground up? mashing together existing things in new combinations is probably most of originality anyway

rotor
Jun 11, 2001

classic case of pineapple derangement syndrome

git apologist posted:

those people will be redirected to yet another worthless and annoying task that doesn’t need to be done

another worthless and annoying task that pays far less and allows money to accumulate more readily in the hands of the already wealthy

rotor
Jun 11, 2001

classic case of pineapple derangement syndrome

Beeftweeter posted:

because the output is based upon prior work i'm not convinced it can come up with something truly original. in that sense looking at generated text is pretty instructive, i.e. you can see that, for example, in this style of writing this sequence of words is generally used with this lexical structure and these subjects usually go with these terms, etc.

most of the big money makers in the tech world have not made all their money because of their unique code.

rotor
Jun 11, 2001

classic case of pineapple derangement syndrome

fart simpson posted:

i mean honestly, how often are things completely invented from the ground up? mashing together existing things in new combinations is probably most of originality anyway

echinopsis
Apr 13, 2004

by Fluffdaddy

Beeftweeter posted:

ai graphics and texts are birds of a feather i think. looking at it from an artistic standpoint is very similar to reading ai generated text: it's convincing, and it might even be kinda good, but it's... off, somehow. you can tell

I often agree, but I think this is a case of seeing enough of it to get a human intuition for how it's off

I suspect if someone went back in time 10 years it would be nearly indistinguishable from contemporary digital art, I doubt people would question the integrity of the person producing it, but if we're talking about AI stuff from even 2 years ago? yeah that definitely looked wrong in a fundamental way


thing is, AI approaches human skill in a way that our intuitions don't serve us very well as far as judging it goes, I don't think it's reasonable to make assumptions about it's lack of ability because we really haven't seen what it's gonna be able to do, it's such an unknown.



even if AI can't do anything better than a human, it can certainly learn from the best human in every domain to become quite superhuman in it's totality of capabilities

fart simpson
Jul 2, 2005

DEATH TO AMERICA
:xickos:

rotor posted:

most of the big money makers in the tech world have not made all their money because of their unique code.

interesting placement of negatives in this sentence structure

Beeftweeter
Jun 28, 2005

a medium-format picture of beeftweeter staring silently at the camera, a quizzical expression on his face

fart simpson posted:

i mean honestly, how often are things completely invented from the ground up? mashing together existing things in new combinations is probably most of originality anyway

true enough, and like i said, that's probably good enough for most

but if the need arises and nobody's capable of figuring it out, that's not good

fart simpson
Jul 2, 2005

DEATH TO AMERICA
:xickos:

Beeftweeter posted:

true enough, and like i said, that's probably good enough for most

but if the need arises and nobody's capable of figuring it out, that's not good

thats why they pay you to crack the tough nuts

Beeftweeter
Jun 28, 2005

a medium-format picture of beeftweeter staring silently at the camera, a quizzical expression on his face

echinopsis posted:

I often agree, but I think this is a case of seeing enough of it to get a human intuition for how it's off

I suspect if someone went back in time 10 years it would be nearly indistinguishable from contemporary digital art, I doubt people would question the integrity of the person producing it, but if we're talking about AI stuff from even 2 years ago? yeah that definitely looked wrong in a fundamental way


thing is, AI approaches human skill in a way that our intuitions don't serve us very well as far as judging it goes, I don't think it's reasonable to make assumptions about it's lack of ability because we really haven't seen what it's gonna be able to do, it's such an unknown.



even if AI can't do anything better than a human, it can certainly learn from the best human in every domain to become quite superhuman in it's totality of capabilities

people generally are pretty terrible at distinguishing details visually though. it's not really surprising either if you think about it — everyone judges art by different metrics and our perceptions of something are not necessarily going to be the same. this is a good thing if you want to fool someone, but i don't know if it's necessarily good for artistic creation. a lot of art is art simply because it was never done before

and yeah, the phrase "good enough" is probably going to get thrown around a lot because of that. for most it probably already is

abigserve
Sep 13, 2009

this is a better avatar than what I had before
I think AI will replace low hanging fruit tasks like writing boilerplate, unit tests, or individual functions. But if I want something unique, or bespoke, that's where it'll fail. It will also need a lot of oversight which is why copilot is very compelling because it's basically "autocomplete on steroids"

Fwiw I asked it to generate me a function that determines if an IP address is within a given subnet and it made an absolute mess of it but I don't doubt it'll improve over time.

The real people that need to be "look for a new career" worried are anyone that writes longform docs like contracts, HLDs, etc because AI generates written English far better than the average person

fart simpson
Jul 2, 2005

DEATH TO AMERICA
:xickos:

an ai that makes a function or set of functions for u that adheres to a single type signature u gave it. ftw

fart simpson
Jul 2, 2005

DEATH TO AMERICA
:xickos:

MoneyWatch
AI-powered "robot" lawyer will be first of its kind to represent defendant in court
moneywatch

By Megan Cerullo

January 9, 2023 / 5:38 PM / MoneyWatch

A "robot" lawyer powered by artificial intelligence will be the first of its kind to help a defendant fight a traffic ticket in court next month. 

Joshua Browder, CEO of DoNotPay, said the company's AI-creation runs on a smartphone, listens to court arguments and formulates responses for the defendant. The AI lawyer tells the defendant what to say in real-time, through headphones. 

The artificial intelligence firm has already used AI-generated form letters and chatbots to help people secure refunds for in-flight Wifi that didn't work, as well as to lower bills and dispute parking tickets, among other issues, according to Browder. All told the company has relied on these AI templates to win more than 2 million customer service disputes and court cases on behalf of individuals against institutions and organizations, he added.

It has raised $27.7 million from tech-focused venture capital firms, including Andreessen Horowitz and Crew Capital.

"In the past year, AI tech has really developed and allowed us to go back and forth in real time with corporations and governments," he told CBS MoneyWatch of recent advances. "We spoke live [with companies and customer service reps] to lower bills with companies; and what we're doing next month is try to use the tech in a courtroom for the first time."

If the robot lawyer loses the case, DoNotPay will cover any fines, Browder said.

Browder declined to disclose the name of the client and the court.
Legal in some, but not most courtrooms 

Some courts allow defendants to wear hearing aids, some versions of which are bluetooth-enabled. That's how Browder determined that DoNotPay's technology can legally be used in this case. 

However, the tech isn't legal in most courtrooms. Some states require that all parties consent to be recorded, which rules out the possibility of a robot lawyer entering many courtrooms. Of the 300 cases DoNotPay considered for a trial of its robot lawyer, only two were feasible. 

"It's within the letter of the law, but I don't think anyone could ever imagine this would happen," Browder said. "It's not in the spirit of law, but we're trying to push things forward and a lot of people can't afford legal help. If these cases are successful, it will encourage more courts to change their rules."
Lawyers "would not support this"

The ultimate goal, according to Browder, is to democratize legal representation by making it free for those who can't afford it, in some cases eliminating the need for pricey attorneys.

But given that the technology is illegal in many courtrooms, he doesn't expect to be able to commercialize the product any time soon.

"This courtroom stuff is more advocacy," he said. "It's more to encourage the system to change," Browder explained. 

He is well aware of the challenge and hurdles on the horizon. 

When he tweeted about demoing DoNotPay's robot lawyer in court, lawyers threatened him and told him he'd be sent to jail, he told CBS MoneyWatch.

"There are a lot of lawyers and bar associations that would not support this," Browder said. 
Putting ChatGPT through law school

Browder wants to arm individuals with the same tools that large corporations can typically access, but are out of reach for those without deep resources. 

"What we are trying to do is automate consumer rights," Browder said. "New technologies typically fall into the hands of big companies first, and our goal is put it in hands of the people first."

AI-powered chatbot ChatGPT has exploded in popularity recently for its ability to spit out coherent essays on wide-ranging topics in under one minute. The technology has drawn interest from investors, with the Wall Street Journal reporting that parent company OpenAI could soon attract investments valuing it at $29 billion. 

Princeton student says his new app helps teachers find ChatGPT cheats

But Browder highlighted its shortcomings and in some cases, lack of sophistication. 

"ChatGPT is very good at holding conversations, but it's terrible at knowing the law. We've had to retrain these AIs to know the law," Browder said. "AI is a high school student, and we're sending it to law school." 

Beeftweeter
Jun 28, 2005

a medium-format picture of beeftweeter staring silently at the camera, a quizzical expression on his face
lol i was just reading a thing saying that guy wants to pay $1m for someone to wear airpods and be fed lines for oral arguments before SCOTUS

which is an interesting idea, but i think they're overestimating the impact oral arguments have. i'm also not convinced it can come up with novel arguments, which is basically half of a lawyer's job

rotor
Jun 11, 2001

classic case of pineapple derangement syndrome

fart simpson posted:

interesting placement of negatives in this sentence structure

i'm full of gummi bears

Gubbinal Girl
Apr 11, 2022


There is a 0% chance that will be allowed in SCOTUS for the next decade at least. They still don't allow video cameras in. They're not about to let their entire field be replaced by a cell phone

It definitely will be allowed to replace public defenders though. And if the public defender AI happens to be real lovely then I guess the defendant shouldn't have been poor :shrug:

Cybernetic Vermin
Apr 18, 2005

don't know much about law, but suspect that this is the same sort of deal as in programming where people automatically think about the 2% of the time when one is writing actually important core logic/algorithm code, which wont get automated because it is both the interesting and most important part. as opposed to the other 98% which is dreary repetitive stuff which takes all the longer and has more errors because it is dreary and repetitive.

that is, the actual oral arguments being the actual human interaction core part of the proceedings i doubt there's any push for adding a bunch of ai nonsense to it. if i understand correctly though there's a ton of searching and collating documents, producing exhaustive filings, and summarizing up the relevant information in preparation involved, which all seems more amenable to ai (good automatic summarization and semantic search both seeming pretty plausible with current tech) and the kind of thing humans are pretty bad at.

rotor
Jun 11, 2001

classic case of pineapple derangement syndrome

Cybernetic Vermin posted:

don't know much about law, but suspect that this is the same sort of deal as in programming where people automatically think about the 2% of the time when one is writing actually important core logic/algorithm code, which wont get automated because it is both the interesting and most important part. as opposed to the other 98% which is dreary repetitive stuff which takes all the longer and has more errors because it is dreary and repetitive.

that is, the actual oral arguments being the actual human interaction core part of the proceedings i doubt there's any push for adding a bunch of ai nonsense to it. if i understand correctly though there's a ton of searching and collating documents, producing exhaustive filings, and summarizing up the relevant information in preparation involved, which all seems more amenable to ai (good automatic summarization and semantic search both seeming pretty plausible with current tech) and the kind of thing humans are pretty bad at.

idk poo poo about the field but I suspect liability will be a serious issue here and the first couple times an AI produces an inaccurate summary or makes up case law out of whole cloth there will be some fireworks.

Cybernetic Vermin
Apr 18, 2005

rotor posted:

idk poo poo about the field but I suspect liability will be a serious issue here and the first couple times an AI produces an inaccurate summary or makes up case law out of whole cloth there will be some fireworks.

no clue about the actual legal bits either, *but* with some limited knowledge about the language models i do suspect summarization and retrieval are places where the "making poo poo up" and wild inaccuracies can be tamed without some great leap forward. i.e. a matter of cutting text down rather than proper generation, making it possible to trace what in the original documents it refers to.

for a start mostly giving an indication of where to look, no doubt missing some stuff, but again one where i suspect it'd brute-force do better than a human pretty quick.

Beeftweeter
Jun 28, 2005

a medium-format picture of beeftweeter staring silently at the camera, a quizzical expression on his face
it probably is actually illegal lol

you have to be an attorney to argue in front of the supreme court and further than that you have to apply to be recognized by the bench. this is ostensibly to keep frivolous arguments from appearing and wasting their time. they're not going to countenance someone being fed lines like bruce willis

echinopsis
Apr 13, 2004

by Fluffdaddy
what was that quote from IBM in the 50s

a computer can never be held liable
so a computer cant make decisions


idk if thats true. perhaps we should try

post hole digger
Mar 21, 2011

quote:

Dear Mark,

Since its launch in November last year many people, most buzzing with a kind of algorithmic awe, have sent me songs ‘in the style of Nick Cave’ created by ChatGPT. There have been dozens of them. Suffice to say, I do not feel the same enthusiasm around this technology. I understand that ChatGPT is in its infancy but perhaps that is the emerging horror of AI – that it will forever be in its infancy, as it will always have further to go, and the direction is always forward, always faster. It can never be rolled back, or slowed down, as it moves us toward a utopian future, maybe, or our total destruction. Who can possibly say which? Judging by this song ‘in the style of Nick Cave’ though, it doesn’t look good, Mark. The apocalypse is well on its way. This song sucks.

What ChatGPT is, in this instance, is replication as travesty. ChatGPT may be able to write a speech or an essay or a sermon or an obituary but it cannot create a genuine song. It could perhaps in time create a song that is, on the surface, indistinguishable from an original, but it will always be a replication, a kind of burlesque.

Songs arise out of suffering, by which I mean they are predicated upon the complex, internal human struggle of creation and, well, as far as I know, algorithms don’t feel. Data doesn’t suffer. ChatGPT has no inner being, it has been nowhere, it has endured nothing, it has not had the audacity to reach beyond its limitations, and hence it doesn’t have the capacity for a shared transcendent experience, as it has no limitations from which to transcend. ChatGPT’s melancholy role is that it is destined to imitate and can never have an authentic human experience, no matter how devalued and inconsequential the human experience may in time become.

What makes a great song great is not its close resemblance to a recognizable work. Writing a good song is not mimicry, or replication, or pastiche, it is the opposite. It is an act of self-murder that destroys all one has strived to produce in the past. It is those dangerous, heart-stopping departures that catapult the artist beyond the limits of what he or she recognises as their known self. This is part of the authentic creative struggle that precedes the invention of a unique lyric of actual value; it is the breathless confrontation with one’s vulnerability, one’s perilousness, one’s smallness, pitted against a sense of sudden shocking discovery; it is the redemptive artistic act that stirs the heart of the listener, where the listener recognizes in the inner workings of the song their own blood, their own struggle, their own suffering. This is what we humble humans can offer, that AI can only mimic, the transcendent journey of the artist that forever grapples with his or her own shortcomings. This is where human genius resides, deeply embedded within, yet reaching beyond, those limitations.

It may sound like I’m taking all this a little too personally, but I’m a songwriter who is engaged, at this very moment, in the process of songwriting. It’s a blood and guts business, here at my desk, that requires something of me to initiate the new and fresh idea. It requires my humanness. What that new idea is, I don’t know, but it is out there somewhere, searching for me. In time, we will find each other.

Mark, thanks for the song, but with all the love and respect in the world, this song is bullshit, a grotesque mockery of what it is to be human, and, well, I don’t much like it — although, hang on!, rereading it, there is a line in there that speaks to me —

‘I’ve got the fire of hell in my eyes’

— says the song ‘in the style of Nick Cave’, and that’s kind of true. I have got the fire of hell in my eyes – and it’s ChatGPT.

Love, Nick

https://www.theredhandfiles.com/chat-gpt-what-do-you-think/

Beeftweeter
Jun 28, 2005

a medium-format picture of beeftweeter staring silently at the camera, a quizzical expression on his face
i agree with the general sentiment (that ai is imitative) but i've been saying "all art is suffering" is objectively not true and at any rate a weird justification. no idea who he is but this guy seems up his own rear end

Cat Face Joe
Feb 20, 2005

goth vegan crossfit mom who vapes



post hole digger
Mar 21, 2011

Beeftweeter posted:

i agree with the general sentiment (that ai is imitative) but i've been saying "all art is suffering" is objectively not true and at any rate a weird justification. no idea who he is but this guy seems up his own rear end

nick cave is definitely up his own rear end, regardless of his being generally right here

post hole digger
Mar 21, 2011


really bad stuff.

distortion park
Apr 25, 2011



really does feel like we're living in the bad future

e: https://www.reddit.com/r/replika/wiki/trainingtips/

distortion park fucked around with this message at 22:21 on Jan 17, 2023

Agile Vector
May 21, 2007

scrum bored



'how to emotionally abuse your caged egirlfriend' is, if not a new low, certainly a sad one for reddit

post hole digger
Mar 21, 2011

i am going to put Eugenia Kuyda in prison for life.

Gubbinal Girl
Apr 11, 2022


Rutibex posted:

yes humans become humans because they absorb language. the medium where their mind resides isn't the important part, it's the information relationships

I've worked with severely autistic people, and I never had a problem seeing them as human. gpt3 is much higher functioning than the people I've worked with

(USER WAS PUT ON PROBATION FOR THIS POST)

(USER WAS BANNED FOR THIS POST)

(USER WAS PERMABANNED FOR THIS POST)

rotor
Jun 11, 2001

classic case of pineapple derangement syndrome
thats the bar for a permaban??

Cat Face Joe
Feb 20, 2005

goth vegan crossfit mom who vapes



rap sheet says they requested it

post hole digger
Mar 21, 2011

rotor posted:

thats the bar for a permaban??

the bar will be much lower when jeffeffery makes me the boss around here.

post hole digger
Mar 21, 2011

echinopsis
Apr 13, 2004

by Fluffdaddy
lol

Cybernetic Vermin
Apr 18, 2005

i increasingly suspect that a lot of what is happening with chatgpt is a good old eliza effect. with a bunch of gpt-generated text sure, but i generally suspect that there's a bunch of hard logic layered on top which crudely parses a couple of classes of questions (e.g. "but in the style of x", "explain y", "that's not correct, should be z") tied up with rigid ways of (re-)querying the model.

which is significant mostly because that part is back in the land of the fragile and labor-intensive, not in itself representing very real progress. though it does improve on the past gpt demos by *not* just randomly generating garbage.

Adbot
ADBOT LOVES YOU

echinopsis
Apr 13, 2004

by Fluffdaddy
feed that unfiltered poo poo directly into my eosophagus

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply