\
  The most prestigious law school admissions discussion board in the world.
BackRefresh Options Favorite

"AI" is borderline just a scam now lol

guys like altman are just knowingly lying about current and ...
Citrine appetizing stage french chef
  11/26/25
That's the art of the deal
Navy crackhouse masturbator
  11/26/25
...
Citrine appetizing stage french chef
  11/26/25
I can't even get grok to get my new imagine wife to show me ...
brass stag film
  11/26/25
...
Maniacal dark mother set
  11/26/25
...
free-loading black public bath
  11/26/25
It's Over
Drab Gaming Laptop Bbw
  11/26/25
sounds like it aced the turing test.
slate contagious kitty
  11/26/25
...
Trip range halford
  11/26/25
...
constantscreen the great
  11/29/25
SV is a uniquely scammy culture where tech cos have been all...
Trip range halford
  11/26/25
Ilya Sutskever's new company has 30 billion in VC funding an...
Citrine appetizing stage french chef
  11/26/25
In my adult life I've made something like 1,000x more money ...
Trip range halford
  11/26/25
TWTR was incredibly lulzy. For years it was mind-boggling th...
Drab Gaming Laptop Bbw
  11/26/25
...
NSAM's Schizoid Conspiracy Advent Calendar
  11/29/25
...
house-broken fantasy-prone hissy fit
  11/26/25
180. Accurate, intelligent, concise. The only other thing...
Onyx roast beef
  11/26/25
Good post. SV is good at displacing shittier industries gene...
Hideous Casino Jap
  11/26/25
Um yeah it has already. A handful of things actually. Junior...
crawly aqua stead
  11/26/25
I follow the legal industry fairly closely and am not aware ...
Trip range halford
  11/26/25
(MoFo tranny partner)
Patel Philippe
  11/26/25
CRISPR came out in 2012 and it brought a lot of investment t...
''"'''"''"''''"''
  11/26/25
...
NSAM's Schizoid Conspiracy Advent Calendar
  11/29/25
So good for us regular people who need to stay employed?
Deranged nursing home rigpig
  11/26/25
no because it will take a while to take advantage of current...
Opaque bipolar station
  11/26/25
are pretty much all programmers fucked?
insanely creepy trump supporter
  11/26/25
to be clear, everyone is fucked unless you're in the elite t...
Opaque bipolar station
  11/26/25
Wage slaves are always and everywhere fukt 4 lyfe
Louis Poasteur
  11/26/25
If we could just get it where it would stop hallucinating la...
poppy cumskin
  11/26/25
Going to be funny when your AI watches Secretary and starts ...
free-loading black public bath
  11/26/25
This will absolutely backfire on me and lead to my death
poppy cumskin
  11/26/25
I understand why it might feel that way from your perspectiv...
misanthropic trailer park gunner
  11/26/25
doesn't the LLM model have a theoretical and a practical lim...
insanely creepy trump supporter
  11/26/25
doubt it. that shit will probably just keep SCALING AND SCAL...
misanthropic trailer park gunner
  11/26/25
can't tell if this is sarcasm
insanely creepy trump supporter
  11/26/25
I’m partially kidding. I don’t think it will be ...
misanthropic trailer park gunner
  11/26/25
Obviously if you can predict what a person will say in any s...
,.,....,..,.,.,,,,..,..,.,..,.,.,.,...
  11/26/25
are the non-diminishing returns in the (online) room with us...
Citrine appetizing stage french chef
  11/26/25
nah they'll come though I promise
misanthropic trailer park gunner
  11/26/25
https://x.com/GoySuperstar/status/1993719651495547061 btw...
Citrine appetizing stage french chef
  11/26/25
"We know that LLMs can't generalize" what does ...
Opaque bipolar station
  11/26/25
it means that LLMs can't generalize, Champ LLMs can only ...
Citrine appetizing stage french chef
  11/26/25
yes but give me a specific example of what you are talking a...
Opaque bipolar station
  11/26/25
well, LLMs *aren't* trained on the entirety of human knowled...
Citrine appetizing stage french chef
  11/26/25
you are right that they aren't trained on "all of human...
misanthropic trailer park gunner
  11/26/25
they can't generalize to all legal cases they just can't ...
Citrine appetizing stage french chef
  11/26/25
I'm not a lawyer, I need an actual example. If you're tr...
Opaque bipolar station
  11/26/25
try it with whatever domain you're an expert in. they're alr...
Citrine appetizing stage french chef
  11/26/25
I'm a programmer so automation for 90% of us will probably b...
computer online
  11/29/25
lol this is not what cant generalize means and it’s al...
misanthropic trailer park gunner
  11/26/25
GoySuperstar is the go-to source for AI analysis
chrome point place of business
  11/26/25
its a complete scam if youre talking about the companies act...
exciting coldplay fan native
  11/26/25
...
Hideous Casino Jap
  11/26/25
lol what it can do is already more impressive than robocops ...
misanthropic trailer park gunner
  11/26/25
like what?
exciting coldplay fan native
  11/26/25
Lol. Look I'm not trying to go full peter thiel here (I act...
Trip range halford
  11/26/25
but what if we end up building flying cars soon with all of ...
misanthropic trailer park gunner
  11/26/25
Then that's great. Right now I've got a website that gave m...
Trip range halford
  11/26/25
nanotech was based on speculative capability claims, many of...
misanthropic trailer park gunner
  11/26/25
It was also the domain of a set of fields and industries tha...
Trip range halford
  11/26/25
techno utopia is BS yes. doesn't mean AI isn't legit.
misanthropic trailer park gunner
  11/26/25
"its useful for "what temperature should i cook a ...
Opaque bipolar station
  11/26/25
It’s just too spammy right now. Everything feels sligh...
Soggy parlour
  11/26/25
many products are scams for sure. However there are lots tha...
razzle-dazzle newt
  11/26/25
good point imo. scientists are lazy....
Hideous Casino Jap
  11/26/25
It's interesting how cynical so much of the board is about A...
Hot principal's office jew
  11/26/25
There is no "conventional wisdom." It's just peopl...
Hideous Casino Jap
  11/26/25
conventional wisdom doesn't even support AI though. there ar...
misanthropic trailer park gunner
  11/26/25
The people who are decisively wrong are the ones who think i...
Hideous Casino Jap
  11/26/25
yeah I agree with that. there are multiple arguments that go...
misanthropic trailer park gunner
  11/26/25
because they're lawyers one of my favorite all time xo po...
Opaque bipolar station
  11/26/25
"AI" is just freshman linear algebra, sophomore st...
chest-beating curious becky
  11/26/25
...
Citrine appetizing stage french chef
  11/26/25
I suppose you also think jet engines are "just" so...
misanthropic trailer park gunner
  11/26/25
(charlatan) https://www.youtube.com/watch?v=Fk0QC7wf2do ...
chest-beating curious becky
  11/26/25
I agree with you that tech in general isn't necessarily for ...
misanthropic trailer park gunner
  11/26/25
I do think the “agents” moment is pretty open en...
crawly aqua stead
  11/26/25
AI is only going to get better. GPT 5 seems way better than ...
flesh autistic lay elastic band
  11/26/25
TiTcR
frisky box office multi-billionaire
  11/26/25
...
Ultramarine aggressive fortuitous meteor
  11/26/25
I'm pretty dumb and retarded when it comes to this shit (hon...
iridescent toaster sanctuary
  11/26/25
https://wccftech.com/quantum-computing-will-pop-the-ai-bubbl...
OYT Magnus
  11/29/25
Quantum computing has been a couple years away for like 15 y...
computer online
  11/29/25


Poast new message in this thread



Reply Favorite

Date: November 26th, 2025 10:06 AM
Author: Citrine appetizing stage french chef

guys like altman are just knowingly lying about current and future projected capabilities. it's really obvious by now that LLMs are not going to make any more qualitative leaps, and the "scaling hypothesis" no longer holds

i'm not saying that LLMs won't keep getting marginally better, and that people will find very useful ways of implementing them in practice - they will. but any talk of "AGI" or generalized capabilities in LLMs is at this point knowingly and deliberately misleading and meant to deceive the public and investors

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462237)



Reply Favorite

Date: November 26th, 2025 10:07 AM
Author: Navy crackhouse masturbator

That's the art of the deal

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462243)



Reply Favorite

Date: November 26th, 2025 10:31 AM
Author: Citrine appetizing stage french chef



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462348)



Reply Favorite

Date: November 26th, 2025 10:35 AM
Author: brass stag film

I can't even get grok to get my new imagine wife to show me her tits

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462370)



Reply Favorite

Date: November 26th, 2025 11:16 AM
Author: Maniacal dark mother set



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462463)



Reply Favorite

Date: November 26th, 2025 11:21 AM
Author: free-loading black public bath



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462482)



Reply Favorite

Date: November 26th, 2025 12:01 PM
Author: Drab Gaming Laptop Bbw

It's Over

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462613)



Reply Favorite

Date: November 26th, 2025 12:12 PM
Author: slate contagious kitty

sounds like it aced the turing test.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462650)



Reply Favorite

Date: November 26th, 2025 1:31 PM
Author: Trip range halford



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462872)



Reply Favorite

Date: November 29th, 2025 8:58 AM
Author: constantscreen the great



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49469756)



Reply Favorite

Date: November 26th, 2025 11:02 AM
Author: Trip range halford

SV is a uniquely scammy culture where tech cos have been allowed to huck futuristic bullshit to a public and a market that doesn't fully understand the tech and is desperate for something to speculate on. Compare this to say, the biotech or pharma fields, where they could theoretically do the same thing but the regulatory environment is oppressive enough to where it's reigned in the culture, so ppl shy away from hype even if that hype wouldn't actually get you cracked down on. (I always say that if Elizabeth Holmes had just stuck to SV bread and butter instead of making a fake blood testing machine, she'd 100% be a billionaire, and probably have moved on to CEO of an incumbent firm, today.)

As big a problem as (maybe bigger than) the technologists themselves is VC. The firms VCs invest in aren't publicly traded, so there's no opportunity to short, resulting in a field where no one shits on anyone else's investment and it's a pure one-way positive feedback loop. Venture capitalists are known to get catty and defensive as fuck at the most gentle/accidental passing criticism from other VCs, because that's not how the game is played and everyone should know that a rising tide of hype raises all fraud ships.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462432)



Reply Favorite

Date: November 26th, 2025 11:11 AM
Author: Citrine appetizing stage french chef

Ilya Sutskever's new company has 30 billion in VC funding and when asked how they are going to make money, he straight up did not have any answer. didn't even pretend like they had a plan

meanwhile actual american businesses in industries that actually make and do important stuff can't get any funding at all

"private equity" sure seems like nothing more than an entire "industry" of pump and dump scamming

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462450)



Reply Favorite

Date: November 26th, 2025 11:22 AM
Author: Trip range halford

In my adult life I've made something like 1,000x more money on crypto than the stock market, despite investing MORE in the latter.

I am by no means some deep ideological believer in crypto, but when TWTR reached a market cap of $60B while posting growing YOY losses as their -4% profit margin widened--with no plan for doing anything different; dorsey would just give different answers everytime he was asked, and that other turdskin would just start babbling in hindi or whatever and hope that the white-girl CNBC reporter would start to wonder if she picked out the wrong turd from the conference room and just abashedly back away--I really started to appreciate the comparatively Sound Fundamentals of the BTC fake (and never actually used) internet-money market.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462485)



Reply Favorite

Date: November 26th, 2025 12:04 PM
Author: Drab Gaming Laptop Bbw

TWTR was incredibly lulzy. For years it was mind-boggling that it was included in bigtech indices despite producing minimal revenue

and then it turns out its value prop the whole time was being a shiny object to the world's richest sperg

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462624)



Reply Favorite

Date: November 29th, 2025 9:45 AM
Author: NSAM's Schizoid Conspiracy Advent Calendar (gunneratttt)



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49469806)



Reply Favorite

Date: November 26th, 2025 11:19 AM
Author: house-broken fantasy-prone hissy fit



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462477)



Reply Favorite

Date: November 26th, 2025 12:14 PM
Author: Onyx roast beef

180. Accurate, intelligent, concise.

The only other thing I would add is that VC switched from a risk sharing model (1970s to around 2008) where partners at the firms would contribute their own capital; to a risk shofting model (2009-Present) where VC funds mostly gather assets and then stuff junk into funds and charge a fee.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462654)



Reply Favorite

Date: November 26th, 2025 12:29 PM
Author: Hideous Casino Jap

Good post. SV is good at displacing shittier industries generally. So GOOGL grew by Ruining newspapers and magazines and shit. FB grew by being better than "scrapbooking" or "yearbooks" or something. AI needs to find something shittier and more useless than AI, and use AI to Ruin it. Discuss.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462684)



Reply Favorite

Date: November 26th, 2025 12:59 PM
Author: crawly aqua stead

Um yeah it has already. A handful of things actually. Junior associates at law firms, entry level accountants, project managers, customer service reps, translators, junior copywriters, early career MBA excel jockeys and junior programmers. It should eventually replace K-12 or at least K-8 teachers, but the unions will slow that process down.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462771)



Reply Favorite

Date: November 26th, 2025 1:08 PM
Author: Trip range halford

I follow the legal industry fairly closely and am not aware of any reduction in associate force stemming from AI. Weirdly, I'm not even aware of anyone *claiming* that, which would be the very first thing I'd do if I had to make normal economically driven layoffs at this point.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462797)



Reply Favorite

Date: November 26th, 2025 9:38 PM
Author: Patel Philippe

(MoFo tranny partner)

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49464510)



Reply Favorite

Date: November 26th, 2025 10:15 PM
Author: ''"'''"''"''''"''

CRISPR came out in 2012 and it brought a lot of investment to pharma. It hasn't produced any healthcare products.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49464642)



Reply Favorite

Date: November 29th, 2025 9:43 AM
Author: NSAM's Schizoid Conspiracy Advent Calendar (gunneratttt)



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49469804)



Reply Favorite

Date: November 26th, 2025 11:13 AM
Author: Deranged nursing home rigpig

So good for us regular people who need to stay employed?

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462458)



Reply Favorite

Date: November 26th, 2025 11:15 AM
Author: Opaque bipolar station

no because it will take a while to take advantage of current capabilities. knowledge workers are still fucked. it's more along the lines of "don't expect some kurzweilian self-perpetuating singularity or massive societal reorganization within the next decade"

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462460)



Reply Favorite

Date: November 26th, 2025 11:16 AM
Author: insanely creepy trump supporter

are pretty much all programmers fucked?

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462466)



Reply Favorite

Date: November 26th, 2025 11:21 AM
Author: Opaque bipolar station

to be clear, everyone is fucked unless you're in the elite tier of your profession. yeah law is a cartel and blah blah blah but the free market will ultimately win (obviously shitlaw and stuff is different, talking about biglaw and the types of "professions" which really just leach off regulations and byzantine corporate structures)

it won't happen overnight though. our Gen is probably still largely fine, maybe we'll have to maek it as policemen at 50 or something. but anybody under 30 and especially those coming out of college now are going to experience a much different job market

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462481)



Reply Favorite

Date: November 26th, 2025 9:36 PM
Author: Louis Poasteur

Wage slaves are always and everywhere fukt 4 lyfe

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49464503)



Reply Favorite

Date: November 26th, 2025 11:18 AM
Author: poppy cumskin

If we could just get it where it would stop hallucinating law then it would be huge. I’ve started trying different threats in prompts and it seems to help a little. Like I’ll tell it a law problem and then say “you’re my genius associate who always double checks his case law because he doesn’t want to go back in the Punishment Closet”

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462473)



Reply Favorite

Date: November 26th, 2025 11:22 AM
Author: free-loading black public bath

Going to be funny when your AI watches Secretary and starts making mistakes on purpose so that you yell at it.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462486)



Reply Favorite

Date: November 26th, 2025 11:23 AM
Author: poppy cumskin

This will absolutely backfire on me and lead to my death

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462489)



Reply Favorite

Date: November 26th, 2025 11:23 AM
Author: misanthropic trailer park gunner

I understand why it might feel that way from your perspective but it isn’t reality. You think it plateaued due to a human cognitive bias toward assuming diminishing returns which doesn’t hold in this case

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462488)



Reply Favorite

Date: November 26th, 2025 11:40 AM
Author: insanely creepy trump supporter

doesn't the LLM model have a theoretical and a practical limit for accuracy and usefulness?

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462537)



Reply Favorite

Date: November 26th, 2025 11:45 AM
Author: misanthropic trailer park gunner

doubt it. that shit will probably just keep SCALING AND SCALING in ways we currently CANT EVEN FATHOM

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462556)



Reply Favorite

Date: November 26th, 2025 11:52 AM
Author: insanely creepy trump supporter

can't tell if this is sarcasm

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462584)



Reply Favorite

Date: November 26th, 2025 12:33 PM
Author: misanthropic trailer park gunner

I’m partially kidding. I don’t think it will be a smooth exponential with unbounded scaling but plateau arguments are almost certainly wrong. I think it will follow q somewhat smooth scaling law until some constraint gets in the way and then architecture changes will cause a phase shift that breaks through the constraints. But it’s already superhuman on most cognitive tasks so idk what you want more from it anyway? Even in the areas experts still outperform it it will usually be due to human representation and semantic quirks. Pretty much all of the things human mathematicians still outperform it on are like that — issues where humans represent things in specific ways so other humans can understand the ideas and also because of human limits in perception and cognitive bandwidth and this throws the model off structurally

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462701)



Reply Favorite

Date: November 26th, 2025 6:34 PM
Author: ,.,....,..,.,.,,,,..,..,.,..,.,.,.,...


Obviously if you can predict what a person will say in any situation, you have a model with human equivalent verbal skills. The LLM training objective isn’t actually limited at that ceiling though, because it also learns to model human errors. Theoretically a very scaled up LLM could be prompted for superhuman behavior, because it can use the learned error model to improve output over its training material.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49464110)



Reply Favorite

Date: November 26th, 2025 11:44 AM
Author: Citrine appetizing stage french chef

are the non-diminishing returns in the (online) room with us right now?

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462552)



Reply Favorite

Date: November 26th, 2025 11:44 AM
Author: misanthropic trailer park gunner

nah they'll come though I promise

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462554)



Reply Favorite

Date: November 26th, 2025 11:47 AM
Author: Citrine appetizing stage french chef

https://x.com/GoySuperstar/status/1993719651495547061

btw i was completely right about models being specifically RL trained to perform on evals, and that's the reason why they're performing increasingly well on evals while overall capabilities are stagnating (duh)

machine learning pumo here didn't believe me, but now we've got the leading guys in the industry admitting it outright

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462562)



Reply Favorite

Date: November 26th, 2025 12:09 PM
Author: Opaque bipolar station

"We know that LLMs can't generalize"

what does this even mean? I don't think anybody ever trained an LLM on Pathfinder 1E rules but it does a decent job of sorting through something that is comically in depth.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462645)



Reply Favorite

Date: November 26th, 2025 12:13 PM
Author: Citrine appetizing stage french chef

it means that LLMs can't generalize, Champ

LLMs can only produce stuff that's within its training distribution. everything outside of its training is Inaccessible for the LLM

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462653)



Reply Favorite

Date: November 26th, 2025 12:22 PM
Author: Opaque bipolar station

yes but give me a specific example of what you are talking about given that LLMs are trained on the entirety of human knowledge and media

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462671)



Reply Favorite

Date: November 26th, 2025 12:34 PM
Author: Citrine appetizing stage french chef

well, LLMs *aren't* trained on the entirety of human knowledge. that's the point. they're trained on the stuff in their training data (a bunch of words that definitely don't include all human knowledge)

for example, in l*w: LLMs do a pretty good job, actually, of evaluating and "reasoning" about cases that have the same structure as other cases in their training data. but any remotely novel kind of legal case, and they totally fail

humans *can* generalize, at least to a far greater extent than LLMs. we can see a completely novel legal case and reason about it pretty well. we can create a mental model of the whole case and think about it accurately and coherently. LLMs can't do that at all

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462704)



Reply Favorite

Date: November 26th, 2025 12:47 PM
Author: misanthropic trailer park gunner

you are right that they aren't trained on "all of human knowledge" although what they are trained on is massive and much larger than what a human could learn in several lifetimes, but there are real knowledge gaps. the idea they cant generalize to novel legal cases is BS though. for instance GPT 4 passed the bar exam and current models have far surpassed gpt 4. its failure mode in law is pretty much the exact opposite of what you said--the issue is that it DOES generalize based on a huge corpus and generate answers that are *plausible* but not authoritative, because it will be something humans decided arbitrarily in case law that has authority, but it could have gone several different ways

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462739)



Reply Favorite

Date: November 26th, 2025 12:55 PM
Author: Citrine appetizing stage french chef

they can't generalize to all legal cases

they just can't do this, you can try it yourself

making shit up that doesn't work and doesn't make sense is not "generalization"

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462760)



Reply Favorite

Date: November 26th, 2025 1:08 PM
Author: Opaque bipolar station

I'm not a lawyer, I need an actual example.

If you're trying to do something truly novel LLMs struggle, but this is like <1% of all knowledge work. To be fair law probably lags behind other fields in data training just because a lot of stuff is just locked behind privatized paywalls, but this will inevitably change over time.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462801)



Reply Favorite

Date: November 26th, 2025 1:35 PM
Author: Citrine appetizing stage french chef

try it with whatever domain you're an expert in. they're already trained on westlaw etc btw

they just can't generalize, it's as simple as that. it's why AI implementation is lagging so much. if LLMs were really capable of generalization, they could be plug and played into virtually any role. but they can't, so they aren't

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462885)



Reply Favorite

Date: November 29th, 2025 9:33 AM
Author: computer online (🧐)

I'm a programmer so automation for 90% of us will probably be the first thing it replaces beyond super low hanging fruit like call centers.

And law, unlike programming for the other 10% who are beyond crudmonkeys like me, doesn't involve creating new stuff -- it is by definition the application and analysis of a set of existing rules. Knowledge professions will die off because there's just rapidly declining value in the knowledge itself, you'll be left with a fraction of the current workforce which basically audits and lightly curates AI generated output.

You should not mistake what people like Karpathy or Ilya are talking about with the replacement of 90% of white collar knowledge profession BS. That's considered the low hanging fruit on the path to AGI, the tricky part as far as that goes is getting rid of cartels, building of trust, and building out the software pipelines to do so. But considering how much companies stand to gain it is a problem that will be solved.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49469789)



Reply Favorite

Date: November 26th, 2025 12:34 PM
Author: misanthropic trailer park gunner

lol this is not what cant generalize means and it’s also wildly false

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462706)



Reply Favorite

Date: November 26th, 2025 1:17 PM
Author: chrome point place of business

GoySuperstar is the go-to source for AI analysis

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462828)



Reply Favorite

Date: November 26th, 2025 12:40 PM
Author: exciting coldplay fan native

its a complete scam if youre talking about the companies actually becoming profitable. its ability to generate pictures and music is really impressive but i dont see how theres big money in that. people will lose interest in that crap if its just computer bs. its useful for "what temperature should i cook a 14lb turkey at" type stuff, but theres no real money in that. itll wipe out a lot of useless white collar jobs that never should have existed i guess, like ib analyst or lawyer. but the idea that its going to lead to robocops staffing mcdonalds drive thrus while proles get their nom on is ridiculous. thats not happening in the next 50-100 years and probably never will in the us. companies like google dropped the ball with making their search engines so shitty (they used to be infinitely useful) and ai fags picked up search from the gutter with their swords.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462722)



Reply Favorite

Date: November 26th, 2025 12:45 PM
Author: Hideous Casino Jap



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462735)



Reply Favorite

Date: November 26th, 2025 12:51 PM
Author: misanthropic trailer park gunner

lol what it can do is already more impressive than robocops staffing mcdonalds

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462748)



Reply Favorite

Date: November 26th, 2025 12:54 PM
Author: exciting coldplay fan native

like what?

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462756)



Reply Favorite

Date: November 26th, 2025 1:01 PM
Author: Trip range halford

Lol. Look I'm not trying to go full peter thiel here (I actually think the tech is impressive, will likely end up an invention whose importance/disruptiveness is roughly the order of the internet, and could still possibly be a truly transformative enlightenment/industrial revolution-level innovation), but watching fags who made a 1480 on the SAT talk about how actually, it's not disappointing at all that we don't have flying cars and have generally, in everyone else's view, vastly underperformed the widely shared technological predictions of our youth -- it's just as impressive that a chat website can spit out partially wrong screeds about information collected from the internet! -- is really weird, and suggests something about these people aside from just being credulous futurists. Maybe a weirdly high fear of death? Of disappointment in society? Of an over-valuing of society over smaller social units (family) and overinvestment in the belief that we were on the cusp of transformation? I dunno.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462777)



Reply Favorite

Date: November 26th, 2025 1:05 PM
Author: misanthropic trailer park gunner

but what if we end up building flying cars soon with all of the extra cognitive power supplied by AI?

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462786)



Reply Favorite

Date: November 26th, 2025 1:25 PM
Author: Trip range halford

Then that's great. Right now I've got a website that gave me some just-okay protein shake recipes (certainly not as good as some I got from trashy listicle men's-health sites) and, when it came time to brine the turkey this year, I went to NYT instead.

Are you maybe a few crucial years younger than me? Were you not around for Nanotechnology?

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462854)



Reply Favorite

Date: November 26th, 2025 1:33 PM
Author: misanthropic trailer park gunner

nanotech was based on speculative capability claims, many of which required breakthroughs in physics, manufacturing and materials that never occurred. It was science-first speculation whose engineering pathways never converged. AI is the opposite. It's trajectory is grounded in empirical scaling already demonstrated across modalities and industries

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462879)



Reply Favorite

Date: November 26th, 2025 1:42 PM
Author: Trip range halford

It was also the domain of a set of fields and industries that did not have--and I am giving a serious and considered numerical estimate here--2% of the fast-talking, pomade-haired, pin-suited straw-hatted hucksterism of SV.

Imagine if The Music Man got to shill techno-utopia as a replacement for the fent epidemic, instead of shitty brass-band instruments as a replacement for pool halls. That's what we're dealing with here.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462909)



Reply Favorite

Date: November 26th, 2025 2:14 PM
Author: misanthropic trailer park gunner

techno utopia is BS yes. doesn't mean AI isn't legit.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49463046)



Reply Favorite

Date: November 26th, 2025 1:17 PM
Author: Opaque bipolar station

"its useful for "what temperature should i cook a 14lb turkey at" type stuff, but theres no real money in that."

that's retarded. that's like the people who said "but there's no money in search lmao" 30 years ago and google now prints dizzying sums of money.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462823)



Reply Favorite

Date: November 26th, 2025 12:48 PM
Author: Soggy parlour

It’s just too spammy right now. Everything feels slightly off

.



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462741)



Reply Favorite

Date: November 26th, 2025 12:49 PM
Author: razzle-dazzle newt

many products are scams for sure. However there are lots that are incredibly useful. I use the following on a daily basis:

NotebookLM

AI call notetaking (Otter or similar)

Claude/ChatGPT/Gemini projects for doing first drafts of stuff where I simply want to drop in voluminious notes/documents and let it do what it do

More broadly, AI is going to be game-changing in may fields, particularly science. It analyzes the literature on a subject and analyzes experimental data much more quickly than previous methods. I'm also told it can design novel drugs/enzymes/ etc. I expect the pace of scientific advances to increase, with unpredictable results.

I expect the capabilities of AI within current use cases will get better, and that new use cases will emerge. I still think it's going to eliminate the need for a lot of human mental work, with unpredictable consequences

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462743)



Reply Favorite

Date: November 26th, 2025 1:04 PM
Author: Hideous Casino Jap

good point imo. scientists are lazy....

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462783)



Reply Favorite

Date: November 26th, 2025 1:09 PM
Author: Hot principal's office jew

It's interesting how cynical so much of the board is about AI when it seems obvious to just about everyone else that this technology will transform life as we know it within the next 5-10 years. I know it's the tone of the board to knock down everything conventional wisdom supports, but that instinct is wrong in this case.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462806)



Reply Favorite

Date: November 26th, 2025 1:14 PM
Author: Hideous Casino Jap

There is no "conventional wisdom." It's just people extrapolating off of the internet. But you could buy stuff on the internet in 2000. It didn't take much imagination to imagine what it would turn into.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462816)



Reply Favorite

Date: November 26th, 2025 1:16 PM
Author: misanthropic trailer park gunner

conventional wisdom doesn't even support AI though. there are midwits all over substack and reddit constantly saying these same things that AI is totally nbd and has plateaued and whatever else. xo just always trends cynical on everything and in this case i'd argue they are aligned with other pseud boards like reddit and substack

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462822)



Reply Favorite

Date: November 26th, 2025 1:18 PM
Author: Hideous Casino Jap

The people who are decisively wrong are the ones who think it's dumb compared to humans. Those people are idiots afaik. But this is a separate from the question of Monetization imho.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462835)



Reply Favorite

Date: November 26th, 2025 1:20 PM
Author: misanthropic trailer park gunner

yeah I agree with that. there are multiple arguments that got brought up itt including people saying the tech itself has plateaued. I agree that most AI startups are flame and that even current profitable AI companies are basically just getting subscription fees right now.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462841)



Reply Favorite

Date: November 26th, 2025 1:18 PM
Author: Opaque bipolar station

because they're lawyers

one of my favorite all time xo poasts I lost track of many years ago was a pretty big thread where everyone shit on the outrageous sum of money facebook spent acquiring instagram. it was 1 billion dollars and instagram would go on to be the very reason facebook thrived even as its core platform became entirely boomers. in fact primary thing facebook was actually good at was picking which companies to buy early on (instagram and whatsapp have been insane returns on investment)

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462833)



Reply Favorite

Date: November 26th, 2025 1:29 PM
Author: chest-beating curious becky

"AI" is just freshman linear algebra, sophomore statistics and $1 trillion in wasted electricity.

Notice the leaps these charlatans make in describing their product.

AGI (like the stat in warcraft) used to be AI used to be Machine 'Learning' used to be Linear Regression.

Name ONE (uno) (1) novel invention or one nontrivial scientific discovery that is for the benefit of humanity and is not sophistry.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462865)



Reply Favorite

Date: November 26th, 2025 1:36 PM
Author: Citrine appetizing stage french chef



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462889)



Reply Favorite

Date: November 26th, 2025 1:37 PM
Author: misanthropic trailer park gunner

I suppose you also think jet engines are "just" sophomore fluid dynamics and thermodynamics, CPUS are "just" semiconductor physics, and MRI is "just" 19th century electromagnetism with 1940s NMR tacked on right?

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462891)



Reply Favorite

Date: November 26th, 2025 1:50 PM
Author: chest-beating curious becky

(charlatan)

https://www.youtube.com/watch?v=Fk0QC7wf2do

heres your average tech evangelist

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462943)



Reply Favorite

Date: November 26th, 2025 2:09 PM
Author: misanthropic trailer park gunner

I agree with you that tech in general isn't necessarily for "the benefit of humanity". we would be fine without most of it, arguably better off. but that doesn't mean it should be trivialized. something can be genuinely impressive and hard to pull off and not make us "better off".

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49463031)



Reply Favorite

Date: November 26th, 2025 1:30 PM
Author: crawly aqua stead

I do think the “agents” moment is pretty open ended at this point. It’s conceivable that models and product evolution will get to a point that an agent can be easily trained to do a lot of useful stuff, but as of today it’s a sufficiently labor intensive and hyperiterative guess and check type of process that it’s not likely to spread far outside of AI evangelist circles. To me the parallels to the e-commerce adoption process between 99-2012 are still the best analogy and though it might not be quite as slow a process I suspect most of the grand visions that evangelists love to project is a 2030s timeline after some correction and on the far side of a period where both the models and app companies prove they can survive and chart a course to durable and substantial profitability, just like we saw with dotcoms/ecommerce and social media and cloud and rideshare apps and streaming.

Just comparing to ecommerce there were these secondary and tertiary problems that had to be properly solved re logistics infrastructure, global supply chains, payment systems, digital marketing and advertising, compute infrastructure etc. that spurred all sorts of massive innovations and spawned entirely new lines of business that were even more profitable and strategic, and that’s before you even get into the massive benefit that accrued to a company like Amazon when the iPhone came along/was so good and mobile quickly became the primary use case. We’re in the 99-00 phase of the investment hype cycle but AI companies are already working out various solutions to eg the energy demand problem as a macro limitation, but it’s problems along those lines that need to be credibly and durably solved to unlock adoption and implementation levels that will actually justify the current insane levels of investment.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462867)



Reply Favorite

Date: November 26th, 2025 1:35 PM
Author: flesh autistic lay elastic band

AI is only going to get better. GPT 5 seems way better than GPT 3. This is like saying in 1993 that you don't see how the internet will really change commerce that much because all you're able to do is send messages without the post office. It's only been around for a few years.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462887)



Reply Favorite

Date: November 26th, 2025 1:36 PM
Author: frisky box office multi-billionaire

TiTcR

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49462890)



Reply Favorite

Date: November 26th, 2025 2:18 PM
Author: Ultramarine aggressive fortuitous meteor



(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49463056)



Reply Favorite

Date: November 26th, 2025 5:32 PM
Author: iridescent toaster sanctuary

I'm pretty dumb and retarded when it comes to this shit (honestly), but I've recently started using AI way more.

It's certainly useful but it still seems a ways off... Right now it's primarily feedback loop is to validate the user and there's a lot of limitations with the paid versions in terms of what it will generate, how it generates it, etc.

For ex. just creating a basic manual regarding expanding into a new practice area required hours of prompting and tweaking and dealing with the website's bullshit to get it to work. Tried this with paid version of both claude and chatgpt and each had their own specific limitations.

But the bigger concern is the constant validation of whatever the user is feeding it since their whole goal is to drive engagement. In some ways it dumbs down the product.

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49463841)



Reply Favorite

Date: November 29th, 2025 9:05 AM
Author: OYT Magnus ( )

https://wccftech.com/quantum-computing-will-pop-the-ai-bubble-claims-ex-intel-ceo-pat-gelsinger/

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49469760)



Reply Favorite

Date: November 29th, 2025 9:22 AM
Author: computer online (🧐)

Quantum computing has been a couple years away for like 15 years or whatever

(http://www.autoadmit.com/thread.php?thread_id=5802800&forum_id=2Ã#49469775)