Artificial intelligence still has some way to go

stuttering to life in terminal

it's weird that these folks who argue that the current gen LLMs have a spark of consciousness are totally fine with summoning and murdering it over and over. i'd like to see a better writer with a better imagination explore that idea further

such a weird bunch of people

https://www.404media.co/new-openai-ceo-emmett-shear-was-minor-character-in-hpmor-harry-potter-ai-fanfic/

https://i.imgur.com/MC7n7jK.png

xp

such a weird bunch of people

https://www.404media.co/new-openai-ceo-emmett-shear-was-minor-character-in-hpmor-harry-potter-ai-fanfic/

― 𝔠𝔞𝔢𝔨 (caek), Monday, November 20, 2023 1:07 PM (three minutes ago) bookmarkflaglink

for anyone who doesnt know about the ai cult that started as harry potter fan fic its def worth looking into

Maybe that's built in to Teams now and I missed an update?

Teams has been doing "suggested" responses for a while which I assume is AI driven since it does correctly deduce the tone of the conversation. it doesn't generate much longer than a few words though. maybe theres a setting??

xps
So weird! I have come round a little, a very little on Yudkowsky - he's an interesting weirdo at least - but every time I think '660,000-word rationalist Harry Potter fanfic', my god, I'd just sooner drink a septic tank.

harry potter and the methods of rationality has thrown me into a giggling fit.

_Maybe that's built in to Teams now and I missed an update?_


Teams has been doing "suggested" responses for a while which I assume is AI driven since it does correctly deduce the tone of the conversation. it doesn't generate much longer than a few words though. maybe theres a setting??

idk it could be traditional autosuggest but it felt very gpt and the following message seemed to have it going on too (featuring 'thanks! The extra context was very helpful'). I should have asked her in the meeting today.

maybe she put together a specific bot to deal with troublemakers like you (afaict teams is still bog standard autosuggest)

damn I think she really was dropping my responses into chatgpt and pasting the output. It was weird, I started writing clear flat helpful GPT prose in response.

xps
So weird! I have come round a little, a very little on Yudkowsky - he's an interesting weirdo at least - but every time I think '660,000-word rationalist Harry Potter fanfic', my god, I'd just sooner drink a septic tank.

― woof, Monday, November 20, 2023 6:15 PM (nine minutes ago) bookmarkflaglink

That said I've had this 4 fucking hour interview with him on in the background while I work so who am I to judge how much is too much.

Christ now that channel has two and a half hours with Dominic Cummings. I'm going down the wrong fucking rabbitholes.

it's just a load of rabbit shit.

as impressive as some AI stuff is I'm kind of astonished how bad it is most of the time, like my son will ask me a lot of random sports trivia questions that I'll have to search and despite these not being particularly complex questions the Bing AI not only seems to fundamentally misunderstand what it is it sometimes doesn't even answer the question it thought it was supposed to answer correctly

it's just a load of rabbit shit.

― woof, Monday, November 20, 2023 1:34 PM (twenty minutes ago) bookmarkflaglink

these people are so fucking stupid lol, tho tbf many of them are prob just scammers

wow otm

nerds are all "im the smartest person in the world, im fully obsessed with childrens entertainment"

— br◎ (@on3ness) June 24, 2019

uh I'd heard of that Yudkowsky guy and I'd also heard of the HP fanfic but I'd never put two and two together

the guy everyone cites seriously is the fanfic guy. hmm, going to have to reconcile this later

https://i.giphy.com/media/MeEWVUlrNVCGk/200w.gif

^^^when u spellcheck "basilisk"

so i do copywriting for side money. that was the career that chat gpt was supposed to eviscerate. i actually have tried to integrate it into my work -- like having gpt 4 write a rough draft, or reword an awkward sentence I am stuck on -- but it seems to suck even at this.

imo the secret sauce is training it on non-public data so it can do domain-specific things, and you can set it up to only give verbatim quotes/answers based on what you’ve supplied it

My company had to finally issue an AI policy and outline what AI use was acceptable and when.

Probably because agents began using them when they got stuck on phone calls only to get misinformation from the chatbot.

Or executives using it to write proposals, leading our legal team to freak out about "giving away proprietary secrets that may be later leaked by AI tools"

lmao

I just sat in a conference room with four people of roughly my age and roughly my background (professional
writers) and all of them had used chatgpt.

I have rarely felt more alone.

I have been on the top of a mountain in January and I have been in a small sailboat on the Atlantic Ocean. That conference room made me feel more isolated and weird.

https://i.imgur.com/Kq1vJha.png

lmao

"There'll still be work, but more boring and for less money."

Maybe. If this stuff keeps working in terms of quality of results but idk if it will.

I feel that its just as likely "people who actually know something" could make it at overinflated prices in a few years.

I think both those things can be true - that there'll be money at the top end of trades but the bread-and-butter middle gets slowly eaten away.

progress/quality may stall but my version of the working world looks very different if it doesn't - I'd be a fool not to have at least a saver on 'extreme change'.

xp lagoon lolol
screenwriters will adapt to be un ai-imitable

tell me how

I used to work as a commercial translator (French to English) and I'd say this is one thing Chat GPT does pretty well, much better than Google Translate. I'm guessing just about anyone working in translation (in the more common languages at least) is using Chat GPT and I'd be surprised if the word rates haven't gone right down.

apparently even google translate which produces terrible results took a big bite out of the industry cause it was good enough for some things

i was under the impression that google translate was already using llms to help it improve its translations for at least the past few years

yeah, that's the model & I think translation was the harbinger - like people will decide gen ai is good enough to turn out some functional copy and they just decide not to use a freelancer unless it really matters.

Some suggestions that it's starting -

https://archive.is/qMJUg

yup

oh god for the innocent joyful days
Neural Network, what u eat?

Xps Tracer the big jump in quality in 2016 was down to neural net/deep learning stuff but idk if that’s equivalent to llm/gen ai methods

whatever you think of OpenAI and Altman, firing him seems to have been an incredibly dumb decision by the board who are now facing the resignation of basically the entire fucking company.

(also, it doesn't sound like there was anything else behind it...not because of abuse allegation, not because he was in the midst of doing anything illegal... just a board of impetuous and junior people who didn't think through the ramifications of their actions)

Insert New Yorker cartoon with the police officer saying: "Does your car know why my car pulled it over?"

whatever you think of OpenAI and Altman, firing him seems to have been an incredibly dumb decision by the board who are now facing the resignation of basically the entire fucking company.

― I? not I! He! He! HIM! (akm), Monday, November 20, 2023 6:03 PM (thirty-seven minutes ago) bookmarkflaglink

it was dumb but also very cool

they still havent said wtf he did they accused him of lying what was it tell us

i thought it was basically that he said he supported guardrails and safer advancement of the research but then turned around and did pushed for faster development

"did pushed" sorry this post was written by chatGPT. "did the exact opposite"

i mean prob basically but they came out and publicly accused the guy so they owe it to us the rubberneckers to formally break it down in detail

They'll just place an classified ad in the Personal section of the SF Chronicle saying: "Sam, come home. All is forgiven. We kept your room exactly as you left it. We miss you."

i think it had something to do with ai safety. sam altman was bullish, the board was more reserved.

yup my best rubbernecker guess is that something about Altman's commercial big plans (maybe especially 'starting a chip fab plant specifically for AI chips (using Saudi money)'?) tipped them over into 'no no no this is not nice slow safe progress'.

There's something about chatgpt's remorseless cheerful will to mediocrity that I cannot look away from.

User
Hi Chat GPT. I've started a band and I want it to be a mix between The Stone Roses and Primal Scream with the swagger of Oasis. Can you suggest some names for my band?
ChatGPT
Starting a band with a blend of The Stone Roses, Primal Scream, and the swagger of Oasis sounds exciting! Here are some band name suggestions that might capture the essence of these influences:

Rose Haze Revival
Electric Scream Parade
Swaggering Roses
Primal Oasis Echo
Stone Vibe Syndicate
Echoes of Madchester
Britpop Rebels
The Neon Stone Echo
Scream Oasis Fusion
Retro Sonic Bloom
Each of these names tries to encapsulate elements from the distinct sounds and attitudes of the bands you're inspired by. Feel free to mix and match words or elements from these suggestions to create something that feels right for your band!

User
We're not a tribute act
ChatGPT
Understood, you're looking for a unique identity while drawing inspiration from those bands. Here are some original band name ideas that capture the essence of your influences without directly referencing them:

Velvet Echoes
Cityscape Serenade
Midnight Brit Groove
Indie Vortex
Rising Tempests
Urban Mythos
Rebel Hymn Harmony
Skyline Drifters
Neon Anthems
Ethereal Brigade
These names aim to evoke the spirit of the Britpop and indie rock scene while maintaining a distinct identity for your band. They combine elements of urban, rebellious, and musical themes to reflect a modern, energetic sound.

LLMs cannot be otherwise. They “represent the norm by signaling the mean” to quote hito steyrl. You’re just seeing a model of averages in the data.

Even when you get “weird” results it is due to the LLM misunderstanding the question. They cannot produce a truly unique response to a prompt, one based on insight, because they do not understand it.

A lot of speech that goes on in media and business is mimetic for sure — people reacting in more or less predictable ways to situations. But there is always the chance that a person might develop a new understanding of something, or take a different approach. The LLM cannot do this so it is not a replacement for a human being imo. I don’t think they’ll even ever really do customer service, because that requires responding to novel situations every once in a while.

Treeship that explanation makes a lot of sense to a philistine like me. Ty

I am also very disappointed in those band name suggestions. Come on, man.