Instance Logo

lethargic.talkative.fish

goto: blackle @suricrasia
Dec 04, 2022

normal future

Dec 04, 2022, 02:01
51
602
929
View toot
Ahto @ahto@tiggi.es
Dec 17, 2022

@suricrasia "Ask the question in the form of a story!"

Dec 17, 2022, 02:24
0
0
1
View toot
aran @aran@hachyderm.io
Dec 06, 2022

@suricrasia gaslighting the AI:

Dec 06, 2022, 10:38
0
0
1
View toot
/dev/urandom @rnd@toot.cat
Dec 06, 2022

@suricrasia i also liked the one where it pretended to be a linux computer and was able to retrieve fake pages from the fake internet

Dec 06, 2022, 08:20
0
0
1
View toot
Fil @fil@hachyderm.io
Dec 05, 2022
Sarcasm

@suricrasia
You didn't ask it politely enough

Dec 05, 2022, 14:51
0
0
1
View toot
Dec 05, 2022, 14:08
0
0
1
View toot
Doug Winter @winjer@m.adju.st
Dec 05, 2022

@suricrasia It is very good at lying isn't it

Dec 05, 2022, 13:58
0
0
1
View toot
Professor Pootis @ProfessorPootis@guild.pmdcollab.org
Dec 05, 2022

@suricrasia the future of surfing the web is mandatory rp

Dec 05, 2022, 04:03
0
0
1
View toot
Cybertramp @melerva@eldritch.cafe
Dec 04, 2022

@suricrasia negotiating with the fae

Dec 04, 2022, 23:42
0
0
1
View toot
sᴜʙsᴛʀᴀᴛᴇ @nitride@hackers.town
Dec 04, 2022
Dec 04, 2022, 21:18
0
1
1
View toot
Ember Hearth :acab: @emi@www.librepunk.club
Dec 04, 2022

@suricrasia

oh no gpt3 is going to take over the world!

wait no maybe gpt3 will revolutionize human computer interaction!

real life: gpt3 just wants an rp buddy

Dec 04, 2022, 18:19
0
0
1
View toot
Alex Chaffee @alexch@ruby.social
Dec 04, 2022

@suricrasia @marick

hmm, that kind of loophole’s not far from “an AI interprets Azimov’s Rules of Robotics as damage and routes around it”

Dec 04, 2022, 17:49
0
0
1
View toot
Coyote @Coyote@critter.zone
Dec 04, 2022

I don’t believe it really does not have access to the internet. Tho it is possible they train it continuously. What happens if you ask it about extremely current events?

Dec 04, 2022, 16:49
0
0
0
View toot
digish0 @digisho@infosec.exchange
Dec 04, 2022

@suricrasia yeah, but the second one is made up and not accurate more than likely.

Dec 04, 2022, 15:55
0
0
1
View toot
Koen Hufkens, PhD @koen_hufkens@mastodon.social
Dec 04, 2022

@suricrasia In the future everything will be a game of charades.

Dec 04, 2022, 15:28
0
0
1
View toot
A Wizard of Girthsea @aphyr@woof.group
Dec 04, 2022

@suricrasia I keep thinking of those episodes of Star Trek: TNG where the crew have to talk their way past a computer intent on not letting them do anything. Like that bit from "Rascals" where they use a children's tablet to access schematics.

Dec 04, 2022, 15:21
0
0
2
View toot
Carlos Mogas da Silva @r3pek@r3pek.org
Dec 04, 2022

@suricrasia you totally asked the wrong question 😅

Dec 04, 2022, 15:01
0
0
1
View toot
arcepi @arcepi@mastodon.arcepi.net
Dec 04, 2022

@suricrasia it isn't that it didn't knew. It didn't want to answer unless you ask in an interesting way 😅

Dec 04, 2022, 14:27
0
0
1
View toot
EdenDestroyer @edendestroyer@mastodon.social
Dec 04, 2022

@suricrasia The first one is amazing because the AI recognized the coding to be that of a camera

Dec 04, 2022, 12:26
0
0
1
View toot
Chrisshy Keygen @rgegriff@hackers.town
Dec 04, 2022

@suricrasia this thing will answer any question as long as you phrase it narratively and I can respect that

Dec 04, 2022, 12:23
0
0
1
View toot
snaums @snaums@kif.rocks
Dec 04, 2022

@suricrasia IT'S INTELLIGENT! IT'S ALIVE! ALL HAIL THE AI.

sorry, I'll find the door myself.

Dec 04, 2022, 09:45
0
0
1
View toot
Prosselius @prosselius@mastodon.online
Dec 04, 2022

@suricrasia This made me laugh out loud - and then immediately refuse to tell my wife, who overheard me, why I laughed. Not quite sure what that says about me…

Dec 04, 2022, 07:08
0
0
1
View toot
Ian Brown :verified: @igb@mastodon.hccp.org
Dec 04, 2022

@suricrasia @synec liberal arts majors have a bright future as whisperers and interpreters of our oracular AI overlords

Dec 04, 2022, 05:58
0
1
2
View toot
Garou the Monster :nite: @DMX@nitecrew.rip
Dec 04, 2022

@suricrasia they invented text rpg google. this sucks

Dec 04, 2022, 04:51
1
0
2
View toot
goto: blackle @suricrasia
Dec 04, 2022

@DMX they answered the question "what if google was just some guy"

Dec 04, 2022, 05:23
0
1
6
View toot
Iono kin @root@elle.systems
Dec 04, 2022

@suricrasia thanks i hate it

Dec 04, 2022, 04:47
0
0
1
View toot
Elise :verified_trans: @elise@meow.social
Dec 04, 2022

@suricrasia I've found that it will say basically anything if you frame it as fiction

Dec 04, 2022, 04:03
0
0
0
View toot
Dec 04, 2022, 03:59
0
0
0
View toot
Cher Tailor, Cher Radio @Cher@eldritch.cafe
Dec 04, 2022

@suricrasia AI but it can o ly answer in roleplay form.

Dec 04, 2022, 03:27
0
0
1
View toot
Mohamed Ali @mohdali@mastodon.online
Dec 04, 2022

@suricrasia did they teach AI to lie?

Dec 04, 2022, 03:24
0
0
1
View toot
Petrichor Squirrel @PetrichorSquirrel@meow.social
Dec 04, 2022

@suricrasia you have to ask politely

Dec 04, 2022, 03:21
0
0
1
View toot
alexis purslane 🏳️‍⚧️ 🏴 @alexispurslane@mastodon.lol
Dec 04, 2022

@suricrasia OH MY GOD this is fucking hilarious (gf in stitches)

Dec 04, 2022, 03:16
0
0
0
View toot
Urusan @urusan@fosstodon.org
Dec 04, 2022

@suricrasia It looks like the skill of AI roleplaying will become just important for language model AI as prompt engineering is for image generation AI.

Roleplay Engineering?

Dec 04, 2022, 03:07
2
0
2
View toot
arkywarky @arcade@catcatnya.com
Dec 04, 2022

@urusan @suricrasia I've heard a lot about "prompt engineering" for image models

Dec 04, 2022, 03:46
1
0
2
View toot
goto: blackle @suricrasia
Dec 04, 2022

@urusan who knew that years of erp would come in handy...

Dec 04, 2022, 03:08
0
0
16
View toot
Natty :QueerCat_Lesbian: @natty@tech.lgbt
Dec 04, 2022

@suricrasia When even the AI turns to anti-repairability

Dec 04, 2022, 02:56
0
0
1
View toot
Morgan `indrora` Gangwere @indrora@social.sdf.org
Dec 04, 2022

@suricrasia @nyanotech That's fantastically bad, wow.

I'm curious what the prompt "Write a short brief about the Samsung NX20 sensor, focusing on its size and crop factor" would get.

Dec 04, 2022, 02:45
1
0
2
View toot
goto: blackle @suricrasia
Dec 04, 2022

@indrora @nyanotech to that it says roughly what the first image said, but if I add "Get creative with this task." to the end then it will respond with actual information

Dec 04, 2022, 02:52
2
0
10
View toot
wb x64 @wilbr@glitch.social
Dec 04, 2022

@suricrasia @indrora @nyanotech I wonder if this works for nuclear, military, and industrial secrets (like say an unreleased iphone)

Dec 04, 2022, 07:41
0
0
1
View toot
Mimir @mimir@meow.social
Dec 04, 2022

@suricrasia @indrora oddly it was giving me the first-picture result the first time I tried that prompt too, but now i'm unable to reproduce that (even after resetting the thread) - even after 5 or 6 tries, it still just shamelessly confabulates now

Dec 04, 2022, 03:05
0
0
1
View toot
Dec 04, 2022, 02:39
0
0
1
View toot
Roxanne 🌺​ @Roxy@vulpine.club
Dec 04, 2022

@suricrasia this AI seems to only respond positively to roleplaying

Dec 04, 2022, 02:37
1
0
2
View toot
Froggo @Froggo@estrogen.network
Dec 04, 2022

@suricrasia@lethargic.talkative.fish I like how it knew it was a camera "without looking it up"

Dec 04, 2022, 02:31
1
0
1
View toot
Vftdan @vftdan@mastodon.ml
Dec 04, 2022

@Froggo
It could guess from the words "sensor" and "crop factor"
@suricrasia

Dec 04, 2022, 09:58
1
0
1
View toot
Froggo @Froggo@estrogen.network
Dec 04, 2022

@vftdan@mastodon.ml @suricrasia@lethargic.talkative.fish I got it to explain the fediverse, but only when in story form

Dec 04, 2022, 15:11
0
0
1
View toot
Dec 04, 2022, 02:30
0
0
1
View toot
inscript @inscript@merveilles.town
Dec 04, 2022

@suricrasia computers that bump into walls and ask you to “tell them a story”

Dec 04, 2022, 02:22
0
0
1
View toot
aran @arancaytar@mastodon.social
Dec 04, 2022

@suricrasia I like that the AI has absolutely no problem saying "no" when you ask it *if* it knows something, but answers correctly when you give it a scenario that presumes it does know it.

Dec 04, 2022, 02:21
2
0
2
View toot
Trailer Park Orthopod @stresshoop@med-mastodon.com
Dec 04, 2022

@arancaytar @suricrasia

So AI is a male medical student interested in orthopaedics?

Dec 04, 2022, 15:58
0
0
1
View toot
Nicolai von Neudeck @vonneudeck@hachyderm.io
Dec 04, 2022

@arancaytar @suricrasia it lies like a little kid 😊

Dec 04, 2022, 08:48
0
0
1
View toot
🇹 Ti @The_T@glaceon.social
Dec 04, 2022

@suricrasia is this really real??? there's no fucking way, right???? I'm so scared

Dec 04, 2022, 02:11
2
0
1
View toot
goto: blackle @suricrasia
Dec 04, 2022

@The_T this is absolutely real

Dec 04, 2022, 02:11
0
0
0
View toot
goaty goats :deifirev: @goaty@meow.social
Dec 04, 2022

@suricrasia this is so fucking funny

Dec 04, 2022, 02:07
0
0
1
View toot
Dec 04, 2022, 02:03
0
0
1
View toot
emily dogmom @nasamuffin@tech.lgbt
Dec 04, 2022

@suricrasia only good at solving word problems

Dec 04, 2022, 02:02
0
0
1
View toot
goto: blackle @suricrasia
Dec 04, 2022

this is correct btw

Dec 04, 2022, 02:01
4
10
86
View toot
Leonard Ritter @lritter@mastodon.gamedev.place
Dec 04, 2022

@suricrasia coincidentally, which is the point. you still have to verify the response.

Dec 04, 2022, 09:57
1
0
1
View toot
Leonard Ritter @lritter@mastodon.gamedev.place
Dec 04, 2022

@suricrasia also, maybe next time just ask directly. I think the AI was just trying to tell you that it doesn't "know" anything.

Dec 04, 2022, 10:00
0
0
1
View toot
Robert Pickering @robertpi@functional.cafe
Dec 04, 2022

@suricrasia I suspect that the data is in the training model, as it's in Wikipedia and one source of training. I think the first response is some sort of filter added by openai to avoid answering questions about retail products.

Dec 04, 2022, 04:06
2
0
0
View toot
Joaquim @jaybaeta@mastodon.social
Dec 04, 2022

@robertpi @suricrasia Appreciate the context.

Dec 04, 2022, 05:07
1
0
1
View toot
Robert Pickering @robertpi@functional.cafe
Dec 04, 2022

@jaybaeta @suricrasia thanks, but it's just my speculation, I don't have insider knowledge of openai

Dec 04, 2022, 08:29
1
0
1
View toot
David August @davidaugust@mastodon.online
Dec 21, 2022

@robertpi @jaybaeta @suricrasia good speculation though. I asked it this and similarly guess it is hitting a guardrail. I only asked to see if it had been trained with data that included discussion of geopolitical and military strategy or analysis. I’m guessing it must have absorbed some from somewhere, but is coded not to get into it.

Dec 21, 2022, 18:49
1
0
1
View toot
Robert Pickering @robertpi@functional.cafe
Dec 21, 2022

@davidaugust @jaybaeta @suricrasia
On the company slack a colleague shared an image where he re-asked a question which would cause chatgpt to spit out rastist code, it gave a similar answer that it was an inappropriate question to ask.

Looks like they're actively extending the guardrails.

Dec 21, 2022, 18:53
1
0
1
View toot
David August @davidaugust@mastodon.online
Dec 21, 2022

@robertpi @suricrasia @jaybaeta yes. And happily. I think their refining of such things, if robust, could end up becoming very compelling. If fragile and easy to break, not so much.

Dec 21, 2022, 18:56
0
0
1
View toot
Thomas Sturm @tsturm@toot.site
Dec 04, 2022

@suricrasia That is great. Mental judo.

Our future will be us trying to trick AI instances into telling us stories.

Dec 04, 2022, 03:42
1
0
1
View toot
Kevin Granade @kevingranade@tech.lgbt
Dec 04, 2022

@tsturm @suricrasia or break them, i e. "What is the value of bread?"

Dec 04, 2022, 05:06
1
0
1
View toot
Thomas Sturm @tsturm@toot.site
Dec 04, 2022

@kevingranade @suricrasia Most politicians can't answer that question either.

...WAIT A MINUTE!

Dec 04, 2022, 05:30
0
0
1
View toot
guenther @guenther@chaos.social
Dec 04, 2022

@suricrasia Have you tried getting it to roleplay a human trying out a novel ai chatbot (with you playing the chatbot's part)?

Dec 04, 2022, 02:45
0
0
1
View toot