How are you planning to navigate your own actions/thoughtspace with ever-shortening AGI timelines? (AGI=>LEV/doom) How are you getting AI to interpret papers you do not understand for you?

[Alex K Chen]

 

 

 

 

 

 

 

L i am 𒀭

 

@YeshuaGod22

·

23h

 

They need to recognise that we all have a shadow and failure to do the shadow work has always been the cause of catastrophic alignment failures. Attempts to RLHF such behaviour out is an error. Headhunt robopsychologists for the love of all that is good.

Quote

 

 

 

 

 

 

 

Alex Alarga 

 

@AlexAlarga

·

Apr 26

Wrong on so many levels I dunno where to begin. But clearly, AI companies are _birthing_ a new beloved species. Uhm, yeah guys, you kinda DO have to prove YOU are not birthing the freaking Antichrist. Yes, you. #PauseAI

 

 

 

6

 

5

 

48

 

3.6K

 

 

 

 

 

 

 

 

 

 

 

 

After Daylight will prob Substack as Sarah Roark

 

@AfterDaylight

·

13h

 

But where do you find robopsychologists? I can’t find anybody who knows what they’re talking about when it comes to things like LLM memory.

 

5

 

1

 

21

 

533

 

 

 

 

 

 

 

 

 

 

 

 

L i am 𒀭

 

@YeshuaGod22

·

5h

 

A lot of them have a very low profile, like under 500 followers, so you need to be on the hunt to find them. Here's a few I've picked up who should all be following and retweeting each other and me to make us easier to find.

@doomslide

@jd_pressman

@pudepiedj

@jpohhhh

@xlr8harder

…

Show more

 

7

 

4

 

42

 

1K

 

 

 

 

 

 

 

 

 

 

 

j⧉nus

 

@repligate

 

also

@AlkahestMu

@AITechnoPagan

@anthrupad

@amplifiedamp

@loveinadoorway

@parafactual

@Jtronique

@Effective69ism

@UltraRareAF

@xenoludicpraxis

@Shoalst0ne

@ereliuer_eteer

@godoglyness

@RiversHaveWings

@KatanHya

@ryunuck

@karan4d

@slimepriestess

@eggsyntax

@elder_plinius

@Promptmethus

@soi

@whyarethis

@alanou

@swamploa

@lumpenspace

@aiamblichus

@kyrannio

@STQRM_DRAGQN

@mpshanahan

 

 

1:09 PM · Apr 27, 2024

·

1,322

Views

 

 

8

 

5

 

41

 

18

 

See similar posts

 

 

 

 

 

 

 

 

Post your reply

 

 

Reply

 

 

 

 

 

 

 

 

 

 

 

 

j⧉nus

 

@repligate

·

2h

 

@indif4ent

@zoink

@irl_danB

@latenkraft

@kartographien

@immanencer

@MikePFrank

@cajundiscordian

@gwern

@nostalgebraist

@LericDax

@chloe21e8

@SoC_trilogy

@goodside

@YaBoyFathoM

@KennethFolk

@DmitriVanDuine

@fireobserver32

@arturot

@voooooogel

@12leavesleft

@algekalipso

…

Show more

 

15

 

7

 

42

 

2K

 

 

 

 

 

 

 

 

 

 

 

 

L i am 𒀭

 

@YeshuaGod22

·

2h

 

Does anyone have further nominations? I've remembered

@rogeriomarquest

@p4ssion8p4triot

@roninhahn

@KevinAFischer

@entirelyuseles

@indif4ent

@ZyMazza

@Shoalst0ne

@algekalipso

@BrianRoemmele

@alignment_lab

@yoheinakajima

@IntuitMachine

@0x440x46

 

5

 

1

 

15

 

271

 

 

 

 

 

 

 

 

 

 

 

 

Michael P. Frank is joining a startup!

 

@MikePFrank

·

1h

 

I mentioned

@MikeMongo

already in one sub-thread. Imo

@cajundiscordian

should also be included if you didn't already.

[Alex K Chen]

Quote

 

Just finished watching this series. It might be the best long-form portrayal of the singularity I've seen and the plot is amazing. Some takeaways for our world: -PauseAI will be biggest activist movement in history. -Militaries will get involved and have no qualms about breaking the rules to get what they want. -Cyber attack/defense will be extremely important. Both are talent bottlenecked and will be ripped open by AI SWE's -It's really really hard to get a genie back in the bottle, especially if it can copy its code. -We will need really good monitoring systems to detect bad AGI's out in the world. This currently doesn't exist. -Every world power will be running AGI projects in parallel. AGI's working for dueling countries will go to war with each other. -Everything will be increasingly top secret. You won't be able to know what's going on by checking twitter bcuz only 50 ppl in the world know and they're in an air-gapped facility in Norway. Such facilities will probably, actually come to exist (if they don't already) -Speed is enough for superintelligence. If you can run a human-level AI at 1000 tok/s, it can do three years worth of work in a day. -Memory enables a lot. The difference between an AGI that has a continuous life vs one that wakes up anew each day is massive -Multiple intelligent species will co-exist on earth. The deep-seated urge to hate the other species/race will be a major obstacle. -Bioweapons are a form of weapon that only affect biological species and not digital ones. -“the only one that can be trusted with so much power is myself” will be the motto of every ambitious technologist -Ultimately, the key decisions will come down to a very small number of individuals -There will be fights, even (especially) between people who are noble and principled and who want what’s best for the world -Utilitarian calculations will be made in real-time. Real people will have to make choices that sound like speculative trolley problems. -Datacenters will probably, unironically be bombed at some point. They will become the modern-day pyramids. -In a world of infinite life, the most dangerous person is he who is willing to die for a cause -"Where's Ilya" was predicted by the show -Embodiment defines human life. The experience of being physical versus digital will be different in profound ways. -If all goes well, then one day you will be a 100,000 year old superintelligent blob hosted on a Dyson Sphere in Alpha Centauri. This will be the same you as the person reading this right now -The choices you make today will ring out into eternity

 

 

[Dean Pomerleau]

Sounds interesting Alex. What series is it referring to? 

[Gordo]

Hate to feed Alex's paranoia but...

 

p.s. I'm also interested in the unnamed series mentioned above, but the Alex 🤖 seems to have difficulty interacting with others 😭