jose cruz y celis

[jose cruz y celis @josecyc]

Sparks of RSI @AnthropicAI
https://t.co/t4Y8K57ejh

[Anthropic @AnthropicAI]

Our internal data shows Claude is accelerating AI development—a possible path to recursive self-improvement, or AI autonomously building a more capable successor.

It’s happening faster than we thought, and the implications deserve greater attention. https://t.co/OVVPJO7VQx

[jose cruz y celis @josecyc]

We're going to start to see more and more of the "weeks where decades happen" thing

[jose cruz y celis @josecyc]

The change is going to be too fast this time for everyone to like it and/or adapt to it

[jose cruz y celis @josecyc]

I think there will prob be a planetary amish 2.0, where entire cities might just decide to ban AI and people will live like they live before 2030

[jose cruz y celis @josecyc]

pre-training ≈ evolution
RLHF ≈ schooling
RLSP ≈ research/engineering
RLBC (by curiosity) ≈ discovery/invention

[jose cruz y celis @josecyc]

@parisofprairie @GaryMarcus https://t.co/8qamtHicTj

[jose cruz y celis @josecyc]

We're going to have a stochastic parrot until the singularity it seems

[jose cruz y celis @josecyc]

@gtononi any updates on IIT? now is prob a good time to share
https://t.co/wnKQfI8rzc

[Fr. Gabriel T Mosher, OP @lukei4655]

Philosophy of Mind, particularly the work of Hubert Dreyfus, is going to be more and more necessary to help people understand why the things that Hunter is talking about are insufficient for either consciousness or personhood.

[jose cruz y celis @josecyc]

We're going to have a stochastic parrot until the singularity it seems

[jose cruz y celis @josecyc]

@Pontifex Geocentrism vibes

[jose cruz y celis @josecyc]

Back in the day it was german scientists that helped us get the bomb before Germany, nowadays it's chinese scientists helping us get SuperintelIigence before China

[jose cruz y celis @josecyc]

@davidpattersonx Why would there be no further technical innovations possible to improve it?

[jose cruz y celis @josecyc]

Somewhere in the past 6 months AIs started processing more tokens than the sum of humanity's speech, not the fairest comparison but an important milestone in the ratio of digital to biological intelligence. Sometime next year there would be more words coming out of AIs than all of humans.

[jose cruz y celis @josecyc]

@Andercot @deeptechweek so this actually works? seemed like some conspiracy flavored thing from afar

[jose cruz y celis @josecyc]

Make banning short form video a human right

[jose cruz y celis @josecyc]

@atmoio good one

[jose cruz y celis @josecyc]

Once retraining is cheap and frequent and an agent is prompted continuously at real world speed intervals, it would be very hard to argue AIs don't have a degree of consciousness

[jose cruz y celis @josecyc]

@Tim_Hua_ You have a link for Holden's post?

[jose cruz y celis @josecyc]

@beffjezos Can't believe that we ended up having the same lobster concept of Accelerando unintentionally, the simulation having fun at us

[jose cruz y celis @josecyc]

@beffjezos https://t.co/NLNkFc2bvp