Metastable @metastable_1

Experiencing vertigo while standing on the precipice. metastable1.tumblr.com/forecasting Joined December 2020

Tweets

3K
Followers

128
Following

100
Likes

15K

Siméon @Simeon_Cps

2 days ago

The ask is based tbh. Even if the premise likely never comes true, the symbolic power of such a statement would be massive. This statement is also easy to agree with if one thinks we have double digits percent chance to blow ourselves up with current level of safety…

Michaël (in London) Trazzi @MichaelTrazzi

3 days ago

527 146 1K 1.3M 552

Download Image

4 10 99 10K 6

Download Image

Steven Byrnes @steve47285

3 weeks ago

There’s a funny thing where economics education paradoxically makes people DUMBER at thinking about future AI. Econ textbooks teach concepts & frames that are great for most things, but counterproductive for thinking about AGI. Here are 4 examples. Longpost: THE FIRST PIECE of…

20 37 276 47K 172

Metastable @metastable_1

2 weeks ago

Unlike many people, Gary Marcus is willing to bet money on his beliefs. Very respectable.

Gary Marcus @GaryMarcus

3 weeks ago

Unlike many people, Gary Marcus is willing to bet money on his beliefs. Very respectable.

6 2 107 8K 6

Download Image

0 0 1 26 0

Stefan Schubert @StefanFSchubert

2 weeks ago

That a university lecturer would say this is appalling. While it's unusually direct, it's not uncommon. It's not just science-denialism - it's a direct attack on people for making accurate scientific claims (that IQ is in some sense real). It's underrated how bad that is.

Dennis M. Hogan @dennismhogan

2 weeks ago

400 2K 51K 1.6M 744

24 39 694 53K 42

Metastable @metastable_1

2 weeks ago

"Καλῶς ἔχει", said little Johnny.

0 0 0 31 0

Download Image

Metastable @metastable_1

2 weeks ago

The old self is dying, and the new self struggles to be born: now is the time of monsters.

0 0 0 27 0

Metastable @metastable_1

2 weeks ago

October 2022. Flan-PaLM 540B, instruction fine-tune of PaLM. "challenging open-ended questions":

0 0 0 30 0

Download Image

davidad 🎇 @davidad

a year ago

Here’s GPT-4 performance on PIAAC literacy in 2023. Something very interesting here is that GPT-4 underperforms Level 4 adults on *Level 3 questions*, despite outperforming Level 4 adults on Level 4 questions, and matching or outperforming Level 2-3 adults on all questions.