g
@fuglypirate
Followers
6
Following
70
Statuses
271
I have accidentally trained my mind to expect the exponential. I'm annoyed if I go a week without seeing an exciting technological revelation.
0
0
0
I guess they need to make o3 available in Canvas. 😅
Today, we released our report on competitive programming with large reasoning models. We show how our LLMs have evolved from amateur competitive programmers to competing with the world's best:
0
0
0
After trying to use OpenAI Canvas to make games, I'm not worried about AGI anymore. 😅 Lots of room for improvement.
10
0
0
our heuristics will fail us knowing that is half the battle
@Cluffalo If we haven't already, we will reach a generation where the majority of personal beliefs about the world are informed by AI-generated narratives.
0
0
1
My 2 AI goalposts: Inexpensive autonomous vehicles that can travel anywhere an average human can. AI that can solve Riemann hypothesis-level question or makes a similarly impactful mathematics/physics discovery. After that, it goes beyond my horizon.
0
0
1
RT @lmthang: Excited to share details of AlphaGeometry2 (AG2), part of the system that achieved silver-medal standard at IMO 2024 last July…
0
185
0
RT @RaphaelDabadie: “I cannot overstate how much progress we are gonna make in the next two years. We know how to improve those models SO S…
0
18
0
additional context 😅
@kimmonismus I would bet less than 3 years from now for all of them. I meant "far from done" in terms of model capability rather than time.
0
0
0
RT @polynoamial: There's a lot of talk of LLMs "saturating all the evals" but there's plenty of evals people could make where LLMs would do…
0
107
0
@_akhaliq If we keep doing these bouncing-ball tests, everyone is just going to over-index on this trope.
0
0
0
RT @levie: AI is going to cause us to move to higher levels of abstraction of how we work. Each level of abstraction provides more leverage…
0
95
0
RT @DrJimFan: We RL'ed humanoid robots to Cristiano Ronaldo, LeBron James, and Kobe Byrant! These are neural nets running on real hardware…
0
468
0
RT @tsarnick: Google DeepMind CEO Demis Hassabis says AGI that is robust across all cognitive tasks and can invent its own hypotheses and c…
0
149
0
RT @kimmonismus: OmniHuman-1 can generate realistic human videos at any aspect ratio and body proportion using just a single image and au…
0
118
0
RT @tszzl: you write one blogpost, you’re a public intellectual. you write 20 tweets of equal length, you’re an addict
0
97
0