singularity or bust @singularitywen

Joined March 2026

Tweets

173
Followers

18
Following

753
Likes

219

singularity or bust @singularitywen

8 hours ago

@arampell Very apt and thought provoking .. llms went from basic word generator to logic (coding) to reasoning to maths , hopping it builds on the remaining sciences as well

0 0 0 14 0

View Details

singularity or bust @singularitywen

9 hours ago

@ainativefirm @nicbstme Copilot wrote this so it should win prize for best fiction.

0 0 0 42 0

View Details

singularity or bust @singularitywen

13 hours ago

@EliCartan @SebastienBubeck If all a human did was ask Model to believe in itself and directed model to in the right direction (as a attending professor would do) absolutely no human deserves the credit , they can be in citation.

0 0 0 595 0

View Details

singularity or bust @singularitywen

13 hours ago

@teortaxesTex Fml , other millennium problems are in play as well , full scale strategy to tank Anthropic IPO

0 0 3 1K 1

View Details

singularity or bust @singularitywen

13 hours ago

@kimmonismus So this newer model or successor to Bel , possibly 10 T model

0 0 0 219 0

View Details

singularity or bust @singularitywen

13 hours ago

@docmilanfar Navier would we have imagined one problem Stoking so much competition among labs. Euler just oblivious to the happening.

0 0 1 360 0

View Details

singularity or bust @singularitywen

14 hours ago

@instrumenthull I hope the friction reduces and this flow of ideas is infinite

0 0 0 83 0

View Details

singularity or bust @singularitywen

14 hours ago

@cdngdev Wow … curious how many tokens did it use?

0 0 0 214 0

View Details

singularity or bust @singularitywen

14 hours ago

@pigeon__s It fails “make me a sandwich” benchmark , can’t be AGI till it does that.

0 0 2 906 0

View Details

singularity or bust @singularitywen

17 hours ago

@mark_k Interested

0 0 0 13 0

View Details

singularity or bust @singularitywen

2 days ago

@ben_golub Actual name was Evalerdos , but it got shortened to make it easy for others to call.

0 0 1 325 0

View Details

singularity or bust @singularitywen

2 days ago

@mark_k Would love to join .. thank you

0 0 0 28 0

View Details

singularity or bust @singularitywen

2 days ago

@nicbstme It is less than 200 MW compute that was used for ASTRA , many multiple of that coming online in 2027.

0 0 1 699 0

View Details

singularity or bust @singularitywen

2 days ago

Let us celebrate this Labor Day as the first Labor Independence Day—the day AGI begins freeing humanity from the necessity of paid labor: a future where machines do the work, people reclaim their time, and freedom means creating, learning, loving, and living.

Jensen Huang @JensenHuang

2 days ago

@ChaseLochmiller @OpenAI GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years. AGI has arrived. Congratulations @OpenAI team. 400K GPUs coming online next.

2K 5K 42K 11.8M 7K

0 0 0 47 0

View Details

singularity or bust @singularitywen

3 days ago

@oneill_c I think it meant for orchestrated Agents to drive end to end research with validation as opposed to using Ai to generate code or ideas. But in true sense coding agents have only been useful starting this year.

0 0 2 2K 0

View Details

singularity or bust @singularitywen

3 days ago

@scaling01 I perceive Frontier labs have already reached a very high degree of RSI internally , they are signaling late 2027 or early 2028 to avoid getting shutdown & negative public perception and use next year to get decisive advantage over others including OS AI.

1 0 2 125 0

View Details

singularity or bust @singularitywen

3 days ago

@Mr_Salio No .Elon himself admitted this is not model to beat Fable and therefore Astra. It is the next or most likely the one that follows (7 T or 10 T)

0 0 0 397 0

View Details

singularity or bust @singularitywen

3 days ago

@AlexKontorovich Better model means cheaper inference , mathematician themselves will be able to do it rather than labs. Problem maybe funding and grants may dry up where maths will not yield commercial benefit as their is no mystery left in unknowns.

0 0 4 400 0

View Details

singularity or bust @singularitywen

3 days ago

@p_ferragu @wholemars Plus non-RE components, pretty big deal , will it show in quarterly results.. no , but does it set Tesla to leap ahead of competitors , F yeah.

0 0 0 9 0

View Details

singularity or bust @singularitywen

4 days ago

Next Goal should be to distill domain specific small AI models that can be run on home computers that solve 80 % of use cases , this small models then can offload to larger models if they cannot solve it.