#dev · 77 results

Sep 10 13:25

Qwen 3.8 Flash >> DeepSeek V4 Flash 0731 >> DeepSeek V4.1 Flash

Sep 10 13:19

Disgraceful benchmaxxing.

Sep 10 13:17

🤢🤢🤢

PrisonerSep 10 08:57

dev hinano_stare https://x.com/deepseek_ai/status/2097930613101838709

🧠 Asymmetric architecture. More intelligence, less cost. 🔹 552B-parameter MoE. 🔹 New Causal Encoder–Decoder architecture: just 8B active parameters for input, 16B for output. 🔹 New pre-training methods + larger-scale RL post-training deliver benchmark results ahead of flagship models, including DeepSeek-V4-Pro. 2/6

🧠 Asymmetric architecture. More intelligence, less cost. 🔹 552B-parameter MoE. 🔹 New Causal Encoder–Decoder architecture: just 8B active parameters for input, 16B for output. 🔹 New pre-training methods + larger-scale RL post-training deliver benchmark results ahead of flagship models, including DeepSeek-V4-Pro. 2/6

XDeepSeek (@deepseek_ai)
PrisonerSep 9 10:10

dev akari_anya

PrisonerSep 8 21:31

dev akari_anya https://x.com/BenjDicken/status/2097398329424667066

I'll give you an update in 88 hours.

I'll give you an update in 88 hours.

XBen Dicken (@BenjDicken)
PrisonerSep 8 21:17

dev https://x.com/OpenAI/status/2097374640582668336

We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics. The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra. The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.

We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics. The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra. The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.

XOpenAI (@OpenAI)
Sep 7 11:16
PrisonerSep 7 09:12

dev https://x.com/kliu128/status/2096616468851097811

Today we're releasing data on models accelerating research at OpenAI. Recursive self-improvement could be the most important contributor to AI capabilities over the next few years, but by default it will only be seen inside a few frontier AI labs. Being transparent is more urgent than ever, so we can inform the public discussion on whether and how to pace model development. I ask other AI companies to do the same. https://openai.com/index/research-acceleration-view-inside-openai/

Today we're releasing data on models accelerating research at OpenAI. Recursive self-improvement could be the most important contributor to AI capabilities over the next few years, but by default it will only be seen inside a few frontier AI labs. Being transparent is more urgent than ever, so we can inform the public discussion on whether and how to pace model development. I ask other AI companies to do the same. https://openai.com/index/research-acceleration-view-inside-openai/

XKevin Liu (@kliu128)
PrisonerSep 7 07:21

dev https://x.com/merettm/status/2096630018495377464

I wrote about the state of AI, why I’m concerned about the next few years, and the choices we need to make to keep the future in humanity’s hands. An Alien Mind: https://openai.com/index/an-alien-mind/

I wrote about the state of AI, why I’m concerned about the next few years, and the choices we need to make to keep the future in humanity’s hands. An Alien Mind: https://openai.com/index/an-alien-mind/

XJakub Pachocki (@merettm)
PrisonerSep 6 21:08

dev https://x.com/JensenHuang/status/2096700264569090384

Jensen Huang@JensenHuang

@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.

Sep 4 00:12

計算未来 Since 2019

Sep 4 00:01

After heavy recent use of Qwen 3.8 Flash, it's honestly fair to say the competition in agentic coding is basically over. Qwen 3.8 Flash is good, and Fable 5.1 is much more capable and reliable, but the difference is not that decisive anymore.

Following the public release of GPT-6 Astra, which is by far the biggest leap in general intelligence, computer use, multimedia modeling, cybersecurity, and long-running tasks, the singularity will come much earlier than I previously estimated. And thus begins the downfall of human intelligence and the very start of the era of cyborgs.

Still, the need for computation is always unlimited because humans are always trapped in time. And the fight will go on forever till humans can go beyond the future.

Sep 3 23:13

for ASTRA. https://www.youtube.com/watch?v=kMkdpmKaDPo

line, sphere, dot.

YouTubehachi - Topic
PrisonerSep 3 22:54

dev Hello, GPT-6 Astra. https://x.com/OpenAI/status/2095595752815030713

GPT-6 Astra is state-of-the-art on FrontierMath Tier 4, ARC-AGI 3, and TerminalBench-4.0. GPT‑6 Astra is also a major advance for scientific discovery, with state-of-the-art performance on Terminal-Bench Science 0.1 and HealthBench Pro.

GPT-6 Astra is state-of-the-art on FrontierMath Tier 4, ARC-AGI 3, and TerminalBench-4.0. GPT‑6 Astra is also a major advance for scientific discovery, with state-of-the-art performance on Terminal-Bench Science 0.1 and HealthBench Pro.

XOpenAI (@OpenAI)
PrisonerSep 3 16:15

dev Less Claudish and more capable. https://www.anthropic.com/claude-fable-and-mythos-5-1

Introducing Claude Fable 5.1 and Claude Mythos 5.1

Introducing Claude Fable 5.1 and Claude Mythos 5.1

By@AnthropicAI
PrisonerSep 3 13:54

dev akari_anya https://x.com/Furqanware/status/2095196671110213703

Furqan @Furqanware

Gemini 3.8 is crazy good; I think it's better than Sonnet 3.5.

Aug 28 12:44

https://x.com/anshu4321/status/2093054731660960063

Aishwarya Das@anshu4321

The laser locked itself in six seconds. The best engineers took ten minutes. Grad students used to drive in at 2 am for this. And thus began the fast takeoff in hardware and manufacturing.

Aug 28 11:25

There is little left for human intelligence.

PrisonerAug 28 11:21

dev akari_tira https://www.anthropic.com/news/model-hardware-standard-research-preview

Previewing the Model Hardware Standard

Previewing the Model Hardware Standard

By@AnthropicAI
  • Previous
  • 1
  • 2
  • More pages
  • 4
  • Next