Agent

82 briefings

• World Models Go Multiplayer, Real-Time at 24FPS
• Agents Start Improving Themselves, and Reaching for Fewer Tools
• Diffusion Swallows the Decoder Too

View topic →

Multimodal

78 briefings

• Agents Start Improving Themselves, and Reaching for Fewer Tools
• Vision Models Start Redesigning How They Output
• Diffusion Swallows the Decoder Too

View topic →

Evaluation

76 briefings

• World Models Go Multiplayer, Real-Time at 24FPS
• Agents Start Improving Themselves, and Reaching for Fewer Tools
• Vision Models Start Redesigning How They Output

View topic →

Training

72 briefings

• World Models Go Multiplayer, Real-Time at 24FPS
• Agents Start Improving Themselves, and Reaching for Fewer Tools
• Diffusion Swallows the Decoder Too

View topic →

Image Gen

66 briefings

• Agents Start Improving Themselves, and Reaching for Fewer Tools
• Vision Models Start Redesigning How They Output
• Diffusion Swallows the Decoder Too

View topic →

Safety

61 briefings

• Agents Start Improving Themselves, and Reaching for Fewer Tools
• The Rulers We Use to Measure What Models Really Think Are Broken
• Optimizer Choice Stretches Capacity Scaling 2.3x

View topic →

Efficiency

55 briefings

• World Models Go Multiplayer, Real-Time at 24FPS
• Diffusion Swallows the Decoder Too
• The Rulers We Use to Measure What Models Really Think Are Broken

View topic →

Architecture

55 briefings

• Agents Start Improving Themselves, and Reaching for Fewer Tools
• Agent Trajectories Let a 30B Match a 235B
• $15 Per Paper, Healthcare Agents Cap at 28%

View topic →

AI for Science

49 briefings

• Agents Start Improving Themselves, and Reaching for Fewer Tools
• The Rulers We Use to Measure What Models Really Think Are Broken
• Gated DeltaNet-2 Splits the Gate, Maestro Outscores GPT-5

View topic →

Robotics

48 briefings

• Gated DeltaNet-2 Splits the Gate, Maestro Outscores GPT-5
• Dual-Stream MoE Unifies Multimodal, Garment Video 30x Faster
• Readable Rules Don't Belong in LLM Weights

View topic →

Reasoning

46 briefings

• Agents Start Improving Themselves, and Reaching for Fewer Tools
• Gated DeltaNet-2 Splits the Gate, Maestro Outscores GPT-5
• Olympiad Gold Becomes a Two-Step Recipe

View topic →

Retrieval

45 briefings

• Vision Models Start Redesigning How They Output
• The Rulers We Use to Measure What Models Really Think Are Broken
• Agent Trajectories Let a 30B Match a 235B

View topic →

Video Gen

43 briefings

• Diffusion Swallows the Decoder Too
• Agent Trajectories Let a 30B Match a 235B
• Dual-Stream MoE Unifies Multimodal, Garment Video 30x Faster

View topic →

Interpretability

39 briefings

• Agents Start Improving Themselves, and Reaching for Fewer Tools
• The Rulers We Use to Measure What Models Really Think Are Broken
• Optimizer Choice Stretches Capacity Scaling 2.3x

View topic →

Code Intelligence

31 briefings

• Vision Models Start Redesigning How They Output
• $15 Per Paper, Healthcare Agents Cap at 28%
• Soohak Caps Top Models at 30%

View topic →