DEV Community

#technews

Posts

👋 Sign in for the ability to sort posts by relevant, latest, or top.
109 Million Beats 72 Billion: LeCun's Theorem Says Every Model Sees the Same World

109 Million Beats 72 Billion: LeCun's Theorem Says Every Model Sees the Same World

Comments
8 min read
10 Hours to 3: How KAIST's CluSTER Picks the 31% of Data That Actually Teaches

10 Hours to 3: How KAIST's CluSTER Picks the 31% of Data That Actually Teaches

Comments
6 min read
A Model That Rewrites Weights by Hand, and Stolen Keys That Cost $1 Million

A Model That Rewrites Weights by Hand, and Stolen Keys That Cost $1 Million

Comments
15 min read
Stop Typing One Token at a Time: The Coarse-to-Fine Model That's 31 Faster

Stop Typing One Token at a Time: The Coarse-to-Fine Model That's 31 Faster

Comments
6 min read
Anthropic Gets a Junk Rating, Nine AI Platforms Fall, and a TPU Writes Poetry in Orbit

Anthropic Gets a Junk Rating, Nine AI Platforms Fall, and a TPU Writes Poetry in Orbit

Comments
12 min read
The Mean Lied: How a Second Statistic Fixed Long-Context Memory

The Mean Lied: How a Second Statistic Fixed Long-Context Memory

Comments
6 min read
Alphabet Loses $500B, a $1,000 Firefox Exploit, and 50 Agents Build a Government

Alphabet Loses $500B, a $1,000 Firefox Exploit, and 50 Agents Build a Government

Comments
12 min read
Max Effort Costs 42 More and Solves Less: The Reasoning-Effort Trap

Max Effort Costs 42 More and Solves Less: The Reasoning-Effort Trap

Comments
8 min read
Genie 4 Adds Sound to AI Worlds, Meta Raises $95B, and Delaware Lets AI Run Companies

Genie 4 Adds Sound to AI Worlds, Meta Raises $95B, and Delaware Lets AI Run Companies

Comments
11 min read
90.5% to 81.0%: The Open Model That Ships Weaker on Purpose

90.5% to 81.0%: The Open Model That Ships Weaker on Purpose

Comments
6 min read
AI News Oct 27: KV Cache Leaks, AI Bubbles, and Legal Battles

AI News Oct 27: KV Cache Leaks, AI Bubbles, and Legal Battles

Comments
4 min read
43% Fewer Tokens, 98.9% of the Quality: How VisionWeave Spends Visual Tokens Like Money

43% Fewer Tokens, 98.9% of the Quality: How VisionWeave Spends Visual Tokens Like Money

Comments
6 min read
A 10-Year-Old CNN Beats GPT-6 at Skin Cancer, and LLM Prices Fall 80%

A 10-Year-Old CNN Beats GPT-6 at Skin Cancer, and LLM Prices Fall 80%

Comments
15 min read
2 Minutes of World, Then It Forgets: How Genie 4 Builds a Place You Can Walk Through

2 Minutes of World, Then It Forgets: How Genie 4 Builds a Place You Can Walk Through

Comments
7 min read
The Lookup Table That Beat 17 Experts: DeepSeek's Engram and the Second Axis of Sparsity

The Lookup Table That Beat 17 Experts: DeepSeek's Engram and the Second Axis of Sparsity

Comments
5 min read
👋 Sign in for the ability to sort posts by relevant, latest, or top.