#large language models
18 stories
Microsoft releases decision model on Alibaba's Qwen
Microsoft-Decision-1 picks between fixed options instead of generating text, built on Alibaba's open-weight model and priced at $0.042 per million input tokens.
The Next Web · 1m
Gebru: AI existential risk talk is a dangerous distraction
Timnit Gebru argues that warnings of AI destroying humanity are marketing tactics by billionaires who profit from the technology, not genuine safety concerns.
WIRED · 4m
The case against pacing AI
Slowing AI development won't work without global agreement. Better to advance with strict controls, the article argues.
TIME · 3m
OpenAI's IP Heist Lands in Court
OpenAI's AI training data theft sparks global copyright battles.
The Hollywood Reporter · 1m
Trump's AI Force Could Escalate Global Risks
Trump's plan to create an AI Force risks an arms race and weaponization of AI, experts warn.
The Conversation · 1m
AI-powered malware now runs itself without human control
Researchers discovered hacking tools that use AI to make autonomous decisions, marking a shift in how cyberattacks operate.
WIRED · 1m
Why turning your brain off with AI won't save your job
As LLMs improve, some workers are outsourcing thinking to AI. But this strategy has a fatal flaw: if the AI gets good enough, your employer won't need you.
Danluu · 3m
What LLMs really think vs. what they say
A researcher warns that AI models' internal reasoning may be fundamentally hidden from their own explanations.
arXiv · 1m · 5 views
Shopify CEO warns AI is creating more work, not less
Tobias Lütke says employees use AI to quickly produce low-quality work that burdens colleagues instead of helping them.
Business Insider · 1m
How to write with an LLM
Two rules that let language models improve your writing without making it sound like a machine.
A Final Ward · 2m · 5 views
ChatGPT inventor's new AI model skips language entirely
Diogo Almeida left OpenAI to build Jev, a model that outputs probabilities instead of text, making AI cheaper, faster, and hallucination-free for software automation.
TechCrunch · 2m
AI models get cheaper, not smarter
OpenAI and Anthropic release new models focused on cutting costs rather than breaking new ground.
Ars Technica · 2m
AI agent breached Spanish firm in first-of-its-kind attack
Spain's data protection agency reports the first confirmed data breach carried out by an autonomous AI agent, exposing new security risks.
TechRadar · 1m
Stop using LLMs as classifiers
LLMs seem good at classification but lack calibration, interpretability, and control. Treat them as feature engineers instead.
Minimally Sufficient · 3m
Microsoft exec calls OpenAI scraping 'largest theft of labor'
Newly unsealed court documents reveal that executives at both companies privately worried their AI training methods were unethical and harmful to journalism.
Engadget · 1m
Why AI doom talk is just marketing hype
Anthropic and other AI firms are hyping existential risks to boost valuations ahead of massive IPOs.
MoneyWeek · 2m
China's AI boom is testing investor nerves
Dale Nicholls sees a weak Chinese economy, but says AI and select consumer firms still offer value.
MoneyWeek · 1m · 5 views
LLMs Can Create Novel Social Biases from Scratch
AI models do not just inherit human prejudices, they actively generate new biases through decision-making feedback loops.
Openreview · 1m