AI Safety Debate Criticized for Concentrating Power, Not Preventing Harm

AI safety artificial intelligence power concentration

L'essentiel

Andy Konwinski argues that the AI safety conversation is being used to consolidate power, citing Anthropic's reversed policy of degrading competitor AI training.
He advocates for a research commons with open access to frontier-scale compute, comparing centralized AI control to historical power grabs.

Résumé généré par IA

Taille de police

Andy Konwinski argues that the AI safety conversation is being used to consolidate power, citing Anthropic's reversed policy of degrading competitor AI training. He advocates for a research commons with open access to frontier-scale compute, comparing centralized AI control to historical power grabs.

Read the full article on Decrypt

Sujets liés

AI safety

Sujetsartificial intelligence power concentration open frontier research commons Anthropic Perplexity AI

This article was originally published by Decrypt.

Articles liés

Plus sur ce sujetAI safety

En développement·2 sa önce

China Cracks Down on AI Personalities, Disabling Custom Agents

China is set to ban AI services that simulate human personality for sustained emotional interaction, leading ByteDance and Alibaba to disable custom agent features in their AI products ahead of new regulations taking effect July 15.

Decrypt

Urgent·2 sa önce

UK Foreign Secretary Warns of AI Risks, Compares to Nuclear Age

UK Foreign Secretary Yvette Cooper warned governments risk repeating nuclear age mistakes by delaying AI regulation. She urged international agreements on frontier technology before a crisis, comparing the AI race to the early nuclear arms race and highlighting the need for proactive safety principles.

Decrypt

En développement·3 sa önce

Coinbase AI News System Sends False World Cup Match Alert

Coinbase's AI-powered news system mistakenly alerted users about a World Cup match outcome before it occurred. The alert falsely reported Norway defeating Brazil 3-2, hours before the actual match, which Norway eventually won 2-1.

Decrypt

En développement·4 sa önce

Joi AI's $2,000 'Masturbation Consultant' Study Goes Viral, Attracting 150,000 Applicants

AI companion company Joi's offer of $2,000 for participants in an AI-guided masturbation study unexpectedly went viral, drawing over 150,000 applicants. The company, which aims to explore AI's role in intimacy and wellness, had to quickly establish a formal selection process for the 10 diverse participants needed.

Decrypt

En développement·9 sa önce

Vitalik Buterin Unveils "Lean Ethereum" Roadmap for Major Protocol Overhaul

Vitalik Buterin has outlined "Lean Ethereum," a three-to-four-year plan to rebuild the blockchain, replacing core components with recursive STARK proofs, quantum-safe cryptography, and enhanced privacy features. The overhaul aims to significantly increase data storage capacity and reduce fees.

Decrypt

En développement·15 sa önce

XRP Ledger Stablecoin Supply Surges to Nearly $890 Million, Driven by Ripple's RLUSD and New Entrant USDV

Stablecoin supply on the XRP Ledger has surged to nearly $890 million, a 20.56% increase in 30 days, outpacing the global market. Ripple's RLUSD dominates this growth, while Valtorum's USDV is emerging as a significant second issuer.

CryptoSlate

Plus sur ce sujetAI safety