LWiAI Podcast #243 - GPT 5.5, DeepSeek V4, AI safety sabotage

# Last Week in AI: Major Model Updates and Safety Concerns The AI world saw significant developments with new versions of major AI systems—including GPT 5.5 and DeepSeek V4—that promise improved capabilities for everyday tasks like writing, analysis, and problem-solving. Meanwhile, researchers raised alarms about potential "safety sabotage," highlighting growing concerns that people might deliberately compromise the safeguards built into AI systems. These developments reflect the ongoing tension between rapid AI advancement and the need to ensure these powerful tools remain reliable and trustworthy.