I expect rapid progress but not towards general superintelligence

DeepSeek 4.1 Flash performs comparably to frontier models at dramatically lower cost, making high-volume AI development affordable for builders.

Ben Thompson argues Apple is diverging from hacker/developer needs in the AI era, favoring integrated consumer experiences over customizability.

MIT Technology Review argues that AI's refusal mechanisms are fundamentally unreliable because harmful capabilities remain intact while safety measures are probabilistic add-ons.

A mathematician critiques OpenAI's preprint claiming to prove the Partition Principle doesn't imply the Axiom of Choice, calling it poorly written with strange structure and questionable references.

Cryptographer Matthew Green warns there's a 15% chance AI advancements could break public-key encryption, requiring advance preparation for cryptographic standards.