DeepSeek built a Claude Code rival, then quadrupled its prices
THE SO WHAT
DeepSeek’s 4x price hike on V4-Pro and V4-Flash in days shows how volatile AI economics are when demand spikes and capabilities harden. If your unit economics assume today’s token prices, stress-test them against 2–4x scenarios and consider hedging with open models or on-prem where feasible.
READ THE SOURCE
MORE FROM THE WIRE
Reed Hastings says AI may push CEOs to cut their workforce in half — or expand their reach
When a scaled operator talks about AI enabling either 50% headcount cuts or massive reach expansion, the subtext is leverage, not inevitability. Boards should be asking CEOs to quantify where AI augments revenue per employee versus where it just masks underinvestment in process and product.
OpenAI shake-up continues with second major departure of the week
Two senior revenue leaders exiting as OpenAI heads toward an IPO means its go-to-market motion is being rebuilt in-flight. If you're a customer or partner, expect pricing, packaging, and account coverage to move around over the next 6–12 months—lock in terms where you can.
Applied AITwitch will train Amazon’s AI on your streams unless you find the setting
Default-on data harvesting for AI training is the new dark pattern—buried opt-outs shift value from creators to platforms by design. If your business depends on UGC or prosumer content, expect rising pressure for transparent data terms and consider making “AI training controls” a competitive feature, not a compliance afterthought.
Applied AIOpenAI previews Ultrafast, an API tier powered by Cerebras that runs GPT-5.6 Sol up to 14× faster and generates up to 750 output tokens per second
A 14× speed bump and ~750 tokens/sec on GPT-5.6 Sol via Cerebras-backed Ultrafast turns latency from UX tax into a design variable. If you're building agents or real-time copilots, revisit which workflows you kept human-in-the-loop purely because models were too slow.