Tag
A federal judge blocked New York state's ban on algorithmic rent-pricing software, striking down a law aimed at curbing the use of AI-driven tools that coordinate rental prices among landlords.
Pawel Huryn tested how much API usage is actually bundled in popular AI subscriptions and priced the calls at list API rates, revealing that plans like Claude Max 20x and Meta's Power Usage tier subsidize usage far more than advertised multipliers suggest.
Phil Schmid comments on an unusual AI product usage policy that allows unlimited human usage while restricting agent usage, noting it as an interesting first.
作者认为 OpenAI 削减 200 美元订阅计划只是开始，AI 订阅价格最终都会向 API 价格看齐，并警告用户不要通过订阅间接补贴 AI 公司的数据中心建设成本。
Greg Isenberg shares 12 observations on Google's Gemini 4 Argon announcement, covering quantum software optimization, data center memory savings, C/C++-to-Rust code rewriting, million-token outputs, tiered cyber-safety access, bug-finding capabilities, and new pricing and monitoring implications.
An AI founder (gabriel) shares frustration that enterprise customers can't get model and budget commitments because AI model prices are halving every two weeks, down from every quarter earlier this year.
The article speculates on future AI pricing after discounts disappear, arguing that as AI reaches human-expert quality, the real consideration becomes the value of users' time.
The speed of AI cost reduction has surpassed all major technologies in history, with prices dropping about 47% per quarter since 2023, which is multiple times faster than declines in DNA sequencing, computing, and other sectors.
The article questions the financial sustainability of AI subscription services, exploring how much users should pay to cover the costs of running large language models like ChatGPT or Claude.
The article draws parallels between the historical impact of cheap paper on double-entry bookkeeping and OpenAI's recent 80% price reduction for the GPT Luna model, speculating on future technological innovations.
Seattle City Council passed the Fair Pricing and Transparency Act to ban surveillance pricing in groceries, prohibiting personalized pricing based on consumer data. Supported by Consumer Reports, the bill aims to protect shoppers from unfair data-driven price differences.
This article explains the cost factors for tasks on Opus 5.5, detailing how turns, cache reads, output tokens, and model selection influence token expenses.
Chinese AI models are undercutting US models by up to 90% in price, causing a dramatic shift in market share and echoing historical subsidy strategies in global business.
The article discusses how reductions in AI costs might shift usage from transactional queries to continuous monitoring, with implications for businesses, though challenges in interpreting automated reports remain.
Frontier AI is facing a pricing crisis as token usage surges 25-fold, while mid-tier models now deliver 90% of flagship capability at one-sixth the cost, reshaping the economics of advanced AI.
Meta is incentivizing users to share prompts and outputs for its new AI model Muse Spark by offering up to 95% discounts, aiming to gather training data for future improvements while addressing data privacy and competitive pricing in the AI industry.
A tweet highlights a 50% discount on GLM-5.3-Flash API pricing, making AI intelligence more affordable for two weeks.
Nvidia is raising prices for AI servers by 15% due to increased memory costs, shifting the supply bottleneck from GPUs to memory manufacturers like SK Hynix, Samsung, and Micron.
The article discusses a price war in the AI model market, where aggressive undercutting by providers like Meta and Google is creating unsustainable economics, potentially leading to a market correction as enterprises reassess spending on AI tokens.
OpenAI has lowered the pricing for the GPT-5.6 Sol AI model, and Vercel's AI Gateway is offering an additional 50% discount on these new rates, further reducing costs for input and output tokens.