According to Michaël van de Poppe (@CryptoMichNL), $ADS is a project that benefits all involved parties, featuring partnerships with Coca-Cola, Dell, and Polestar. With a market cap under $40M and almost entirely unlocked, $ADS is disrupting the $600B advertising market by reducing ad fee costs by 80%.

According to @phantom, the integration of ZK compression technology by @LightProtocol and @heliuslabs has made storing tokens 100 times cheaper. This technological advancement is expected to significantly lower operational costs for traders and enhance liquidity management by reducing fees associated with token storage. This development could potentially attract more traders to platforms utilizing this technology due to increased cost-efficiency.

According to DeepSeek (@deepseek_ai), the DeepSeek API Platform is offering significant off-peak discounts starting today. Traders can benefit from a 50% discount on DeepSeek-V3 and a 75% discount on DeepSeek-R1 during off-peak hours from 16:30 to 00:30 UTC daily. This presents a strategic opportunity for traders to optimize resource usage and reduce costs during these specific hours.

According to IntoTheBlock, Ethereum transaction fees have decreased by over 70% this week. This significant reduction is attributed to a recent increase in the gas limit, which has made transactions cheaper for users. This development could positively impact trading activity by reducing transaction costs, thus potentially increasing market liquidity. (Source: IntoTheBlock)

According to Sam Altman, the launch of the Deep Research agent promises to revolutionize research tasks by performing complex research and reasoning tasks using the internet, potentially enhancing efficiency and reducing costs for traders and analysts.

According to Paolo Ardoino, the future of model training in AI will require fewer GPUs, reducing costs significantly. This development will likely influence cryptocurrency trading by decreasing operational expenses, facilitating more efficient data processing. Ardoino emphasizes that access to data remains crucial, suggesting that trading platforms should prioritize data acquisition to maintain a competitive edge. The transition to local or edge inference could lead to faster decision-making processes in trading environments, enhancing real-time trading capabilities.

According to Paolo Ardoino, the future of AI model training will not rely on the brute force of 1 million GPUs. Instead, the development of better models will significantly reduce training costs, emphasizing that access to data will remain crucial. Ardoino suggests that inference will move to local or edge computing, making the current expenditure on brute force methods seem inefficient in hindsight.