In this episode of the MadTech Podcast, ExchangeWire head of marketing Grainne Reid is joined by CEO Rachel Smith and Ci En Lee, media strategist at Publicis Singapore to discuss the latest in the ad tech and marketing landscapes. 

They examine the latest with the TikTok US ban amidst Trump’s tariffs, the impact ‘woke’ marketing, and DeepSeek’s current position in the global AI race. 

Trump to extend TikTok ban enforcement deadline after China tariffs derail deal (CNN) 

President Donald Trump announced that he is again postponing enforcement of the TikTok sale-or-ban law for 75 days. The delay comes after Trump’s tariff announcement derailed a deal that had been set to transfer control of the app’s US operations to American ownership. 

Marketing’s ‘woke’ rebrand has ultimately helped the far right (The Guardian) 

In recent years, brands started marketing their values of diversity and inclusivity, along with their services/products. Now as we’re met with brands adjusting to new policies and stripping away these values. Eugene Healey argues that “Social progress once came hand-in-hand with economic progress. Now, instead, social progress has been offered as a substitute for economic progress.” He notes that multinational brands have been promoting social transformation, while simultaneously being one of the largest contributors to declining living standards. He says that marketers must reckon with how they’ve trivialised activism by turning it into comms strategy, and co-opted movements only to abandon them. 

DeepSeek and Tsinghua Developing Self-Improving AI Models (Bloomberg) 

DeepSeek is working with Tsinghua University on reducing the training its AI models need in an effort to lower operational costs. The company collaborated with researchers from the Beijing institution on a paper detailing an approach to reinforcement learning to make models more efficient. This new method aims to help AI models better adhere to human preferences by offering rewards for more accurate and understandable responses. It has been proven effective, but expanding it to more general applications has proven challenging. DeepSeek’s team is currently working on solving this with something it calls self-principled critique tuning. DeepSeek is calling these new models DeepSeek-GRM - short for “generalist reward modeling”.