The situation remains normalizing, all facshed up. At least some of the folks working at Palantir are starting to understand what their work is doing to their world. According to reports in Wired Magazine and 404Media, employees of the meta-surveillance giant are not told how their products are extensively used by ICE to target and round-up real people. Listeners will be relieved to know that the CEOs of Apple, Anthropic, and OpenAI have all condemned ICE violence while couching their critiques with sycophantic praise for Trump. Meanwhile, TikTok users have reported they are unable to write the word "Epstein" or express criticism of US Leader Donald Trump. TikTok was recently forced to sell controlling interest of US operations to a group of investors loyal to Trump. TikTok introduced new terms of service allowing it to be even more brazen with personal privacy, ostensibly on behalf of Trump, than the former owner, ByteDance, was on behalf of the People's Republic of China. Same form-fields, bigger pile. It's hard to take the tech part seriously except it makes up the world we work in and it can be deadly serious, depending on who uses it for work.

There is no sugar coating how brutal the last week has been. ICE agents murdered yet another regular person in the American federal government's relentless and brutal crackdown on Minneapolis Minnesota. This week's extra-judicial killing was broadcast live to the Internet by dozens of bystanders with mobile phones as was the previous one, just a week before. The mental and emotional burden of accepting this outrageous behaviour is the new normal for the American federal government makes it hard to prioritized our work and coverage of tech. It's especially difficult because so many of the companies our work relies on are lending or selling their tools to the now outwardly authoritarian federal government. This week we note how Canadian social media tool Hootsuite is providing social media services to the DHS and the outrage . We also talk about the Electronic Frontier Foundation's efforts to push digital security training to help people protect themselves in the digital space.

Amidst the chaos, Google is facing another anti-trust suit, this one filed by Californian consumers citing Google's dominance of the search space. Google is, meanwhile, appealing another anti-trust ruling that forced it to share search data with competitors suggesting the courts neglected to consider several of the issues properly. Every time Google goes to court, secrets spill out in court filings. This time some of those secrets address how Google looks at spam. Google worries that if spammers learn how Google deals with them, result quality will degrade. The documents go on to mention how user-side data is used to build and train GLUE statistical models and RankEmbed models. The UK and France are considering banning young people from using social media due to the harms social media can do to teens. Australia banned users under age 16, forcing the removal of 5million accounts (from a population of 28million). OpenAI is introducing test ads to the lower tiers of ChatGPT in the coming weeks. The Financial Times reports they expect to make somewhere in the "low billions" in ad revenues in 2026. Google is not expected to introduce ads to the Gemini environment in 2026. These ads are going to be impression based rather than pay per click. Organic search is up and, (surprise), clicks are down only slightly at a decrease of 2.5% year over year according to a large scale study using Similarweb data. All this and a lot more in a densely packed episode as our New Roman Times start to change forever.

Elon Musk's AI-Grok continues to churn out non-consensual images of people including sexualized images of children, an issue that continues into this week. Google is integrating core search signals into its AI Experience in AI Mode and AI Overviews. This, along with Personal Intelligence, the ring that binds all your Google apps that's coming to Gemini and AI Mode soon, will work to "significantly mitigate" known limitations of LLMs. Google is also introducing Universal Commerce Protocol, the new open standard for agentic commerce connecting stores and products to AI results sets. Far from hurting SEO, UCP is likely to introduce new venues for SEO services. Meanwhile, Google's Danny Sullivan took aim at the AIO belief you should create content in easily digestible chunks to better feed LLMs. The theory suggests that smaller sentences and broken down concepts will be easier for LLMs to contextualize and regurgitate. In reality, Google would rather content be created to assist, help, or inform page visitors. They especially don't want people to create two versions of content, one for LLMs and one for the web. Apple and Google make it official, Gemini is the white-labeling itself as the thinking bit of Siri and Apple Intelligence. Google has removed some AI health summaries after an investigation by the Guardian newspaper that found false and misleading information that could put people at risk. In other news, news publishers expect to see traffic drop 43% by 2029 according to a Reuters Institute Report. Stick around to see if you can actually find the final story.