📫 Subscribe | 🗣️ Hire us to speak | 🎓 Online Courses | 📅 Events

Late last week, Anthropic announced that it will be embedding imperceptible text watermarks and file metadata into the outputs of their Claude models worldwide to meet transparency requirements imposed by the EU AI Act. The technology triggers detection signatures even during minor tasks like proofreading, light copyediting, basic formatting, or translating human-written drafts. However, Anthropic notes that the watermarks lose reliability if the generated text is heavily rewritten, mixed with other copy, or kept very short, as AI developers work toward a Dec. 2 deadline to bring legacy models into compliance.

This week we learned that Microsoft is unifying its consumer and work Copilot applications into a single streamlined interface ahead of a broader “super app” rollout planned for later this quarter. The consolidated platform brings Microsoft 365 productivity tools like Word, Excel, and Outlook directly to consumer users, while retiring standalone features such as group chats and Deep Research. Though both account types will now share the same updated app interface, Microsoft ensures that work and personal data boundaries, tenant controls, and compliance protections remain completely separate.

Last week, Suno launched Studio 2.0 for its Premier subscribers, introducing MIDI support, advanced stem separation, high-quality audio exports, and a new AI chat bar capable of designing custom plugins. The update marks a significant expansion into full-scale music production software and arrives alongside new industry licensing partnerships with BMG and Warner Music Group. However, these developments come as the company continues to navigate ongoing copyright litigation and legal challenges from major labels, including Universal Music Group and Sony Music.

Lean TECHniques works alongside Midwest companies to leverage technology for business impact. Through advisory, coaching, and embedded delivery, Lean TECHniques helps organizations develop high-performing teams, create environments that fuel innovation, and deliver best-in-class customer experiences. Core areas of expertise include software, cloud, artificial intelligence, product management, and technology leadership.

ShieldFont is an open-source project that lets writers protect their content from unauthorized AI scraping without hiding it from human readers. A custom typeface (built with Danish foundry Playtype) swaps key content words in your text — nouns, verbs, adjectives, adverbs — with other real words right in the HTML source code. The font itself then "translates" those words back to the original on screen, so human readers see your actual writing. Any AI scraper reading the raw HTML sees fluent-but-false sentences instead.

We (Snider and his 8-year-old son) spent part of our weekend turning pets (and people) into Pokemon cards. Here’s one of our border collie Finn.

Professor Snider is commissioner of a few fantasy football leagues, meaning he needs a tool to manage who is in each league, track whether they have paid their dues, manage draft dates and write email communications to league members. In years past, this was done on a spreadsheet. But it’s 2026, so Snider used Lovable to vibe-code a tool that helps him to all of this. We won’t share the link publicly (because there’s a cost associated with using the tool), but we are looking for a couple people to test the app. Reply to this email if you’re interested.

Transformer: The “T” in GPT. It’s the neural network architecture behind basically every major AI language model today (GPT, Claude, Gemini, you name it). Introduced by Google researchers in 2017, the key trick to transformers is “attention”: instead of reading a sentence word by word in order, the model looks at all the words at once and weighs how much each one matters to every other one. That's how knows “it” refers to “the dog” three words earlier, or that “bank” means something different in “river bank” versus “bank account.”

Before transformers came on the scene, AI language models processed text sequentially and struggled with long-term context. By contrast, transformers process everything in parallel, which is both why they understand context so well and why they scale so efficiently on modern hardware that is designed specifically for parallel computation. That scalability is a big part of why the last few years of AI progress happened as fast as they did: throw more data and computing power at the same architecture, and it keeps getting better at processing language.