Anthropic jumps in capabilities, the FTC is worried about kids and AI, and a public assassination reminds us we have more pressing challenges than AI taking all of our jobs.

Anthropic’s Ups and Downs

Down: Anthropic’s landmark $1.5B settlement offer for copyright infringement was expected to be accepted this past week…but it wasn’t. Instead a federal judge said no, so now we’ll have to wait at least another two weeks to see what happens.

Up: Claude now works directly with Office documents – and can create and modify Word, Excel, PowerPoint, and PDF files. This is big, because it moves well beyond the chat interface and into work that people actually need to get done. If you use Office as part of your work, I highly encourage you to watch the brief video at the link. Although I’m sure actual results won’t always be as good as the demo, this is impressive stuff.

Up: And Microsoft took notice. According to preliminary reports, Microsoft has decided to partner with Anthropic to include Claude in Microsoft Copilot. Specifically because Claude’s ability to work with Microsoft Office documents is better than OpenAI’s GPT-5.

Agent-Washing Becomes a Self-Fulfilling Prophecy

Google Cloud released their annual report, the ROI of AI 2025 based on a survey of 3,466 leaders, mostly C-level, in companies with $10M+ in revenue. The survey reported that 52% of companies using genAI have deployed AI agents.

Ridiculous! Anyone following this closely knows that true AI agent adoption in the workplace is in the single digits.Of course, if we count Claude’s newfound ability to work with documents as an AI agent, that number might grow pretty quickly.

It’s not (all) Google’s fault – their definition of an AI agent is a good one. The leaders (mostly C-level) either didn’t read it, didn’t understand it, or are so absorbed in the hype they didn’t want to sound like they were behind. An assistant? Not an agent. RAG? Not an agent. A GPT? Not an agent. An LLM with a custom prompt? Definitely not an agent.

“Agentic AI has become one of the most discussed — and most misunderstood — topics in the AI space. Influencers pushing content and startups selling complexity have flooded the term with noise. As a result, most people talking about agents are either guessing or bluffing.” [emphasis mine]
– Devansh, Artificial Intelligence Made Simple

True AI agent use in enterprises is not 50%, it’s nascent. The tech and its capabilities are well ahead of adoption. Plus, models aren’t reliable enough for applications where accuracy is crtitical, which is most applications in business. For now, agents are mostly limited to coding, customer service, and Deep Internal Research (Deep Research but on internal content rather than the internet).

Adoption will continue to increase, but not at the rate the startups or AI labs are promising.

The FTC Wakes Up

It’s been pretty clear that these AI chatbots are problematic because of harmful information, engagement maximization, persuasive ability, and sycophancy (insincere flattery). This is particularly true for children. Good news: the FTC wants to know what the companies peddling these products are doing to protect kids from harm. The FTC sent a letter to seven of the major US chatbot companies (since they’re focusing on apps used by children, notably Anthropic was not one of the seven) asking for information about design decisions and efforts to protect children. Some of the specifics requested include how they monetize user engagement, screen for negative impacts, and monitor and enforce compliance with usage policies (i.e., underage use).

It’s about time! Regulation won’t solve this problem, but it’s clear that it’s needed. We’ve seen ample evidence that these companies are not applying adequate protections themselves.

Parents: Don’t look to the government to protect your kids. Stay involved. Know what they’re using on their devices. Talk to them about AI and especially AI “companions.”

An Even Better Image Model

The model wars continue. Bytedance (the owner of TikTok) released an updated version of its image generator, Seedream 4.0. Rumor has it that it’s better than the best, Google’s Imagen 4 Ultra, which was released just a couple of weeks ago. Imagen is technically in beta, we’ll see if they make enough improvements for the official release to best Seedream.

---

---

I’ve already explained what matters above. This week, there’s something else that matters much more: how the people of this great country will choose to deal with evil.

The public assassination of Charlie Kirk last week was a tragedy. Violence is not the answer to resolving differences. Charlie was the embodiment of the answer. The answer is the foundation of this country: free speech and open debate. This was so important to the founding fathers that it’s the first amendment to the Constitution.

(Note: if you’re thinking, if free speech was so important, why wasn’t it put in the Constitution in the first place? It’s because the founders debated whether it was appropriate for the Constitution to list rights at all. Some felt doing so would imbue more power to government than warranted. But they put their division side and added the Bill of Rights – the first 10 amendments to the Constitution – three years after the Constitution was ratified.)

It’s as ironic as it is sad that Charlie’s life was cut short by someone who was so upset that he couldn’t win the debate, he decided to remove the ability to debate altogether.

So I’m going to use a word that’s not used very often, but we have to acknowledge it if we want this Union to last another 250 years: evil. That is the word that describes what took place at Utah Valley University on Wednesday.

The assassin was evil, because his motivations and his acts were evil. I think Charlie himself would have agreed that we are all capable of evil, and it is only faith in Jesus Christ that can redeem us of that evil. My prayer is that his killer accepts Jesus’ offer and claims that redemption. Alongside my prayers for the Kirk family and for Charlie’s legacy to live on.

I want the country that Charlie envisioned. The country that he was trying to create. I want the glorious USA to be a country where we can disagree without being divisive, where we can debate without violence, and where we can dissent without hate.Do not take your freedom of speech lightly. It is what allows us to move forward, explore new ideas, to become better as individuals and better as a nation. It is a rare commodity, and it was purchased at a high price. Let’s all use that freedom of speech to build up, not to put down; for good, and not for evil.

“I may disapprove of what you say, but I will defend to the death your right to say it.”
– Evelyn Beatrice Hall, in her biography summarizing Voltaire’s beliefs on the matter.

Clearly, Charlie Kirk would agree. He embodied it every day. May we live up to his worthy example.