Anthropic is taking an unusual approach to AI safety. It released a dark warning about AI risks, alongside an updated ethical “constitution” written by its in-house philosopher, Amanda Askell. That document tells its AI, Claude, to use its own judgment when balancing usefulness, honesty, and safety. In plain terms, Anthropic is betting that AI can develop something like “wisdom.” The twist: the company is letting the AI help decide what ethical behavior looks like.

For builders, the stakes are clear. If this works, AI companies may be able to grow faster without relying on rigid safety rules that break easily. If it fails, regulators could crack down harder, legal risks could rise, and customers may start demanding proof that systems are being tested and governed. The message is simple: build safety into your product now—with clear controls, shutdown options, and response plans—because shipping without guardrails won’t be an option for long.

Your TAM (Total Addressable Market) is not the size of the problem. It's not a percentage of a trillion-dollar industry. And it's definitely not a number you can pull from a Google search and shrink down to fit your pitch.
In this video, we break down exactly how to build a solid TAM slide for your pitch deck—one that investors will actually take seriously. We’ll debunk common mistakes, walk through bottom-up calculations, and show you how to estimate SAM and SOM the right way.

Juries are about to hear the first major test case claiming that social media features like autoplay, infinite scroll, and algorithm-driven feeds are addictive to kids and harm their mental health. Instead of blaming specific posts, the lawsuits focus on product design—basically arguing that platforms should slow down engagement, not police every piece of content. Snap and TikTok have already settled. Meta and Google deny the claims and say they’ve built strong protections for teens.

If jurors agree with the plaintiffs, the case could force companies to turn over internal documents and give lawyers leverage to push for large settlements and new “safety-by-design” rules through courts and lawmakers. If the companies win, it could shut down similar lawsuits and keep recommendation algorithms protected under free-speech and Section 230 laws. Either way, the outcome will shape how much responsibility courts place on social media platforms to protect kids.