Issue 322: Moving too fast and breaking too much

Tech's been moving fast. But you gotta wonder…is moving fast actually working or just breaking things that can’t be fixed? This week, we’re taking a look at what’s moving and what’s working, which are sometimes not the same things. On the pod, HumanX’s Stefan Weitz helped us figure out if 2025 was really the year of AI agents, or if the hype outweighed the growth. Speaking of AI agents, we have a blog article on the expense of AI becoming your second brain (spoiler: you pay with your first brain). Plus, we have stories from around the web about frustration as a product and AIs escaping sandboxes, which will make you wonder if the Internet is broken or if that’s just part of its product design nowadays.

But not everything is moving fast and breaking itself. On Leaders of Code, Netlify’s Dana Lawson shared with our CPTO Jody Bailey how they make their distributed engineering teams work successfully with the help of AI agents. And if you’re ever in doubt about what does and doesn’t work in the tech world, you can always look to open source. Chainguard’s Dan Lorenc joined the podcast to discuss keeping open-source projects alive by moving fast and fixing them. If you don’t want to break your codebase, maybe consider slowing down with your AI agents, as recommended by one dev in our stories from the web this week. You can take a page from ENIAC’s engineers, who are celebrating its 80th anniversary this week.

And on our sites? Well, our users are wondering if they’re allowed to move fast and break things. Are you breaking your brain by using AI to understand research papers? Are unpaid contractors allowed to break the things they’ve built? Is this steampunk world moving too slow by not having electricity? You better get a move on, because we have all those answers and more already down below.