Rob Phillips

I don’t know why there is any surprise. Here’s OpenAI’s product strategy for the next 2 years: - you will be able to upload anything to ChatGPT - you will be able to link any external service like Gmail, Slack - ChatGPT will have persistent memory, no more multiple chats unless you want it - ChatGPT will have a consistent, user customizable personality including political bias - ChatGPT will be able to respond by text, voice, images (diagrams and video still ?? In this timeframe) - ChatGPT will become much much faster until you feel it’s a real person (>50ms response time) - Hallucinations and non factual errors will decline rapidly - as self moderation improves, question rejection will decline

This may be the most controversial thing I’ve posted. But I think it needs to be said. We are having an urgent conversation about slowing AI in medicine. Instituting rigorous safety measures. Thorough vetting before deployment. Many respected voices are calling for caution, and their instincts are grounded in a tradition of patient safety that I deeply respect. But I want to pose a question that I haven’t seen anyone ask. What is the cost of slowing down? Not the cost to technology companies. The cost to patients. We talk about AI safety as if the alternative is a well-functioning system. It isn’t. The current system produces error rates that have barely improved in decades. M&M cases that rarely lead to broad physician education. Community hospital physicians reliant on self-education of highly variable quality. Emergency physicians who never receive feedback on whether their practice patterns are calibrated — whether they order too many CTs or too few, admit too aggressively or too conservatively. Practice patterns that drift with fatigue across a single shift and across an entire career. These aren’t hypothetical harms. They’re the measured, documented, persistent background rate of medical error that we have normalized. Half of chronic disease medications aren’t taken correctly. Twenty percent of prescriptions are never filled. Up to half of adverse drug reactions are preventable. Thirty to eighty percent of hypertension patients discontinue treatment within the first year. This is the baseline. This is what we’re protecting when we slow AI deployment. Calculus measures continuous change. If we modeled the rate of improvement in patient care as a function over time, slowing AI adoption doesn’t just delay improvement by a fixed amount. It changes the integral. The cumulative patient harm prevented shrinks. Every month of delayed deployment represents ongoing harm from errors that a more capable system could have caught. We aren’t comparing AI with safety checks versus AI without safety checks. We’re comparing AI deployed in 2028 after rigorous vetting versus the current system continuing to produce the same error rates it has produced since 2000. The question is whether the cumulative harm from that delay exceeds the harm AI might introduce. I am not arguing against safety measures. I’m arguing that the cost of delay must be measured against a baseline that is far worse than most people acknowledge. We apply compassionate use and emergency authorization frameworks to drugs when the background mortality rate justifies accelerated deployment. We should at least ask whether AI in medicine has reached that threshold. The instinct to slow down feels responsible. But if slowing down means patients continue dying from errors that AI could prevent — errors we’ve failed to fix for decades through every other means — then the calculus of caution isn’t as simple as it appears. Sometimes the most dangerous thing you can do is nothing.

incredible that we built all this RAG and vector database stuff and it turns out that grep from 1973 works better than all that

Introducing Moldable, the world's first personal software factory. You have powers now, go use them. + guess what? Everything's local on your machine... so you own the data, the conversations, everything. Yeah, I know. This is like Claude Code for the rest of us.

Your computer, finally personal. Today we're launching Glaze, the second product in Raycast's history. It's a big moment for us, and I want to share the thinking behind it. Something is fundamentally changing about software. We see it every day inside our own team. People who never wrote a line of code are now contributing directly to our codebase. The barrier between "having an idea" and "making it real" is collapsing. And that changes everything. For six years, we've obsessed over what makes a great desktop app. The speed. The polish. The feeling of something that truly belongs on your computer. We've poured that into Raycast, and hundreds of thousands of people use it every day. But all that knowledge was locked inside our team. With Glaze, we're commoditizing it. Everything we've learned about building beautiful, capable desktop apps is now available to everyone. Tell Glaze what you want and it builds a real app that lives in your dock or taskbar. It launches instantly, works offline, and taps into the full power of your desktop. Beautiful by default and personal when you want it to be. It's fun for individuals and works just as well for teams. Our support team built a Glaze app connected to GitHub that runs their entire extension review workflow. Others have built dozens of internal tools. When you can shape software around how your team actually works, everything clicks. Here's what gets me most excited: we think Raycast becomes even more important in a world full of Glaze apps. Glaze apps will be deeply integrated with Raycast, connecting them all together in ways nobody else can do. The two products make each other better. A small team started building Glaze from scratch last summer. What they've shipped in that time still blows my mind. When we started Raycast, we set out to change how people use their computers. Glaze is the next chapter of that mission. We're opening the private beta today, March 4th. Mac only to start. Existing Raycast users will get priority access soon. We can't wait to see what you create and I’ll share some of my apps over the next couple of days. 💠

Introducing Moldable, the world's first personal software factory. You have powers now, go use them. + guess what? Everything's local on your machine... so you own the data, the conversations, everything. Yeah, I know. This is like Claude Code for the rest of us.