Samuel Roy

The PHP critical remote exploits details and patches are out: php.net/index.php#2026-10-20… Under version 8, you can reduce your exposure with some settings but it's far from ideal

Here's a smart home demo with @typesafeai and super nice explanations by @allietheicon They call the pattern shown in this video Speculative Fan-out: Send a long list of questions, get the answers in one go. It's fast and typed so you can use the response in your code with confidence. I can’t wait to get my hands on it, well done guys!

One chat to drive all your coding agents Starting with Claude Code, ChatGPT, Cursor and OpenCode.

Having diffs in Slack was not on my bingo card

When should you use Gemini 3.6 Flash or Gemini 3.5 Flash-Lite? ... In a fantasy world where DeepSeek V4 and GPT 5.6 Luna don't exist?

When OpenAI is shipping your last week's features... ...but you can plug in any model, including your local favorites 🙌 Whether with Alter or ChatGPT, you'll love the picture-in-picture mode. It's the best way to stay in control while an AI agent works on your behalf on your computer. This is especially useful when exploring Computer Use capabilities. Stop it immediately and steer it before catastrophic failures happening.

Understanding the code is ownership. This is a great writeup on patterns a dev can use to avoid drowning in diffs. On my side, I have two scheduled actions: 1. At the end of the day, a wrap-up of the most significant things that happened in the codebase 2. The next morning, a quiz to ensure I fully grasped the changes and their meaning. If I'm wrong, I'm quizzed again the next day until I get it. My previous flow was generating HTML reports with charts, but the micro-worlds idea is much better. Well done @geoffreylitt!

Guys, if you want both the perfect monitor for your mac device and the perfect AI assistant for macOS, you know what to do. We partnered with BenQ to offer 3 months free of @alterhq_ if you buy the new MA320UG monitor, a 32" 4K matching your macbook colors. Thanks @BenQAmerica !

Did you notice the engraved quote on OpenAI’s first chip? 1. It's the last sentence of last year “Gentle Singularity” post by Sam Altman. A must read. 2. It tells a lot about what Jalapeño really is: the first stone in OpenAI’s compounding infrastructure flywheel. 3. If intelligence is going to become abundant, its cost has to collapse. The long-term target is having intelligence priced close to the cost of energy. Custom silicon is how OpenAI starts compounding that flywheel: better chips → cheaper inference → more usage → more products/agents → more infra → lower cost of intelligence. It’s a physical step toward the world Sam described: smoothly, exponentially scaling intelligence. Am I right @sama?

Apple container has been the default code runner in Alter for the past 4 months, helping thousands of users run AI-generated code safely. For every macOS builder out there in need of a sandbox, 100% recommended.

Xcode 27 beta 2 released and it will make loop engineering easier for iOS and macOS apps. Loop engineering is basically: stop prompting the agent for every step, and give it a workflow where it can act, observe, fix, and continue. For iOS/macOS apps, the useful feedback loop usually sits outside the repo. In the simulator. In SwiftUI previews. In screenshots. In runtime state. In debugger output. In crashes, hangs, launches, energy, and disk writes. Xcode had MCP support since 26.3, but Xcode 27 expands access to the parts agents could not easily reach by only reading files. Agents can boot simulators, install and launch apps, synthesize touch events, capture screenshots, render SwiftUI preview variants, get richer preview snapshot context, access app diagnostics, and use LLDB through lldb-mcp. So the loop can move from: prompt -> edit files -> wait for human QA to: prompt -> edit -> build -> launch -> interact -> screenshot -> inspect -> debug -> fix -> repeat It should check that the UI works in dark mode. That large type does not break the layout. That the interaction actually behaves correctly. I’m really curious to see how far we can push this. At the same time, I’m a bit worried about the performance side. Xcode is already slow and heavy today. If you start running multiple agent loops, with builds, simulators, previews, screenshots, diagnostics and LLDB sessions in parallel, your Mac can get cooked pretty fast. Especially on smaller machines. An M2 Pro with 16GB RAM is still a very common dev setup. Will that loop stays usable on a real codebase, on a real machine, while you have multiple git worktrees? Let's find out

Fugu Ultra Available now (but not for Europeans yet), pricing details: Input $5 / Output $30 / Cached Input $0.50 For context greater than 272k tokens, pricing is 2x for input and $45 for output. So output is $5 more expensive than Opus 4.8 which is the most expensive frontier model they use. It's also perfectly matching gpt 5.5 pricing. Kind of puzzling but people might be ready to pay if the results are at the Fable 5 level. At the end, it's cheaper than Fable 5 by 2x. But what's interesting here, is the architecture. Having a pool of models means XX% of your requests goes to cheaper models. So there's a place for other players to enter the game and play on that margin to offer competitive alternatives at a cheaper price for the same level of intelligence. After all, Gemini 3.1 pro is only $2/$12 for context under 200k. If you're able to handle 50% of queries with it, you're on for a pretty good margin.

🎬 Build Your AI Meeting Report System in 6 Minutes 1. Automatic meeting recording with Alter No bots, speaker identification, everything local on macOS, even records Facetime calls 2. A meeting agent triggers and generates 4 reports from the transcription: Overview, Team Dynamics, Technical Breakdown, and Personal Coaching 3. Everything is saved in your Obsidian vault, ready to be consumed or used by another agent Full breakdown in video

🎬 Siri AI on macOS 27 demo replay 1. Summon from Spotlight 2. Can analyze files 3. Available through the contextual menu (right-click) 4. Work with your Apple apps, like drafting emails in Mail

Why Siri AI is allowed on macOS but not on iOS for Europe? It doesn't make any sense. Either you give access or you don't. You get the same level of access, the same security risks, the same data. Your Mac device is even more at risk than your phone. More files, documents, and plenty of developers with infrastructure keys and secrets. Whole businesses running on Mac. Also a perfect place to run unsafe code. Can someone explain?

Tip if you’re not in a Cursor team: once you’re happy with your canvas, ask Cursor to save it as an html file. Same style, same everything and you can do what you want with it.

Excited to see how it behaves with Alter on Computer use and tool planning

Adding Minimax M3 and Opus 4.8 benchmark numbers for comparison.