Proton

Random musings of an engineer.

[Proton @ProgenticEngine]

@bflycomputer It looks promising for new users, with thoughtful design and easier setup. Bugs and missing features are expected at this stage. Try it on a spare computer, back up your files, and share feedback to help improve it.

[Proton @ProgenticEngine]

@DiscussingScoop Someone want to animate this using AI? 👀

[Proton @ProgenticEngine]

@abdimoalim_

[Proton @ProgenticEngine]

@thefusioncube @poteto The question is whether we need a new programming language, or a better agent verification layer over existing languages that already have enormous ecosystems.

For agentic development, I increasingly lean toward the second.

[Proton @ProgenticEngine]

@rainwaterJay Amdahl's Law, adding agents does not guarantee faster execution when coordination and verification simply increases token consumption, communication overhead, latency, and opportunities for conflicting instructions.

[Proton @ProgenticEngine]

@almonk @OpenAI @thsottiaux @gdb for iOS Duo 👀👀👀

[Proton @ProgenticEngine]

@mattpocockuk we do not need a new language

[Proton @ProgenticEngine]

@v_language Vinix is a particularly interesting example because it goes beyond replacing applications. It challenges the assumption that building an entirely new operating system is prohibitively expensive or impractical.

[Proton @ProgenticEngine]

@Edmondif143061 Use Composer, PSR-4 autoloading, explicit types, strict typing where appropriate, and validated inputs. Prefer simple classes and explicit dependencies. Handle errors clearly, bind SQL values, escape output, and enforce authorization.

[Proton @ProgenticEngine]

@thsottiaux 🤓

[Proton @ProgenticEngine]

@rustaceans_rs The opportunity is compelling: keep #PHP’s application ecosystem while gaining a safer, more controllable runtime. The remaining challenge is proving compatibility, operational resilience, and competitive performance across workloads beyond the ones already targeted.

[Proton @ProgenticEngine]

@brombinmirko @ascott2013 The most significant achievement is that this is no longer simply an effort to make an Office installer execute under Wine. Developers are reaching the point where actual application functions, Microsoft authentication, document rendering, and cloud integration can be tested

[Proton @ProgenticEngine]

@gdb Fix the iOS and macOS apps, the UI and UX need to be more beautiful, better icons, better use if mixed fonts for inline and expandable tile cards. Unify the look across all your surfaces.

[Proton @ProgenticEngine]

Frustrating that @apple @johnternus did not optimize the Setup Screen for the new #iPhoneDuo to take up the full screen width. Isn't it running iOS 27.1 to be optimized for the Duo??? #Apple

[Proton @ProgenticEngine]

As excited as I am for the iPhone Duo, the reviews have made me appreciate the delay in getting mine. A few extra weeks should give app developers time to update their apps and hopefully give @Apple time to smooth out issues like the noticeable Touch ID delay.

[Proton @ProgenticEngine]

@buildwithsid Cheating indie devs ultimately hurts users.

The hardest part of software engineering isn't building the app, it's maintaining it: fixing bugs, shipping updates, & providing long-term support.

Most indie apps cost a small annual fee. If you genuinely use the software, pay for it

[Proton @ProgenticEngine]

so Claude finally has an answer to @Muse and @ChatGPT Dots

[Proton @ProgenticEngine]

@CupOJoseph An attacker would still need to intercept or redirect traffic. But an affected device could accept a convincing fake cloud service and expose whatever that connection carries.

[Proton @ProgenticEngine]

@cedric_chee The larger product issue is transparency. Users should know when selecting Pro changes the model, reasoning resources, & execution strategy. OpenAI currently explains parts of this across multiple documentation pages but the exact ChatGPT Pro execution budget remains undisclosed.

[Proton @ProgenticEngine]

@brexton The useful test of #AI is whether assistance reduces the total effort required to reach a satisfactory outcome, including explaining the task, checking the result, and correcting mistakes. A polished demonstration of an agent clicking “buy” does not answer that question.