<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: mgaunard</title><link>https://news.ycombinator.com/user?id=mgaunard</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 13 May 2026 06:37:03 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=mgaunard" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by mgaunard in "Why senior developers fail to communicate their expertise"]]></title><description><![CDATA[
<p>Even with AI, there is a clear difference between juniors and seniors.<p>None of the things I can think of have anything to do with avoiding problems.<p>To some degree, having 5+ agents working on different projects is similar to leading a team of 5+ people. The skills translate well.<p>The senior is also able to understand what the agents do, review and challenge it. Juniors often can't.<p>And finally, the senior has a deeper understanding of what the business and problem domain are, and can therefore guide the AI more effectively towards building the right thing.</p>
]]></description><pubDate>Tue, 12 May 2026 18:27:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=48112286</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=48112286</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48112286</guid></item><item><title><![CDATA[New comment by mgaunard in "Why TUIs are back"]]></title><description><![CDATA[
<p>You're not switching, it's all in the same pane, doesn't mess with windowing or focus, your keyboards shortcuts still work etc.</p>
]]></description><pubDate>Mon, 04 May 2026 09:17:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=48006327</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=48006327</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48006327</guid></item><item><title><![CDATA[New comment by mgaunard in "Why TUIs are back"]]></title><description><![CDATA[
<p>As I said, the advantage of the TUI is that they avoid context-switching.</p>
]]></description><pubDate>Mon, 04 May 2026 01:54:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=48003754</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=48003754</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48003754</guid></item><item><title><![CDATA[New comment by mgaunard in "Why TUIs are back"]]></title><description><![CDATA[
<p>Power users have always preferred the command-line, since expressing what you want to do as a programming language is of course much more powerful and productive than clicking menus.<p>To avoid context-switching from the command-line, many essential UIs were made text-only. Another route would have been to integrate the command-line within graphical applications, but few did it -- the main example that comes to mind is Jupyter.</p>
]]></description><pubDate>Sun, 03 May 2026 20:54:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=48001371</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=48001371</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48001371</guid></item><item><title><![CDATA[New comment by mgaunard in "How fast is a macOS VM, and how small could it be?"]]></title><description><![CDATA[
<p>That's also what Colima does.<p>OrbStack isn't open-source though and I can't justify buying a license for every single person in my company just for something functionally equivalent but performing better.<p>These kinds of things should just be provided by Apple as a first-class thing.</p>
]]></description><pubDate>Sat, 02 May 2026 19:51:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=47989838</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47989838</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47989838</guid></item><item><title><![CDATA[New comment by mgaunard in "How fast is a macOS VM, and how small could it be?"]]></title><description><![CDATA[
<p>My only experience with VMs on macOS is colima+docker, and it's relatively painful and inefficient (but usable).</p>
]]></description><pubDate>Sat, 02 May 2026 11:38:01 +0000</pubDate><link>https://news.ycombinator.com/item?id=47985504</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47985504</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47985504</guid></item><item><title><![CDATA[New comment by mgaunard in ""People who don't use AI will be left behind""]]></title><description><![CDATA[
<p>I find that good people get better with AI, but I'm not sure more average people really do.<p>I've seen some produce stuff without really understanding it, barely review anything, and pretty much suffer from imposter syndrome.</p>
]]></description><pubDate>Wed, 29 Apr 2026 19:57:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=47953683</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47953683</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47953683</guid></item><item><title><![CDATA[New comment by mgaunard in "Zero-copy protobuf and ConnectRPC for Rust"]]></title><description><![CDATA[
<p>The converter is generated automatically based on the differences between the two schemas.<p>Takes zero effort other than CPU cycles.</p>
]]></description><pubDate>Fri, 24 Apr 2026 18:59:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=47894426</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47894426</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47894426</guid></item><item><title><![CDATA[New comment by mgaunard in "Zero-copy protobuf and ConnectRPC for Rust"]]></title><description><![CDATA[
<p>I already said you can UUIDs and schemas, and even dynamic conversion between mismatched schemas.<p>Doing plain C structs doesn't prevent any of this.</p>
]]></description><pubDate>Wed, 22 Apr 2026 11:17:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=47861953</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47861953</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47861953</guid></item><item><title><![CDATA[New comment by mgaunard in "Zero-copy protobuf and ConnectRPC for Rust"]]></title><description><![CDATA[
<p>You can also just use an array which sets a max capacity, and either use a null-terminator or a separate size field.<p>In practice you probably want to have both, and choose what's most practical based on the message.</p>
]]></description><pubDate>Wed, 22 Apr 2026 11:15:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=47861923</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47861923</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47861923</guid></item><item><title><![CDATA[New comment by mgaunard in "Zero-copy protobuf and ConnectRPC for Rust"]]></title><description><![CDATA[
<p>I handle it in different ways by topic.<p>For topics which are sending the state of something, a gap naturally self-recovers so long as you keep sending the state even if it doesn't change.<p>For message buses that need to be incremental, you need to have a separate snapshot system to recover state. That's usually pretty rare outside of things like order books (I work in low-latency trading).<p>For requests/response, I find it's better to tell the requester their request was not received rather than transparently re-send it, since by the time you re-send it it might be stale already. So what I do at the protocol level is just have ack logic, but no retransmit. Also it's datagram-oriented rather than byte-oriented, so overall much nicer guarantees than TCP (so long as all your messages fit in one UDP payload).</p>
]]></description><pubDate>Mon, 20 Apr 2026 09:34:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=47832054</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47832054</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47832054</guid></item><item><title><![CDATA[New comment by mgaunard in "Zero-copy protobuf and ConnectRPC for Rust"]]></title><description><![CDATA[
<p>It's 2026 and I'm still defining my own messaging and wire protocols.<p>Plain C structs that fit in a UDP datagram that you can reinterpret_cast from is still best. You can still provide schemas and UUIDs for that, and dynamically transcode to JSON or whatever.</p>
]]></description><pubDate>Mon, 20 Apr 2026 09:12:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=47831895</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47831895</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47831895</guid></item><item><title><![CDATA[New comment by mgaunard in "Show HN: I built a Cargo-like build tool for C/C++"]]></title><description><![CDATA[
<p>I find that conan2 is mostly painful with ABI. Binaries from GCC are all backwards compatible, as are C++ standard versions. The exception is the C++11 ABI break.<p>And yet it will insist on only giving you binaries that match exactly. Thankfully there are experimental extensions that allow it to automatically fall back.</p>
]]></description><pubDate>Thu, 09 Apr 2026 19:21:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=47708474</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47708474</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47708474</guid></item><item><title><![CDATA[New comment by mgaunard in "S3 Files"]]></title><description><![CDATA[
<p>It also means that you need to pay for EFS, which is outrageously expensive, to use S3, whose whole purpose is to be cheap.</p>
]]></description><pubDate>Wed, 08 Apr 2026 06:41:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=47686225</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47686225</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47686225</guid></item><item><title><![CDATA[New comment by mgaunard in "S3 Files"]]></title><description><![CDATA[
<p>Zero mention of s3fs which already did this for decades.</p>
]]></description><pubDate>Tue, 07 Apr 2026 20:11:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=47680740</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47680740</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47680740</guid></item><item><title><![CDATA[New comment by mgaunard in "AWS engineer reports PostgreSQL perf halved by Linux 7.0, fix may not be easy"]]></title><description><![CDATA[
<p>Well, you can always pin to a core and move other threads out of that core.<p>That's what you'd do if manually scheduling. Ideally the dynamic scheduler would do that on its own.</p>
]]></description><pubDate>Sun, 05 Apr 2026 21:57:13 +0000</pubDate><link>https://news.ycombinator.com/item?id=47654308</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47654308</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47654308</guid></item><item><title><![CDATA[New comment by mgaunard in "AWS engineer reports PostgreSQL perf halved by Linux 7.0, fix may not be easy"]]></title><description><![CDATA[
<p>The scheduling is based on how much the LWP made use of its previous time slices. A spinning program clearly is using every cycle it's given without yielding, and so you can clearly tell preemption should be minimized.</p>
]]></description><pubDate>Sun, 05 Apr 2026 21:55:37 +0000</pubDate><link>https://news.ycombinator.com/item?id=47654291</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47654291</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47654291</guid></item><item><title><![CDATA[New comment by mgaunard in "AWS engineer reports PostgreSQL perf halved by Linux 7.0, fix may not be easy"]]></title><description><![CDATA[
<p>Most people looking for performance will reach for the spinlock.<p>The expectation is that the kernel should somehow detect applications that are spinning, and avoid preempting them early.</p>
]]></description><pubDate>Sun, 05 Apr 2026 06:30:52 +0000</pubDate><link>https://news.ycombinator.com/item?id=47646678</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47646678</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47646678</guid></item><item><title><![CDATA[New comment by mgaunard in "NHS staff refusing to use FDP over Palantir ethical concerns"]]></title><description><![CDATA[
<p>Getting a minor wound bandaged up is not what A&E is meant for, it's for life-threatening injuries.<p>Going to A&E and waiting there also means you're losing 4 to 6 hours.<p>111 you just get some robot asking you a never-ending list of inane questions before someone tells you to either self-care at home or go to A&E.<p>A pharmacist should be able to administer the supplies they sell, particularly wound dressing and care. It's a requirement in some other European countries like France (where pharmacists are doctors), but in the UK the reality is that most are unable to do so.</p>
]]></description><pubDate>Fri, 03 Apr 2026 23:36:54 +0000</pubDate><link>https://news.ycombinator.com/item?id=47633769</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47633769</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47633769</guid></item><item><title><![CDATA[New comment by mgaunard in "NHS staff refusing to use FDP over Palantir ethical concerns"]]></title><description><![CDATA[
<p>Not all disinfectants are child safe, and the wound was serious enough to require steri-strips (an alternative to sutures) -- it was not a matter of a bandaid.</p>
]]></description><pubDate>Fri, 03 Apr 2026 23:24:04 +0000</pubDate><link>https://news.ycombinator.com/item?id=47633651</link><dc:creator>mgaunard</dc:creator><comments>https://news.ycombinator.com/item?id=47633651</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47633651</guid></item></channel></rss>