Hacker News: igor47

New comment by igor47 in "Our position on open-weights models"

igor47 — Mon, 27 Jul 2026 23:38:14 +0000

Addressed in the source, but there's an asymmetry favoring the attackers. They only have to find one exploit once. The defenders have to be perfect all the time.

New comment by igor47 in "Jellyfin founder Andrew leaves team"

igor47 — Tue, 21 Jul 2026 02:18:05 +0000

I wonder if it was just a start up issue? Some kind of initial scan of your database?

New comment by igor47 in "Inside the 24 hours that led to pulling Fable 5"

igor47 — Mon, 15 Jun 2026 06:11:49 +0000

But it's not though. The root of the problem is politics, and the root of that problem is self -interest.

New comment by igor47 in "SpaceX: Flying High on Impunity"

igor47 — Fri, 05 Jun 2026 05:54:21 +0000

I'm pretty mad about the blatant manipulation of index funds by spacex, and I looked into how to avoid the IPO. If you're like me, the amount of unrealized gains from index funds over the past few years mean the tax penalty of exiting these funds is just too big (the majority of my money is not in tax advantage accounts like a 401k). If not for the tax reasons, I would probably go with a direct index. Vanguard and Fidelity both have em. But honestly even at inflated float ratios this stock makes a very small part of the overall index. What I hate is forcibly being made party to such blatant manipulation, but like any other "concentrated benefits, diffuse costs" situation, it's probably too low a real cost for most people to deal with.

New comment by igor47 in "Amazon Has Axed Its New Stargate Series"

igor47 — Wed, 03 Jun 2026 14:58:23 +0000

Multiple comments on this thread about how bad it would be if the main character was a woman. Why is that so bad?

New comment by igor47 in "Preparing for KDE Plasma's Last X11-Supported Release"

igor47 — Tue, 02 Jun 2026 15:52:39 +0000

Yeah agreed. I switched to kde from gnome a few months back, and it's amazing how much better it's been in a thousand little ways.

New comment by igor47 in "Defeating Git Rigour Fatigue with Jujutsu"

igor47 — Mon, 25 May 2026 04:14:19 +0000

Are you me? I do feel like I'm starting to forget git as a result of my happy jj use. Thankfully some repos use git submodules, which keeps me at least a little connected

New comment by igor47 in "Defeating Git Rigour Fatigue with Jujutsu"

igor47 — Mon, 25 May 2026 04:11:28 +0000

When I'm working in git, I always start work by creating a new branch with a name. Sometimes the branch becomes something different as I work and then I might rename it or more often just keep a stale name around. But in git commit descriptions come later.

In jj, it's the opposite. I start with a change, and I often describe it right away. Branches (bookmarks) come at the end.

You could, in jj, tag a new empty change with a bookmark as soon as you create it. You don't have to advance the bookmark -- that the first change in a sequence of changes is tagged with a bookmark is probably as much information as you need?

New comment by igor47 in "Code review is not about catching bugs"

igor47 — Sun, 24 May 2026 02:43:20 +0000

Code review is about getting a team on the same page more broadly. The question posed in the article -- whether the change should be part of the product -- is certainly part of that. It's important to have a shared understanding on a team re what the code actually does and should do. But you can also use code review to align on architecture, best practices, road map.

I also find that code review helps a team feel like a team. It's nice to know anyone actually cares enough to pay attention to what you do, and feedback is an amazing gift that's really underrated.

New comment by igor47 in "Launch HN: Superset (YC P26) – IDE for the agents era"

igor47 — Sun, 24 May 2026 01:06:15 +0000

"just" is doing a lot of heavy lifting here. I always automate my repo setup, usually with mise tasks, but there's often assumptions baked in, eg around which ports the app will listen on. It's also still hard to get the agents to use "mise run test" instead of invoking pytest directly. This sounds harder to deal with when you have multiple copies of the test or dev stack running together on the same machine.

New comment by igor47 in "Web Serial Support in Firefox"

igor47 — Fri, 22 May 2026 00:59:32 +0000

Nice. I ran into this issue a few months ago:

https://igor.moomers.org/posts/wled-christmas-lights#flashin...

Nice to hear it works now without the very hacky work around

New comment by igor47 in "Where to buy a non-Apple, non-Google smartphone"

igor47 — Sun, 17 May 2026 00:18:39 +0000

Or maybe just an android emulator? I wonder if this would actually work...

New comment by igor47 in "The hypocrisy of cyberlibertarianism"

igor47 — Sat, 09 May 2026 19:17:38 +0000

Life without a smartphone increasingly challenging. You have to use either Google or Apple. I use a de googled Android lineage phone but this is always getting harder, as numerous threads on this site will attest. Plus literally every employer I've ever had has used Google services, plus lots of other sites I might have to use implement recaptcha or otherwise invisibly to me share my data or data about me with Google. Also, even if I do figure out a way to stay off Google's radar, they're a powerful force which shapes my world. They hire lobbyists to influence policy which affects me, build data centers which raise my cost of electricity, or sell killer robots to evil people.

I think where people go wrong is treating Google the way they treat their weird neighbor Bob. Bob's damage is limited. Google is an immense, powerful, alien entity, far beyond the control of any person, and with its own inscrutable goals which are the not goals of literally any person alive or dead.

I genuinely don't understand the desire to leave this entity unmoored to wreck what havoc it may.

New comment by igor47 in "The hypocrisy of cyberlibertarianism"

igor47 — Sat, 09 May 2026 17:46:36 +0000

But in a democracy, you at least have input! Google is also a coercive force with no real checks on its power, but it doesn't care about anything you have to say. That's the difference, that's it, right there. The answer to abuse of power is not to just unleash raw power, its to subordinate and restrict it. That's what government is for. When you find yourself arguing that power you participate in is bad and shouldn't restraint power you have 0 influence in, that's when it's time to wonder if they've gotten to you.

New comment by igor47 in "The hypocrisy of cyberlibertarianism"

igor47 — Sat, 09 May 2026 17:41:49 +0000

On this side of the wall, you and your friends are strong and happy and free in your garden. On the other side, a hellscape filled with giant monsters debating how best to filet you. You will keep ceeding them ground, your garden gets ever smaller. The monsters ate Brian, oops, well that's the consequence of freedom! But you're next, isn't it completely obvious you're next? Why would you unilaterally disarm against the monsters? Why for the love of God why would you say "no the monsters are good actually!"

New comment by igor47 in "The hypocrisy of cyberlibertarianism"

igor47 — Sat, 09 May 2026 17:33:03 +0000

Strong agree. That passage seems to me to be decrying the friction of the real world, whereas it's become increasingly clear to me just how valuable friction is in the world, and how inextricably tied the tech companies war on friction to the bad outcomes technology seems to engender.

New comment by igor47 in "Higher usage limits for Claude and a compute deal with SpaceX"

igor47 — Thu, 07 May 2026 01:06:28 +0000

No it's really bad

New comment by igor47 in "Show HN: Home Memory – A local DB of my house, down to cables and pipes"

igor47 — Fri, 17 Apr 2026 15:19:30 +0000

Yeah this is a killer app for mcp servers. My puppy has food allergies and the vet asked me to track his eating and pooping. I ended up building an mcp server to do this data entry and now I track his activity and planning to track also his training progress. It's very different when you can just tell your model "he ate 2 cups of his kibble" or "we practiced stay for 5 minutes" vs doing tedious data entry. As a bonus this helps my partner and me coordinate dog care so we can have fewer conversations about the dog

New comment by igor47 in "Moving a large-scale metrics pipeline from StatsD to OpenTelemetry / Prometheus"

igor47 — Thu, 16 Apr 2026 14:46:45 +0000

A long time ago, I introduced dogstatsd at Airbnb. We had already been using vanilla statsd (with no tag support -- cardinality lived in the metric name!) and this was a low cost migration. More than a decade later, I'm assuming it was difficult to track down and refactor all the places that statsd calls were emitted and using OTLP was an easier route. This is a great example of how technical decisions compound over time.

New comment by igor47 in "Moving a large-scale metrics pipeline from StatsD to OpenTelemetry / Prometheus"

igor47 — Thu, 16 Apr 2026 14:41:03 +0000

So is Victoria metrics?