Hacker News: phtrivier

New comment by phtrivier in "Source code of the Estonian online voting system"

phtrivier — Sun, 18 Oct 2026 11:31:15 +0000

Yes. But I'm slightly less worried about deliberate mishaps (which tends to leave trace or organisation) than pure mistakes (which go unnoticed, except for malicious actors who are better at hiding their trace.)

Hanlon's razor, etc...

New comment by phtrivier in "Source code of the Estonian online voting system"

phtrivier — Sun, 18 Oct 2026 08:22:24 +0000

But then, let's deal with the elephant in the thread : are Estonians trusting their infrastructure enough to be sure that élections won't be tampered by a particular neighboring country with well-known skills in cyber security ?

New comment by phtrivier in "Source code of the Estonian online voting system"

phtrivier — Sat, 17 Oct 2026 20:42:09 +0000

I've been an overseer at election booths several times. I kept a good look at the transparent ballot box, along with complete strangers. In the end, once, I may even have inadvertently recognized my envelope and bulletin during the opening and counting - because we where very very few people to choose that candidate in my polling station :D And I was lucky enough to follow the president of the polling station to the town hall where armies of public servant where doing the addition, and I checked the CSV on next Monday, etc, etc...

New comment by phtrivier in "Source code of the Estonian online voting system"

phtrivier — Sat, 17 Oct 2026 20:37:44 +0000

Nope. Vote tallying is public with random overseers, in transparent boxes watched by a room of people from diverse opinions, with random inspections - precisely because we DO NOT trust the system or the people who count the vote.

Which has always been my biggest beef about digital vote. I'm all trusting the people writing voting software to do their best, actually. Being a Software engineer myself, I'm more concerned about bugs and hacks than deliberate tampering by the government. Especially in an age where AI agents can access pretty much anything ever connected to an electric plug...

New comment by phtrivier in "Source code of the Estonian online voting system"

phtrivier — Sat, 17 Oct 2026 20:32:34 +0000

So there is a tool that remembers who you voted for. That's great to convince your bully that you actually did what they told you.

However, how do you know you're one of the X votes displayed in the final tally ?

New comment by phtrivier in "Source code of the Estonian online voting system"

phtrivier — Sat, 17 Oct 2026 20:30:28 +0000

Exactly, it means someone you could prove to someone else how you voted, which is the sine qua non condition for creating a vote economy. That's actually the main reason behind trying to keep the vote secret, much more than threats of intimidation for opposing parties voters, etc...

New comment by phtrivier in "Source code of the Estonian online voting system"

phtrivier — Sat, 17 Oct 2026 19:25:19 +0000

I know the answer is "you don't", but suppose I vote vote electronically for candidate A. In the end, the candidate loses with 764 votes. How do I convince myself that my vote was counted in the 764 votes, and that it was not a randomly chosen number (with a RNG skewed around "plausible" scores) ?

New comment by phtrivier in "Tesla Roadster"

phtrivier — Fri, 16 Oct 2026 19:55:39 +0000

A though for the people doing the "boring" engineering at Tesla (like, you know, enhancing the range of the vehicles, improving security, making FSD closer to reality, etc...), who used to get some light (remember "battery day" ?), and are now on the backbench while the sycophants gets the budget for vanity projects like this. At least, at Tesla, the sycophants are _smart_ and pull out cool demos. Not the case in every "new clothes" kingdom.

New comment by phtrivier in "Incident with Actions, Pull Requests and Issues"

phtrivier — Mon, 12 Oct 2026 13:41:49 +0000

At this point, a status bar color when gh is down would be enough.

New comment by phtrivier in "Theranos.world"

phtrivier — Fri, 09 Oct 2026 12:32:36 +0000

Ok, I have to ask : is there a way to get a "cartridge" to insert into the theranos machine ?

I would love to have the machine "function" up to a point (except, actually doing the analysis results in a "NotImplementedException" of some kind.à

New comment by phtrivier in "The last time my family was replaced by technology"

phtrivier — Thu, 01 Oct 2026 11:18:46 +0000

Also, it leads to significant lower purchase power, which may indirectly decrease profits, at some point ? ("Firing your customers" works up to a point)

New comment by phtrivier in "The last time my family was replaced by technology"

phtrivier — Thu, 01 Oct 2026 11:13:14 +0000

Out of curiosity, what was the timeline ? Cause I'm afraid that, when you mentioned "I don't know what the other farriers" did, the answer may be "they were sent to the front and died by the 1000s".

Which starts to be my fear for the real "solution" to job displacement by AIs : people will just be sent to some far away conflict where they will transition to a career in "daffodil fertilization" :/

New comment by phtrivier in "Ask HN: What are you reading?"

phtrivier — Wed, 30 Sep 2026 14:32:19 +0000

I meant this one: https://www.imdb.com/title/tt26446278/

New comment by phtrivier in "Ask HN: What are you reading?"

phtrivier — Wed, 30 Sep 2026 10:42:32 +0000

Going through "Here comes the sun" by Bill McKibben - slightly disappointed / surprised, as the book is very much a pamphlet / manifesto in favor of a single idea (that I could paraphrase as "solar is good, solar is _now_ cheap, it's so cheap that it's very cheap, and so good that's it's very good, and we should build more cause it's cheap and good, and if you disagree you're a fossil fuel shill astroturfer and all the land and materials problems are not really problems, and oh, did I mentionned it's cheap and good ?") To which I'll say, _fair enough_, but I was hoping for a slightly more nuanced view of the topic.

Also started "Sutskever's list" by Richard Heimann, which goes into the history of the current AI paradigm "one paper at a time". It's written for a curious audience : people who can vaguely understand the concepts, but not go into details. That means that I kinda "get" some chapters because I already read about the topics, and I'm completely stumped by some. But the "insider baseball" history is illuminating sometimes.

New comment by phtrivier in "Ask HN: What are you reading?"

phtrivier — Wed, 30 Sep 2026 09:20:28 +0000

Just out of curiousity : the recent movie adaptation had a huuuge success in France two years ago (deservedly so, in my opinion), did that cross the Atlantic ?

New comment by phtrivier in "Analysis: EVs are now nine times cheaper than petrol or diesel to drive in UK"

phtrivier — Mon, 28 Sep 2026 10:33:55 +0000

Meta: As plenty of commenters have pointed out, this kind of statements (and similar ones like "PV electricity is the cheapest one", etc...) are entirely true _with some attached caveats_.

For example, an EV is both undoubtably cheaper to operate than a gas powered car ; and infinitely more expensive to operate if you can't afford to buy one in the first place.

Similarly, the price of kWh of PV-generated electricity is both lower than pretty much everything else as this point ; and, basically infinite when the sun is down and your 4h battery has run out.

Is there a proper math concept to fairly express those things ? "Mean", "Average" or "Median" fall short. Mean with very very large error bars ? "Average with substitution" ?

In the end, I can't wait for all those circumvoltions to be irrelevant (when EVs are just cheap to buy too, and electricity has enough storage to bake it into the price, etc...)

New comment by phtrivier in "Pentagon says overreliance on AI contributed to missile strike on Iran school"

phtrivier — Tue, 22 Sep 2026 22:04:30 +0000

There's a lot of room between "no consequence for anyone" and "Trump gets to the Hague because of that".

If I remember correctly, some of the people responsible for torturing iraqui prisonniers did get sentenced. They were definitely scapegoats of larger organisations. But if the "lampistes" get caught, this may have _some_ effects on otherw potential _lampistes_.

Palantie execs will be fine, of course. I'm.not expecting them to grow a conscience at this stage. So they will shrug it off, and go on with creating shareholders value. Until they're sacked by their board for not meeting some quotas, and have to live with _less money_ - which is the closest thing to the seventh circle of hell for those person, anyway.

New comment by phtrivier in "US Revokes Limits on Power Plants' Climate Pollution"

phtrivier — Sun, 20 Sep 2026 20:35:14 +0000

From the wording, it's not clear if the ruling (from 2024) had any impact on power plant running today, or newly built one (it seems like it was mandating carbon capture eventually.)

The "good" news is that coal plants are already suboptimal, economically. This is delaying the closure of existing plans, which is just absurd, but completely on brand for the political party that people voted in office, with complete knowledge of what would happen.

I suppose they also voted for "cheaper groceries" and "no more endless wars", but that will come any time soon.

New comment by phtrivier in "Tech Fascism Has Come for American Democracy"

phtrivier — Thu, 17 Sep 2026 12:09:15 +0000

It's a known problem, for which we've also had a pretty robust workaround since Montesquieu : distribute separate part of power to _many_ craving people who disagree and can lock each other's.

I'm all for adding a branch that would be primarily picked at random (and I'm all for having a branch that is only recruited among peers of expertise, with fréquent rotation as a way to limit the risk of connivence, etc...)

New comment by phtrivier in "Tech Fascism Has Come for American Democracy"

phtrivier — Thu, 17 Sep 2026 07:08:13 +0000

Do you have a link to any study about that ? That would probably not happen in Europe, but at least in the US, ethnicities can be studied, so any bias could be visible statistically ?