<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: BenoitP</title><link>https://news.ycombinator.com/user?id=BenoitP</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sun, 06 Sep 2026 19:13:13 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=BenoitP" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by BenoitP in "Virtual Threads for a scripting language in Java 8 without Loom"]]></title><description><![CDATA[
<p>Another project worth checking would be Quasar (and had virtual threads in 2013):<p><a href="https://docs.paralleluniverse.co/quasar/" rel="nofollow">https://docs.paralleluniverse.co/quasar/</a><p>But I believe the author moved to a Big Corp which sells databases, working on some  big project which took all his time...</p>
]]></description><pubDate>Fri, 04 Sep 2026 14:05:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=49564853</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49564853</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49564853</guid></item><item><title><![CDATA[New comment by BenoitP in "Carbon-aware electricity pricing, measured daily on 38 grids"]]></title><description><![CDATA[
<p>France exports about 60TWh of electricity yearly for about 4B€.<p>Flamanville running at 100% generates 32TWh. It is amortized in 9 years. I'd say it's doing pretty okay even with its ballooning cost.</p>
]]></description><pubDate>Fri, 04 Sep 2026 13:52:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=49564684</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49564684</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49564684</guid></item><item><title><![CDATA[New comment by BenoitP in "Carbon-aware electricity pricing, measured daily on 38 grids"]]></title><description><![CDATA[
<p>Took your link, changed de to en, reduced granularity weekly to daily. It looks way more spiky:<p><a href="https://www.energy-charts.info/charts/energy/chart.htm?l=en&c=EU&interval=day&legendItems=lw3w1&partsum=0&year=2025" rel="nofollow">https://www.energy-charts.info/charts/energy/chart.htm?l=en&...</a><p>Also, both wind and solar (especially solar) change vastly under a sub-daily aggregation.</p>
]]></description><pubDate>Fri, 04 Sep 2026 13:39:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=49564472</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49564472</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49564472</guid></item><item><title><![CDATA[New comment by BenoitP in "Carbon-aware electricity pricing, measured daily on 38 grids"]]></title><description><![CDATA[
<p>It'd be really great if wind turbines across Germany were producing at uncorrelated times.<p>Let's look at the wind right now: uniform wind across Germany:<p><a href="https://earth.nullschool.net/#2026/09/04/1200Z/wind/surface/level/orthographic=-348.66,46.32,2614" rel="nofollow">https://earth.nullschool.net/#2026/09/04/1200Z/wind/surface/...</a><p>Let's look at the wind earlier: uniformly no wind across Germany:<p><a href="https://earth.nullschool.net/#2026/09/02/0000Z/wind/surface/level/orthographic=-348.50,48.23,2614" rel="nofollow">https://earth.nullschool.net/#2026/09/02/0000Z/wind/surface/...</a><p>Hmmmm.</p>
]]></description><pubDate>Fri, 04 Sep 2026 13:30:03 +0000</pubDate><link>https://news.ycombinator.com/item?id=49564356</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49564356</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49564356</guid></item><item><title><![CDATA[New comment by BenoitP in "US diesel prices hit a record high of $5.85 on average"]]></title><description><![CDATA[
<p>Diesel, or rather energy, is also a way bigger part of every single aspect of our modern world. Let's take drinking a can of coke as an example:<p>First degree: It came into your home from the factory through the shop. 80% product + 20% diesel.<p>Second degree: Product is aluminum can + water + ingredients. Aluminium came to be a can through 70% refined aluminum + 15% diesel (transport) + 15% electrical energy (pressing the can). Water was pumped locally: free from the ground + 100% electrical energy. Ingredients came from a farm: 80% local product + 20% diesel.<p>Third degree: Electrical energy came from 40% gas turbines, 30% hydro, 30% wind farm. Ingredients from farms came from soil 40%, 35% nitrogen fertilizer (== gas), 25% diesel (machines)<p>... and so on<p>If you decompose everything all the way, the global tally is either matter, which is given to us by nature for free. Or energy; The sources of which being vastly composed of fossil fuels. Fossil fuel prices affect everything.<p>These fossil resources being finite we can also conclude we're experiencing a centuries-long hay-fire. But hey, we're merely plateauing right now; and most of us won't be here for the harder part of the crash.<p>And I didn't even mention climate change.</p>
]]></description><pubDate>Fri, 04 Sep 2026 13:12:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=49564155</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49564155</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49564155</guid></item><item><title><![CDATA[New comment by BenoitP in "Benchmarking Vector Indexes"]]></title><description><![CDATA[
<p>What's wrong with <a href="https://ann-benchmarks.com/index.html" rel="nofollow">https://ann-benchmarks.com/index.html</a>
?</p>
]]></description><pubDate>Tue, 01 Sep 2026 16:39:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=49524402</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49524402</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49524402</guid></item><item><title><![CDATA[New comment by BenoitP in "Harvest (IBM 7950): Supercomputer for cryptanalysis at the NSA in the Cold War"]]></title><description><![CDATA[
<p>They had vertical integration and made bank. Then each layer got more complex as computing became mainstream, and IBM got out-competed on everyone of them by specialized companies. And then they sold them off. The only layer that remains is the most strategic: the customer relationship.<p>IBM is now just a consultancy.</p>
]]></description><pubDate>Wed, 26 Aug 2026 07:45:51 +0000</pubDate><link>https://news.ycombinator.com/item?id=49445355</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49445355</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49445355</guid></item><item><title><![CDATA[The Proof Machine (2016)]]></title><description><![CDATA[
<p>Article URL: <a href="https://incredible.pm/">https://incredible.pm/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=49068711">https://news.ycombinator.com/item?id=49068711</a></p>
<p>Points: 27</p>
<p># Comments: 4</p>
]]></description><pubDate>Mon, 27 Jul 2026 12:29:24 +0000</pubDate><link>https://incredible.pm/</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49068711</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49068711</guid></item><item><title><![CDATA[New comment by BenoitP in "An Introduction to Formal Logic(2020) [pdf]"]]></title><description><![CDATA[
<p>To the more exercise-inclined, there's this introductory game:<p><a href="https://incredible.pm/" rel="nofollow">https://incredible.pm/</a></p>
]]></description><pubDate>Mon, 27 Jul 2026 12:28:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=49068693</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=49068693</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49068693</guid></item><item><title><![CDATA[New comment by BenoitP in "Tell HN: I'm tired of AI-generated answers"]]></title><description><![CDATA[
<p>Same here<p>A potential solution would be to have a reputation system, with some light pageranking in it. People who care about their reputation are bound to rate other likewise people higher. I don't think such loyalty exists between vibetalkers.<p>It's obvious we're not alone, and 2 distinct clusters would form (with people on a spectrum in between)</p>
]]></description><pubDate>Fri, 22 May 2026 06:59:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=48232874</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=48232874</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48232874</guid></item><item><title><![CDATA[New comment by BenoitP in "Beyond Semantic Similarity"]]></title><description><![CDATA[
<p>Not doing retrieval, but one thing I'd love to try would be active learning passes of querying. Here is a paper for large scale active learning<p><a href="https://cdn.aaai.org/ojs/20591/20591-13-24604-1-2-20220628.pdf" rel="nofollow">https://cdn.aaai.org/ojs/20591/20591-13-24604-1-2-20220628.p...</a><p>Steps would be:<p>- A first hybrid query with bm25+query vector<p>- rank some first results locally (ask the LLM for relevance, usefulness? ask the human?)<p>- query in active learning manner; providing positive and negative examples to orient the direction of better results. It'd silence the polluting components from the original query<p>- iterate<p>If anyone's hiring for remote roles, I'd love to give that technique a try</p>
]]></description><pubDate>Wed, 13 May 2026 07:00:01 +0000</pubDate><link>https://news.ycombinator.com/item?id=48118690</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=48118690</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48118690</guid></item><item><title><![CDATA[New comment by BenoitP in "UpDown: Efficient Manycore based on Many Threading & Scalable Memory Parallelism"]]></title><description><![CDATA[
<p>Yet another manycore proposal, but I feel each time we're getting closer. Bandwidth is one of the few dimensions still growing, and there's something to unlock by shaping computation by this firehose of bits.<p>However it seems the software part is always a blocker, and these architecture only address well a handful of program types.<p>Is this time different? I think it is. The paradigm here is about small threads that yield very often, and we have that way of programming: Erlang, Go, Java's virtual threads.</p>
]]></description><pubDate>Tue, 14 Apr 2026 07:31:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=47762432</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47762432</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47762432</guid></item><item><title><![CDATA[Full System Cost of a Resilient and Carbon Neutral Electricity System [pdf]]]></title><description><![CDATA[
<p>Article URL: <a href="https://unece.org/sites/default/files/2025-09/GECES-21_2025_INF.2%20-%20Understanding%20the%20Full%20System%20Cost%20of%20the%20Electricity%20System.pdf">https://unece.org/sites/default/files/2025-09/GECES-21_2025_INF.2%20-%20Understanding%20the%20Full%20System%20Cost%20of%20the%20Electricity%20System.pdf</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47678127">https://news.ycombinator.com/item?id=47678127</a></p>
<p>Points: 4</p>
<p># Comments: 0</p>
]]></description><pubDate>Tue, 07 Apr 2026 16:49:54 +0000</pubDate><link>https://unece.org/sites/default/files/2025-09/GECES-21_2025_INF.2%20-%20Understanding%20the%20Full%20System%20Cost%20of%20the%20Electricity%20System.pdf</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47678127</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47678127</guid></item><item><title><![CDATA[New comment by BenoitP in "Has electricity decoupled from natural gas prices in Germany?"]]></title><description><![CDATA[
<p>LCOE is good for marginal cost (eg: one more solar panel), but fails dramatically at evaluating systemic costs.<p>A nuclear reactor moves the entire market down, including the costs to the consumer when he buys solar energy.<p>Here is a UN document explaining it: <a href="https://unece.org/sites/default/files/2025-09/GECES-21_2025_INF.2%20-%20Understanding%20the%20Full%20System%20Cost%20of%20the%20Electricity%20System.pdf" rel="nofollow">https://unece.org/sites/default/files/2025-09/GECES-21_2025_...</a></p>
]]></description><pubDate>Tue, 07 Apr 2026 16:44:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=47678055</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47678055</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47678055</guid></item><item><title><![CDATA[New comment by BenoitP in "Jax's true calling: Ray-Marching renderers on WebGL"]]></title><description><![CDATA[
<p>JAX is designed from the start to fit well with systolic arrays (TPUs, Nvidia's tensor cores, etc), which are extremely energy-efficient. WebGL won't be the tool that connects it on the web, but the generation after WebGPU will.</p>
]]></description><pubDate>Thu, 02 Apr 2026 07:24:27 +0000</pubDate><link>https://news.ycombinator.com/item?id=47611091</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47611091</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47611091</guid></item><item><title><![CDATA[New comment by BenoitP in "Jax's true calling: Ray-Marching renderers on WebGL"]]></title><description><![CDATA[
<p>I believe we could get there eventually. For example for collision there is work to make it differentiable (or use a local surrogate at the collision point):
<a href="https://arxiv.org/abs/2207.00669" rel="nofollow">https://arxiv.org/abs/2207.00669</a><p>The robotics will need to connect vision with motors with haptics with 3D modelling. And to propagate gradient seamlessly. For calibrating torque with the the elastic deformation of the material for example. After all matter is not discreet at small scales (staying above the atomic scale)<p>All this will require all modules to be compatible with differentiability. It'll be expensive at first, but I'm sure some optimizations can get us close to the discreet case.<p>Also even for meshes there is a lot to gain with trying to go the continuous way:<p><a href="https://www.cs.cmu.edu/~kmcrane/Projects/DDG/" rel="nofollow">https://www.cs.cmu.edu/~kmcrane/Projects/DDG/</a></p>
]]></description><pubDate>Thu, 02 Apr 2026 07:19:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=47611052</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47611052</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47611052</guid></item><item><title><![CDATA[New comment by BenoitP in "Jax's true calling: Ray-Marching renderers on WebGL"]]></title><description><![CDATA[
<p>Yeah :)<p>I had a lot of fun writing the article! And it is only half a joke<p>My intuition for so-called world models is that we'll have to plug modules, each responsible for a domain (text, video, sound, robot-haptics, physical modelling) It'll require to plug modules in a way that will allow the gradient to propagate. A differentiable architecture. And JAX seems well placed for this by making function manipulation a first citizen. Looking at your testimony comforts me in this view</p>
]]></description><pubDate>Thu, 02 Apr 2026 07:07:45 +0000</pubDate><link>https://news.ycombinator.com/item?id=47610964</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47610964</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47610964</guid></item><item><title><![CDATA[New comment by BenoitP in "Jax's true calling: Ray-Marching renderers on WebGL"]]></title><description><![CDATA[
<p>Damn, I should have spent more time QA-ing that post. I'll try to patch it.<p>You did not miss much though: it just rotates the scene.</p>
]]></description><pubDate>Thu, 02 Apr 2026 06:58:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=47610885</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47610885</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47610885</guid></item><item><title><![CDATA[Jax's true calling: Ray-Marching renderers on WebGL]]></title><description><![CDATA[
<p>Article URL: <a href="https://benoit.paris/posts/jax-ray-marcher/">https://benoit.paris/posts/jax-ray-marcher/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47605494">https://news.ycombinator.com/item?id=47605494</a></p>
<p>Points: 91</p>
<p># Comments: 14</p>
]]></description><pubDate>Wed, 01 Apr 2026 19:36:07 +0000</pubDate><link>https://benoit.paris/posts/jax-ray-marcher/</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47605494</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47605494</guid></item><item><title><![CDATA[New comment by BenoitP in "TurboQuant: Redefining AI efficiency with extreme compression"]]></title><description><![CDATA[
<p>It is AI generated. Or was written by someone a bit far from the technical advances IMHO. The Johnson-Lindenstrauss Lemma is a very specific and powerful concept, when in the article the QLJ explanation is vacuous. A knowledgeable human would not have left the reader wanting for how that relates to the Lemma.</p>
]]></description><pubDate>Wed, 25 Mar 2026 08:02:07 +0000</pubDate><link>https://news.ycombinator.com/item?id=47514583</link><dc:creator>BenoitP</dc:creator><comments>https://news.ycombinator.com/item?id=47514583</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47514583</guid></item></channel></rss>